9285.7-09A
April 1992
Guidance for Data Useability in
Risk Assessment
(Part A)
Final
Notice: Guidance for Radioanalytical
Data Useability in Risk Assessment is
Given in Part B
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
NOTICE
The policies and procedures set forth here are intended as guidance to U.S. Environmental Protection Agency and other
government employees. They do not constitute rulemaking by the Agency, and may not be relied on to create a
substantive or procedural right enforceable by any other person. The U.S. Environmental Protection Agency may take
action that is at variance with the policies and procedures in this guidance and may change them at any time without
public notice.
Copies of the guidance can be obtained from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Phone: 703-487-4650
-------
Contents
CHAPTER 1 INTRODUCTION AND BACKGROUND 1
1.1 CRITICAL DATA QUALITY ISSUES IN RISK ASSESSMENT 1
1.1.1 Data Sources 2
Y.I.2 Detection Limits 2
1.1.3 Qualified Data 2
1.1.4 Background Samples : 2
1.1.5 Consistency in Data Collection 2
1.2 FRAMEWORK AND ORGANIZATION OF THE GUIDANCE 3
CHAPTER 2 THE RISK ASSESSMENT PROCESS 7
2.1 OVERVIEW OF BASELINE HUMAN HEALTH RISK ASSESSMENT AND THE EVALUATION OF
UNCERTAINTY 7
2.1.1 Data Collection and Evaluation 11
2.1.2 Exposure Assessment 13
2.1.3 Toxicity Assessment 15
2.1.4 Risk Characterization 17
2.2 ROLES AND RESPONSIBILITIES OF KEY RISK ASSESSMENT PERSONNEL 18
2.2.1 Project Coordination 18
2.22 Gathering Existing Site Data and Developing the Conceptual Model 18
2.2.3 Project Scoping 18
2.2.4 Quality Assurance Document Preparation and Review 20
2.2.5 Budgeting and Scheduling 21
2.2.6 Iterative Communication 21
2.2.7 Data Assessment 22
2.2.8 Assessment and Presentation of Environmental Analytical Data 23
CHAPTER 3 USEABILITY CRITERIA FOR BASELINE RISK ASSESSMENTS 25
3.1 DATA USEABILITY CRITERIA 26
3.1.1 Data Sources 26
3.1.2 Documentation 29
3.1.3 Analytical Methods and Detection Limits 30
3.1.4 Data Quality Indicators 31
3.1.5 Data Review 34
3.1.6 Reports from Sampling and Analysis to the Risk Assessor , 36
3.2 PRELIMINARY SAMPLING AND ANALYTICAL ISSUES 37
3.2.1 Chemicals of Potential Concern 40
3.2.2 Tentatively Identified Compounds 41
3.2.3 Identification and Quantitation 45
3.2.4 Detection and Quantitation Limits and Range of Linearity 47
3.2.5 Sampling and Analytical Variability Versus Measurement Error 50
32.6 Sample Preparation and Sample Preservation 54
3.2.7 Identification of Exposure Pathways 55
3.2.8 Use of Judgmental or Purposive Sampling Design 55
-------
Contents
(cont'd)
3.2.9 Field Analyses Versus Fixed Laboratory Analyses 57
3.2.10 Laboratory Performance Problems 58
CHAPTER 4 STEPS FOR PLANNING FOR THE ACQUISITION OF USEABLE
ENVIRONMENTAL DATA IN BASELINE RISK ASSESSMENTS 63
4.1 STRATEGIES FOR DESIGNING SAMPLING PLANS 63
4.1.1 Completing the Sampling Design Selection Worksheet 65
4.1.2 Guidance for Completing the Sampling Design Selection Worksheet 72
4.1.3 Specific Sampling Issues 76
4.1.4 Soil Depth Issues 79
4.1.5 Balancing Issues for Decision-Making 80
4.1.6 Documenting Sampling Design Decisions 81
4.2 STRATEGY FOR SELECTING ANALYTICAL METHODS 81
4.2.1 Completing the Method Selection Worksheet 83
4.2.2 Evaluating the Appropriateness of Routine Methods 84
4.2.3 Developing Alternatives When Routine Methods are not Available 87
4.2.4 Selecting Analytical Laboratories 87
4.2.5 Writing the Analysis Request 88
4.3 BALANCING ISSUES FOR DECISION-MAKING 88
CHAPTER 5 ASSESSMENT OF ENVIRONMENTAL DATA FOR USEABILITY IN
BASELINE RISK ASSESSMENTS 95
5.1 ASSESSMENT OF CRITERION I: REPORTS TO RISK ASSESSOR 100
5.1.1 Preliminary Reports 100
5.1.2 Final Report 100
5.2 ASSESSMENT OF CRITERION II: DOCUMENTATION 101
5.3 ASSESSMENT OF CRITERION ffl: DATA SOURCES 101
5.4 ASSESSMENT OF CRITERION IV: ANALYTICAL METHOD AND DETECTION LIMIT 102
5.5 ASSESSMENT OF CRITERION V: DATA REVIEW 102
5.6 ASSESSMENT OF CRITERION VI: DATA QUALITY INDICATORS 103
5.6.1 Assessment of Sampling and Analytical Data Quality Indicators 105
5.6.2 Combining the Assessment of Sampling and Analysis 114
CHAPTER 6 APPLICATION OF DATA TO RISK ASSESSMENTS 117
6.1 ASSESSMENT OF THE LEVEL OF CERTAINTY ASSOCIATED WITH THE
ANALYTICAL DATA 117
6.1.1 What Contamination is Present and at What Levels? 117
6.1.2 Are Site Concentrations Sufficiently Different from Background? 119
6.1.3 Are All Exposure Pathways and Areas Identified and Examined? 120
6.1.4 Are All Exposure Areas Fully Characterized? 120
6.2 ASSESSMENT OF UNCERTAINTY ASSOCIATED WITH THE BASELINE RISK ASSESSMENT
FOR HUMAN HEALTH 121
IV
-------
Contents
(cont'd)
APPENDICES
I. DESCRIPTION OF ORGANICS AND INORGANICS'DATA REVIEW PACKAGES 125
II. LISTING OF COMMON POLLUTANTS GENERATED BY SEVEN INDUSTRIES 153
III. LISTING OF ANALYTES, METHODS, AND DETECTION OR QUANTITATION LIMITS FOR
POLLUTANTS OF CONCERN TO RISK ASSESSMENT 167
IV. CALCULATION FORMULAS FOR STATISTICAL EVALUATION 235
V. "J" DATA QUALIFIER SOURCE AND MEANING 239
VI. "R" DATA QUALIFIER SOURCE AND MEANING 245
VII. SUMMARY OF COMMON LABORATORY CONTAMINANTS, CONCENTRATION REQUIRE-
MENTS, AND RISK ASSESSMENT IMPLICATIONS 249
VIII. CLP ANALYTICAL METHODS SHORT SHEETS AND TCL COMPOUNDS 253
IX. EXAMPLE DIAGRAM FOR A CONCEPTUAL MODEL FOR RISK ASSESSMENT 263
-------
Exhibits
1 Data Useability Criteria to Plan Sampling, Analysis and Assessment Efforts in Baseline Risk Assessment 3
2 Organization of the Guidance 5
3 Data Relevant to Components of the Risk Assessment Process 8
4 Baseline Risk Assessment Process and Typical Sources of Uncertainty 9
5 Range of Uncertainty of Risk Assessment 10
6 Development of Conceptual Site Model 12
7 Generic Equation for Calculating Chemical Intakes 16
8 Roles and Responsibilities of Risk Assessment Team Members 19
9 Example Risk Assessment Checklist for Use in Scoping 20
10 Checklist for Reviewing the Workplan 21
11 Checklist for Reviewing the Sampling and Analysis Plan 22
12 Importance of Data Useability Criteria in Planning for Baseline Risk Assessment 26
13 Data Sources and Their Use in Risk Assessment 28
14 Relative Importance of Documentation in Planning and Assessment 30
15 Relevance of Sampling Data Quality Indicators 31
16 Relevance of Analytical Data Quality Indicators 32
17 Alternative Levels of Review of Analytical Data 34
18 Automated Systems to Support Data Review 35
19 Data and Documentation Needed for Risk Assessment 36
20 Importance of Sampling Issues in Risk Assessment 38
21 Sampling Variability and Measurement Error 39
22 Importance of Analytical Issues in Risk Assessment 40
23 Median Coefficient of Variation for Chemicals of Potential Concern 42
24 Munitions Compounds and Their Detection Limits 43
25 Summary of Most Frequently Occurring Chemicals of Potential Concern by Industry 44
26 Steps in the Assessment of Tentatively Identified Compounds 45
27 Requirements for Confident Identification and Quantitation 45
28 Relative Impacts of Detection Limit and Concentration of Concern: Data Planning 46
29 The Relationship of Instrument Calibration Curve and Analyte Detection 48
30 Example of Detection Limit Calculation 49
31 Measurement of Variation and Bias Using Field Quality Control Samples 51
32 Sampling Issues Affecting Confidence in Analytical Results 52
33 Sources of Uncertainty that Frequently Affect Confidence in Analytical Results 53
34 Sample Preparation Issues 54
35 Information Available from Different Sampling Techniques 54
36 Comparison of Sample Preparation Options 56
37 Identification of Exposure Pathways Prior to Sampling Design is Critical to Risk Assessment 57
38 Strengths and Weaknesses of Biased and Unbiased Sampling Designs 58
39 Characteristics of Field and Fixed Laboratory Analyses 59
40 Strengths and Weaknesses of Field and Fixed Laboratory Analyses 60
41 Examples of Spatially and Temporally Dependent Variables 64
42 Examples of Sampling Designs 65
vu
-------
Exhibits
(cont'd)
43 Applicability of Sampling Designs 66
44 Common Sampling Designs 67
45 Hierarchical Structure of Sampling Design Selection Worksheet 68
46 Factors in Determining Total Number of Samples Collected 72
47 Relationships Between Measures of Statistical Performance and Number of Samples Required 73
48 Number of Samples Required to Achieve Given Levels of Confidence, Power and MDRD 76
49 Confidence Levels for the Assessment of Measurement Variability 77
50 Soil Depth Sampling Worksheet 79
51 Automated Systems to Support Environmental Sampling 81
52 Method Selection Worksheet 82
53 Automated Systems to Support Method Selection 84
54 Common Laboratory Contaminants and Interferences by Organic Analyte 85
55 Common Laboratory Contaminants and Interferences by Inorganic Analyte 86
56 Comparison of Analytical Options for Organic Analytes in Water 90
57 Comparison of Analytical Options for Organic Analytes in Soil 91
58 Comparison of Analytical Options for Inorganic Analytes in Water and Soil 92
59 Comparison of Analytical Options for Organic and Inorganic Analytes in Air 93
60 Data Useabiliry Assessment of Criteria 95
61 Minimum Requirements, Impact if Not Met, and Corrective Actions for Data Useability Criteria 96
62 Corrective Action Options When Data Do Not Meet Performance Objectives 97
63 Data Useability Worksheet 98
64 Relative Importance of Detection Limit and Concentration of Concern: Data Assessment 103
65 Consequences of Alternative Sampling Strategies on Total Error Estimate 104
66 Use of Quality Control Data for Risk Assessment 105
67 Steps to Assess Sampling Performance 110
68 Recommended Minimum Statistical Performance Parameters for Risk Assessment Ill
69 Basic Model for Estimating Total Variability Across Sampling and Analysis Components 114
70 Combining Data Quality Indicators From Sampling and Analysis into a Single Assessment of Uncertainty... 115
71 Data Useability Criteria Affecting Contamination Presence 118
72 Data Useability Criteria Affecting Background Level Comparison 119
73 Data Useability Criteria Affecting Exposure Pathway and Exposure Area Examination „ 120
74 Data Useability Criteria Affecting Exposure Area Characterization 121
75 Uncertainty in Data Collection and Evaluation Decisions Affects the Certainty of the Risk Assessment 122
viu
-------
Tips*
The analytical data objective for baseline risk assessments is that uncertainty is known and
acceptable, not that uncertainty be reduced to a particular level, (p. 3)
To maximize data useability for the risk assessment, the risk assessor must be involved from
the start of the Rl. (p. 7)
All data can be used in the baseline risk assessment as long as their uncertainties are clearly
described, (p. 11)
Uncertainty in the analytical data, compounded by uncertainty caused by the selection of the
transport models, can yield results that are meaningless or that cannot be interpreted, (p. 14)
Uncertainties in toxicological measures and exposure assessment are often assumed to be
greater than uncertainties in environmental analytical data; thus, they are assumed to have a
more significant effect on the uncertainty of the risk assessment, (p. 17)
Analytical data collected solely for other purposes may not be of optimal use to the risk
assessment, (p. 20)
Effective planning improves the useability of environmental analytical data in the final risk
assessment.
(p. 25)
Use historical analytical data and a broad spectrum analysis to initially identify the chemicals
of potential concern or exposure areas, (p. 26)
To expedite the risk assessment, preliminary data should be provided to the risk assessor as
soon as they are available, (p. 35)
To protect human health, place a higher priority on preventing false negatives in sampling
and analysis than on preventing false positives, (p. 41)
Use preliminary data to identify chemicals of potential concern and to determine any need to
modify the sampling or analytical design, (p. 41)
Specific analysis for compounds identified during library search can be requested, (p. 41)
The closer the concentration of concern is to the detection limit, the greater the possibility of
false negatives and false positives, (p. 47)
The wide range of chemical concentrations in the environment may require multiple analyses
or dilutions to obtain useable data. Request results from all analyses, (p. 47)
Define the type of detection or quantitation limit for reporting purposes; request the sample
quantitation limit for risk assessment, (p. 47)
When contaminant levels in a medium vary widely, increase the number of samples or
stratify the medium to reduce variability, (p. 50)
Sampling variability typically contributes much more to total error than analytical variability.
(p. 50)
Field methods can produce legally defensible data if appropriate method QC is available and
if documentation is adequate, (p. 57)
To minimize the potential for false negatives, obtain data from a broad spectrum analysis
from each medium and exposure pathway, (p. 58)
The CLP or other fixed laboratory sources are most appropriate for broad spectrum analysis
or for confirmatory analysis, (p. 58)
Solicit the advice of the chemist to ensure proper laboratory selection and to minimize
laboratory and/or methods performance problems that occur in sample analysis, (p. 58)
Use of the Sampling Design Selection Worksheet will help the RPM or statistician determine
an appropriate sampling design, (p. 65)
' For further information, refer to the text. Page numbers are provided.
ix
-------
Tips
(cont'd)
While other designs may be appropriate in many cases, stratified random or systematic
sampling designs are always acceptable, (p. 65)
If the natural variability of the chemicals of potential concern is large (e.g., greater than 30%),
the major planning effort should be to collect more environmental samples, (p. 72)
At least one broad spectrum analytical sample is required for risk assessment, and a
minimum of two or three are recommended for each medium in an exposure pathway, (p.
73)
Collect and analyze background samples prior to the final determination of the sampling
design since the number of samples is significantly reduced if little background
contamination is present, (p. 75)
Systematic sampling supplemented by judgmental sampling is the best strategy for
identifying hot spots, (p. 75)
Focus planning efforts on maximizing the collection ofuseable data from critical samples, (p.
78)
The ability to combine data from different sampling episodes or different sampling
procedures is a very important consideration in selecting a sampling design but should be
done with caution, (p. 78)
Ensure that critical requirements and priorities are specified on the Method Selection
Worksheet so that the most appropriate methods can be considered, (p. 83)
Use routine methods wherever possible since method development is time-consuming and
may result in problems with laboratory implementation, (p. 83)
Analyte-specific methods that provide better quantitation can be considered for use once
chemicals of potential concern have been identified by broad spectrum analysis, (p. 84)
All results should be reported for samples analyzed at more than one dilution, (p. 85)
Field analysis can be used to decrease cost and turnaround time providing data from a broad
spectrum analysis are available, (p. 89)
Focus corrective action on maximizing the useability of data from critical samples, (p. 97)
Use preliminary data as a basis for identifying sampling or analysis deficiencies and taking
corrective action, (p. 100)
Problems in data useability due to sampling can affect all chemicals involved in the risk
assessment; problems due to analysis may only affect specific chemicals, (p. 100)
Qualified data can usually be used for quantitative risk assessments, (p. 105)
Anticipate the need to combine data from different sampling events and/or different
analytical methods, (p. 107)
Determine the distribution of the data before applying statistical measures, (p. 109)
Determine the statistical measures of performance most applicable to site conditions before
assessing data useability. (p. 110)
Use data qualified as U or J for risk assessment purposes, (p. 113)
The major concern with false negatives is that the decision based on the risk assessment may
not be protective of human health, (p. 117)
False negatives can occur if sampling is not representative, if detection limits are above
concentrations of concern, or if spike recoveries are very low. (p. 117)
False positives can occur when blanks are contaminated or spike recoveries are very high. (p.
118)
Statistical analysis may determine if site concentrations are significantly above background
concentrations when the differences are not obvious, (p. 120)
The primary planning objective is that uncertainty levels are acceptable, known and
quantitatable, not that uncertainty be eliminated, (p. 121)
-------
PREFACE
The U.S. Environmental Protection Agency (EPA) has
established a Data Useability Workgroup to develop
national guidance for determining data useahility
requirements needed for environmental data collection
on hazardous waste sites under the Comprehensive
Environmental Response, Compensation, and Liability
Act of 1980 (CERCLA) as amended by the Superfund
Amendments and Reauthorization Act of 1986 (SARA).
Data useability is the process of assuring or determining
that the quality of data generated meets the intended use.
This guidance has been designed by the Risk Assessment
Subgroup of the Data Useability Workgroup to provide
data users with a nationally consistent basis for making
decisions about the minimum quality and quantity of
environmental analytical data that are sufficient to
supponSuperfundrisk assessment decisions, regardless
of which parties conduct the investigation. This
document is the first part (Part A) of the two-part
Guidance for Data Useability in Risk Assessment. Part
B of this guidance addresses radioanalytical issues.
Risk Assessment Guidance for Superfund (RAGS).
Volume I: Human Health Evaluation Manual. Pan A
(EPA 1989a) serves as a general guidance document for
the risk assessment process. Building upon RAGS, an
"interim final" version of Guidance for Data Useability
in Risk Assessment was issued by the Risk Assessment
Subgroup of the Data Useability Workgroup in October
1990. The guidance was issued as "interim final" in
order to obtain and incorporate comments and criticisms
from data users who tested it in real-world situations.
The authors acknowledge the significant help of all who
have provided comments and criticisms. The results
indicate thatmany people react favorably to the guidance
and find it useful in planning a risk assessment or in
evaluating assessments already underway. Issues were
identified where guidance in the interim final needed to
be supplemented or discussed in more detail. These
issues include providing a more detailed discussion of
sampling strategies, incorporating groundwater factors,
addressing soil depth for exposure, and obtaining
background data. Issues concerning data reporting
formats, validation and use of non-CLP data, and
tentatively identified compounds were also identified.
The final version of the guidance provides greater detail
in the discussion of these and other issues.
This guidance provides direction for planning and
assessing analytical data collection activities for the
baseline human health risk assessment, conducted as
pan of the remedial investigation (RI) process.
Although the guidance addresses the baseline risk
assessment within the RI, it is appropriate for use in
the new Superfund Accelerated Cleanup Model
(SACM) where data needs for risk assessment are
considered at the onset of site evaluation. Site-
specific conditions may often require sampling or
analysis beyond the basic recommendations given in
this guidance. The guidance does not directly address
the use of ecological data for purposes other than
baseline risk assessments for human health, although
some considerations have been included when data may
be used tor both ecological and human health evaluation.
This guidance complements guidance provided in RAGS
(EPA 1989a), Guidance for Conducting Remedial
Investigations and Feasibility Studies Under CERCLA
(EPA 1988a), andData Quality Objectives for Remedial
Response Activities: Development Process (EPA 1987a).
RAGS provides the framework for making data quality
assessments in baseline risk assessments, and this
guidance supplements and strengthens important
technical details of the framework by providing direction
on minimum requirements for environmental analytical
data used in baseline risk assessments. As such, it
complements and builds upon Agency guidance for the
development and use of data quality objectives in all
data collection activities.
This guidance is addressed primarily to the remedial
project managers (RPMs) who have the principal
responsibility for leading the data collection and
assessment activities that support the human health risk
assessment and, secondarily, to risk assessors who must
effectively communicate their data needs to the RPMs
and use the data provided to them. Chemists, quality
assurance specialists, statisticians, hydrogeologists and
other technical experts involved in the RI process can
use this guidance to optimize the useability of data
collected in the RI for use in baseline risk assessments.
Comments on the guidance should be sent to:
Toxics Integration Branch
Office of Emergency and Remedial Response
401 M Street, SW (OS-230)
Washington, DC 20460
Phone: 202-260-9486
XI
-------
ACKNOWLEDGEMENTS
This guidance was developed by an EPA workgroup with membership from EPA Headquarters, EPA Regional offices
and representatives of the contractor community. The EPA Risk Assessment Subgroup of the Data Useability
Workgroup provided valuable input regarding the content, approach and organization of the guidance. Members of the
Risk Assessment Subgroup, responsible for generating this guidance, have experience in human health risk assessment,
remedial project management, chemistry, toxicology, hydrogeology, and quality assurance. Technical review was
provided by lexicologists, chemists, quality assurance specialists, engineers, project managers, and statisticians from
both EPA and contractor staff.
Leadership for development of the "interim final" version of this guidance was provided by Data Useability Workgroup
Region III Co-chairpersons Chuck Sands [currently at the Analytical Operations Branch (AOB)] and Claudia Walters,
and Ruth Bleyler of the Toxics Integration Branch (TIB).
Leadership for development of the "final" version of this guidance was provided by Ruth Bleyler and Lisa Matthews of
TIB and Chuck Sands of AOB. We wish to acknowledge Region V and Region VI for their assistance with the
implementation effort for the final version of the guidance.
Members of the Risk Assessment Subgroup include:
Ruth Bleyler Toxics Integration Branch
Richard Brunker USEPA Region III
Rex Bryan Viar & Company
Matt Charsky Office of Waste Programs Enforcement
Skip Ellis CH2MHILL
Gwen Hooten USEPA Region VIII
Dawn loven USEPA Region III
Peter Isaacson Viar & Company
Cindy Kaleri USEPA Region VI
Jim LaVelle USEPA Region VIII
Jim Luey USEPA Region VIII
Jon Rauscher USEPA Region VI
Chuck Sands Analytical Operations Branch
Robin Smith CH2M HILL
Pat Van Leeuwen USEPA Region V
Chris Weis USEPA Region VIII
Leigh Woodruff USEPA Region X
Additional Workgroup participation includes:
Wayne Berman ICF
Ann Marie Burke USEPA Region I
Dorothy Campbell USEPA Region VIII
Judy Hsieh USEPA Region I
Mark Moese Ebasco
Sheila Sullivan USEPA Region V
Hans Waetjen Office of Waste Programs Enforcement
Xlll
-------
Chapter 1
Introduction and Background
This guidance was developed by the U.S. Environmental
Protection Agency (EPA) for remedial projectmanagers
(RPMs), risk assessors, and contractors. It is published
in two parts; this document is Part A. Part B solely
addresses useability issues in radioanalytical sampling
and analysis for risk assessment. Both parts of this
guidance are designed to assist RPMs in maximizing
the useability of environmental analytical data collected
in the remedial investigation (RI) process for baseline
human health risk assessments. Since RPMs, with
assistance from technical experts, oversee the preparation
of workplans and sampling and analysis plans for RI
data collection, it is important for them to understand
the types, quality and quantity of data needed by risk
assessors, and the impact that their data collection
decisions have on the level of certainty of baseline risk
assessments for human health. This guidance provides
detailed approaches and basic recommendations for
both obtaining and interpreting data for risk assessment
that specifically address:
• HowtodesignRIsampUngandanalyticalactivities
that meet the data quantity and data quality needs
of risk assessors,
• Procedures for assessing the quality of the data
obtained in the RI,
• Options for combining environmental analytical
data of varying levels of quality from different
sources and incorporating them into the risk
assessment,
• Procedures for determining the level of certainty
in the risk assessment based on the uncertainty in
the environmental analytical data, and
• Guidelines on the timing and execution of the
various activities in order to most efficiently
produce deliverables.
Although the guidance addresses the baseline risk
assessment within the RI, it is appropriate for use in the
new Superfund Accelerated Cleanup Model (SACM)
where data needs for risk assessment are considered at
the onset of site evaluation.
Risk assessors should be an integral part of the RI
planning process to ensure that adequate environmental
analytical data of acceptable quality andquantity for the
risk assessment are collected during the RI. This
guidance assists risk assessors in communicating their
environmental analytical data needs to the RPMs. Risk
assessors should work closely with the RPMs to identify
and recommend sampling designs and analytical
methods that will maximize the quality of the baseline
risk assessment for human health within the site-related
and budgetary constraints of the RI, and will produce
consistent risk assessments useful to risk managers.
This guidance provides a number of worksheets and
exhibits that can be used as bases for the organization of
sampling or analytical planning or assessment processes.
However, implementation of guidance will be site-
specific, and site personnel should develop and modify
these guidance materials to best suit the conditions at
their site.
Although ecological data useability is not addressed
specifically in this guidance, the chemical data obtained
from site characterization are useable forcertain elements
of the ecological assessment. In an ecological
assessment, the chemicals of potential concern and their
priorities may be different than those of the human
health risk assessment. For example, iron is rarely of
concern in human health risk assessments, but high
levels of iron may pose a threat to aquatic species. Eco-
guidance documents relevant to risk assessment include
Risk Assessment Guidance for Superfund, Volume II:
Environmental Evaluation Manual(EPA 1989b), ECO
Update (EPA 199 la) and Ecological Assessment of
Hazardous Waste Sites: A Field and Laboratory
Reference (EPA 1989c).
1.1 CRITICAL DATA QUALITY ISSUES
IN RISK ASSESSMENT
Five basic environmental data quality issues are
frequently encountered in risk assessments. This
guidance provides procedures, minimum requirements,
and other information to resolve or minimize the effect
of these issues on the assessment of uncertainty in the
risk assessment. The issues affect both the planning for
and the assessment of analytical data for use in RI risk
assessments. The following sections describe these
issues and their impact on data useability, and highlight
the resolutions of these issues.
Acronyms
CLP Contract Laboratory Program
EPA U.S. Environmental Protection Agency
QAPjP quality assurance project plan
RAGS Risk Assessment Guidance for Superfund
RI remedial investigation
RPM remedial project manager
SACM Superfund Accelerated Cleanup Model
-------
1.1.1 Data Sources
Data users must select sampling and analytical
procedures and providers appropriate to the data needs
of each risk assessment. Practical tradeoffs among
detection limits, response time, documentation,
analytical costs, and level of uncertainty should be
considered prior to selecting sampling designs, analytical
methods, and service providers.
The Contract Laboratory Program (CLP) has been the
principal source of analytical data for investigations at
hazardous waste sites. The CLP requires adherence to
specific data acceptance criteria which results in data of
known analytical quality produced in a standardized
package. Another principal source of analytical data is
the EPA Regional laboratory, which often produces
data similar in quality to that of the CLP. Other
analytical sources, such as field analysis or fixed
laboratories (EPA, state, or private), can also produce
data of acceptable quality. Accordingly, RPMs and risk
assessors should seek the source of data that best meets
the data quality needs of the risk assessment Section
4.2 provides guidance for selecting analytical sources.
Held analytical data have been used primarily to aid in
making decisions during sampling. However, recent
advances in technology, when accompanied by sufficient
and appropriate quality control measures, allow field
analytical data to be used in risk assessments with more
frequency and more confidence than in the past By
using field analyses, RPMs can increase the number of
samples to better characterize the site and significantly
decrease sample turnaround time (to provide real-time
decision-making in the field) as long as acceptable data
quality is maintained. Guidance for assessing the
useability and applicability of field analytical data in the
risk assessment process is also provided in Section 4.2.
For any source of monitoring data, RPMs must ensure
that data quality objectives, analytical methods, quality
control requirements and criteria, level of documentation,
and degree and assignment of responsibilities forquality
assurance oversight are clearly documented in the quality
assurance project plan (QAPjP). In addition, the RPM
is responsible for the enforcement of these parameters.
For non-Superfund-lead analyses, the potentially
responsible party, state, or federal agency determines
and documents these parameters. The QAPjP is then
submitted to the RPM for review. In all cases involving
risk assessment, the RPM shouldalways seek the source
of data that best meets the data quality needs of the risk
assessor. The data source chosen must generate data of
known quality.
1.1.2 Detection Limits
Selecting the analytical method to meet the required
detection limits is fundamental to the useability of
analytical data in risk assessments. In addition, the type
of detection limit, such as method detection limit or
sample quantitation limit used in making data quality
decisions affects the certainty of the risk assessment.
Guidance for making these decisions is provided in
Section 4.2. Preliminary remediation goals, as defined
in Risk Assessment Guidance for Superfltnd (RAGS)
Volume I: Human Health Evaluation Manual, Part B
(EPA 1991b), provide criteria to be considered in
evaluating the adequacy of detection limits.
1.1.3 Qualified Data
Laboratories, and individuals conducting independent
data review, affix coded qualifiers to data when quality
control requirements or other evaluation criteria are not
met Data reviewers assess these and many other
criteria to determine the useability of data. Qualified
data must be used appropriately in risk assessments.
Data are almost always useable in the risk assessment
process, as long as the uncertainty in the data and its
impact on the risk assessmentare thoroughly explained.
Section 5.6 describes procedures for incorporating
qualified data and data of varying analytical quality into
the risk assessment.
1.1.4 Background Samples
In conducting a risk assessment, it is critical to distinguish
site contamination from background levels due to
anthropogenic or naturally occurring contamination in
order to determine the presence or absence of
contamination and to compare with background risk.
Analytical data reported near method detection limits
and sample results qualified during data review
complicate the use of background sample data to
determine site contamination. Planningforthecollection
of a sufficient number of background samples from
representative locations increases the certainty in
decisions about the significance of site contamination.
Section 4.1 discusses how statistical analysis and
professional judgment can be combined to design a
sampling program for collecting adequate background
data.
1.1.5 Consistency in Data Collection
Data collection activities may vary among parties
conducting RIs. Consistency in all Superfund activities
is increasingly crucial. All parties collecting
-------
environmental analytical data for baseline risk
assessments for human health should use guidance
provided in Risk Assessment Guidance for Superfund
(RAGS) Volume I: Human Health Evaluation Manual,
Pan A (EPA 1989a) and this guidance to ensure that
baseline risk assessments for human health are conducted
consistently and are protective of the public health.
1.2 FRAMEWORK AND ORGANIZA-
TION OF THE GUIDANCE
This guidance is organized following the usual sequence
used to determine the useability of environmental
analytical data for baseline human health risk
assessments. Exhibit 1 illustrates the conceptual
framework for the guidance. Six criteria are used to
evaluate data useability for baseline risk assessments
for human health:
• Data sources,
• Documentation,
• Available analytical services in terms of analytical
methods and detection limits.
• Data quality indicators,
• Data review, and
• Reports to risk assessor.
These criteria address the five major data quality issues
described in Section 1.1 and other issues that impact
data useability in the risk assessment. The data useability
criteria are applied in RI planning to guide the design of
sampling plans and select analytical methods for the
data collection effort. The criteria are employed again
to assess the useability of the analytical data collected
during the RI, and of data from other studies and
sources, such as site inspections. This guidance also
describes how to determine the uncertainties in the risk
assessment based on the level of uncertainty of the
environmental analytical data, determined using die
data useability criteria.
«• The analytical data objective for baseline
risk assessments is that the uncertainty is
known and acceptable, not that the
uncertainty be reduced to a particular level.
EXHIBIT 1. DATA USEABILITY CRITERIA TO PLAN SAMPLING,
ANALYSIS AND ASSESSMENT EFFORTS
IN BASELINE RISK ASSESSMENT
DEFINING
DATA USEABIUTY
CRITERIA (3.1)
• Data Sources
• Documentation
• Analytical Methods
and Detection Limits
• Data Quality
Indicator*
• Data Review
• Reports to Risk
Assessor
PLANNING
ASSESSING
DETERMINING
SAMPLING
CONSIDERATIONS
' Preliminary Sampling
Issues (3.2)
' Strategies for
Designing
Sampling Plans (4.1)
ANALYTICAL
CONSIDERATIONS
• Preliminary Analytical
Issues (3.2)
• Strategy for Selecting
Analytical Methods
(4.2)
DATA USEABILITY
CRITERIA (5.0)
• Reports to Risk
Assessor
• Documentation
• Data Sources
• Analytical Methods
and Detection Limits
• Data Review
• Data Quality
Indicators
LEVELS
OF
CERTAINTY
FOR
BASELINE
RISK
ASSESSMENT
(6.1)
-------
Exhibit 2 summarizes the purpose of each chapter of
this guidance and highlights how the chapters can best
assist RPMs and risk assessors. Worksheets, assessment
tables, and other aids are used extensively throughout
the guidance. These are tools that can be used "as is,"
or they can be modified for use or used as the basis for
site-specific worksheets or summaries. Chaptercontents
are summarized below.
• Chapter 2—The Risk Assessment Process: This
chapter explains the purpose and objectives of a
baseline human health risk assessment and
describes the four basic elements of a risk
assessment* data collection and evaluation,
exposure assessment, toxicity assessment, and
risk characterization. The chapter discusses the
uncertainties associated with the risk assessment
process and emphasizes the impact of analytical
data quality on each element The roles and
responsibilities of the RPM, the risk assessor, and
others involved in planning and conducting data
collection activities to support the risk assessment
are described.
• Chapter3—Useability Criteria for Baseline Risk
Assessments: Six criteria are defined in this
chapter for interpreting the importance of sample
collection, analytical techniques, and data review
procedures to the useability of analytical data in
risk assessments. The sampling and analytical
issues that need to be addressed hi using these
criteria are discussed. The chapter stresses the
need to consider and plan for risk assessment data
requirements in the early design stages of the RI.
• Chapter 4—Steps for Planning for the Acquisition
of Useable Environmental Data in Baseline Risk
Assessments: This chapter provides explicit
guidance for designing sampling plans and
selecting analytical methods based on the data
quality requirements of baseline risk assessments.
Worksheets for sampling design selection, soil
depth sampling, and method selection are provided
as part of the step-by-step guidance for making
data collection decisions for individual sites.
• Chapter 5—Assessment of Environmental Data
forUseabilityinBaselineRiskAssessments: This
chapter explains how to assess the useability of
site-specific data for risk assessments after data
collection according to the six criteria defined in
Chapter 3. For each assessment criterion, the
chapter defines minimum data requirements and
explains how to determine actual performance
compared to performance objectives and execute
appropriate corrective actions for data critical to
the risk assessment. The chapter also describes
options available to risk assessors for incorporating
analytical data from different sources and varying
levels of quality into the baseline risk assessment.
• Chapter 6—Application of Data to Risk
Assessments: This chapter details procedures for
determining the overall level of uncertainty
associatedwithtberiskassessment Thediscussion
addresses characterization of contaminant
concentrations within exposure areas, determining
the presence or absence of chemicals of potential
concern, and distinguishing site contamination
from background levels.
• Appendices—The appendices provide analytical
and sampling technical reference materials,
including descriptions of generic organic and
inorganic data review packages; listings of
common industrial pollutants; analytical methods
and detection or quantitation limits (see Section
3.2.4 for definitions); common laboratory
contaminants; calculation formulas for statistical
evaluation; information on analytical data
qualifiers; a summary of Contract Laboratory
Program methods with corresponding Target
Compound List compounds and Target Analyte
List anaytes; and an example of a conceptual site
model.
• Index—The index provides cross-references
throughouttheguidance. This is important because
Chapters 3, 4, and 5 present planning and
assessment issues as complementary discussions
that can be viewed independently.
• Tips—Tips, marked with a »', are incorporated
into the text of the chapters. These tips draw
attention to key issues in the text but are not
intended to summarize the discussion in the chapter.
-------
EXHIBIT 2 . ORGANIZATION OF THE GUIDANCE
Chapter 1
Introduction and Background
• Presents critical data useability issues.
• Specifies audience to be primarily RPMs and risk assessors.
• Defines scope and specifies organization of the guidance.
Chapter 2
The Rick Assessment Process
• Explains the elements of a risk assessment and the impact of analytical data quality on each
element.
• Defines the uncertainties in the risk assessment process.
• Describes the roles of the risk assessor, RPM and others involved with the risk assessment
planning and assessment process.
Chapters
Useability Criteria for Baseline Risk Assessments
• Defines six criteria for assessing data useability: data sources, documentation, analytical
methods/detection limits, data quality indicators, data review, and reports to the risk assessor.
• Applies criteria to sampling and analytical issues.
Chapter 4
Steps for Planning for the Acquisition of Useable Environmental Date in Baseline Risk
Assessments
* Provides guidelines for designing sampling plans and selecting analytical methods.
* Provides worksheets to support sampling design selection, soil depth sampling,
and analytical method selection.
Chapters
Assessment of Environmental Date for Useability in Baseline Risk Assessments
• Describes minimum requirements for useable data.
• Explains how to determine actual performance compared to objectives.
• Recommends corrective actions for critical data not meeting objectives.
• Describes options for combining data from different sources and of varying quality into the risk
assessment
Chapters
Application of Date to Risk Assessments
• Provides procedures to determine the uncertainty of the analytical date.
• Explains now to distinguish site from background levels of contamination and determine the
presence (absence) of chemicals of potential concern.
• Discusses how to characterize contaminant concentrations within exposure areas.
Appendices
• Provide technical reference materials for sampling and analysis.
• Describe date review packages and meanings of selected data qualifers.
21-002-002
-------
Chapter 2
The Risk Assessment Process
This chapter is an overview of the data collection and
evaluation issues thataffect the quality and useability of
baseline human health risk assessments. Ecological
risk assessment is not discussed in this guidance. The
discussion focuses on how the quality of environmental
analytical data influences the level of certainty of the
risk assessment and stresses the importance of
understanding data limitations in characterizing risks to
human health.
The chapter has two sections. Section 2.1 is an overview
of baseline human health risk assessment and the
significance of uncertainty in each stage of the risk
assessment process. Section 2.2 summarizes the roles
and responsibilities of key participants in the risk
assessment process.
2.1 OVERVIEW OF BASELINE
HUMAN HEALTH RISK
ASSESSMENT AND THE
EVALUATION OF UNCERTAINTY
The approach to the baseline human health risk
assessment process used for exposure to chemicals of
potential concern is well established. The National
Research Council (NRC) prepared a comprehensive
overviewof this process (NRC1983), which has become
the foundation for subsequent EPA guidance (EPA
1986a, EPA 1989a, EPA 1989b). RAGS, Part A (EPA
1989a), discusses in detail the human health baseline
risk assessment process which is used in the Superfund
program.
The risk assessment process has four components:
• Data collection and evaluation,
• Exposure assessment,
• Toxicity assessment, and
• Risk characterization.
Exhibit 3 lists information sought in each component of
the baseline risk assessment.
Uncertainty analysis is often viewed as the last step in
the risk characterization process. However, as discussed
in detail in RAGS, Part A, uncertainty analysis is a
fundamental element of each component of risk
assessment, and the results for each component require
an explicit statement of the degree of uncertainty. These
results are the bases for estimating the degree of
uncertainty in the risk assessment as a whole. This
chapter reviews the issues that determine the level of
uncertainty in each component of risk assessment.
*• To maximize data useability for the risk
assessment, the risk assessor must be
involved from the start of the Rl.
The importance of obtaining analytical data that fulfill
the needs of risk assessment cannot be overstated. The
risk assessor must be involved from the start of the risk
assessment process to help establish the scope of the
investigation and the design of the sampling and analysis
program.
All analytical data collected for baseline risk assessment
must be evaluated for their useability. The procedures
for evaluating the adequacy of the data are documented,
along with the resulting estimates of the levels of
certainty. Limitations in the analytical data are not the
only source of uncertainty in risk assessment Exhibit
4 identifies some typical sources of uncertainty, inherent
in each component of the risk assessment, which restrict
the depth and breadth of the evaluation. This guidance
deals only with the uncertainty inherentindata collection
and evaluation. Consult RAGS, Part A, for a more
complete discussion of these and other uncertainties.
Acronyms
ATSDR Agency for Toxic Substances and Disease
Registry
DQO data quality objective
EPA U.S. Environmental Protection Agency
CIS Geographical Information System
HEAST Health Effects Assessment Summary Tables
IRIS Integrated Risk Information System
LOAEL lowest-observable-adverse-effect level
NOAEL no-observable-adverse-effect level
NRC National Research Council
PAH polycyclic aromatic hydrocarbon
PCS polychlorinated biphenyl
QA quality assurance
QAPjP quality assurance project plan
QC quality control
RAGS Risk Assessment Guidance for Superfund
RfC reference concentration
RfD reference dose
RI remedial investigation
RME reasonable maximum exposure
RPM remedial project manager
SAP sampling and analysis plan
SOP standard operating procedure
UCL upper confidence limit
-------
EXHIBIT 3. DATA RELEVANT TO COMPONENTS OF
THE RISK ASSESSMENT PROCESS
Risk Assessment
Component
Data
Data Collection and
Evaluation
Background monitoring data for all affected media.
Environmental data for all relevant media.
List of chemicals of potential concern.
Distribution of sampling data.
Confidence limits surrounding estimates of
representative values.
Exposure Assessment
• Release rates.
• Physical, chemical and biological parameters, for
evaluating transport and transformation of site-
related chemicals.
• Parameters to characterize receptors according to their |
activity, behavior and sensitivity.
• Estimates of exposure concentrations for all
chemicals, environmental media and receptors
at risk.
• Estimates of chemical intake or dose for all
exposure pathways and exposure areas.
Toxicity Assessment
• Toxicity values for all chemicals, exposure
pathways, and exposure areas of concern.
• Uncertainty factors and confidence measures for
RfDs; weight-of-evidence classifications for cancer
slope factors.
Risk Characterization
• Hazard quotients and indices.
• Estimates of excess lifetime cancer risk.
• Uncertainty analysis.
21-002-003
-------
EXHIBIT 4. BASELINE RISK ASSESSMENT PROCESS AND
TYPICAL SOURCES OF UNCERTAINTY
I
Exposure Assessment
Assumptions regarding intake
factors, population characteristics,
and exposure patterns may not
adequately characterize exposure
and may result in underestimates or
overestimates of risk.
The degree to which release or
transport models are represen-
tative of physical reality may
overestimate or underestimate risk.
Inappropriate selection of detection
limit can result in overestimate or
underestimate of risk.
Assumption of 100% bioavail-
ability of chemicals in environ-
mental media (soil in particular) may
result in overestimates of risk.
Assumption that chemicals of
potential concern do not degrade or
transform in the environment may
result in underestimates or
overestimates of risk.
Incremental risks associated with
exposure to site-related chemicals
of potential concern cannot be fully
characterized and may result in
underestimates of risk.
Methods used to estimate inhalation
exposure to volatiles, suspended
particulates or dust may
overestimate intake and risk.
Very few percutaneous absorption
factors are available for chemicals
of potential concern. Exposure
from dermal contact may be over-
estimated using conservative
default values.
Data Collection and
Evaluation
Use of inappropriate method
detection limits may result in
underestimates of risk.
Results may overestimate or
underestimate risk when an
insufficient number of
samples are taken.
Contaminant loss during
sampling may result in
underestimates of risk.
Extraneous contamination
introduced during sampling
or analysis may result in
overestimation of risk.
Risk Characterization
• Risk/dose estimates are
assumed to be additive in the
absence of information on
synergism and antagonism.
This may result in over-
estimates or underestimates
of risk.
• Toxicity values are not
available for all chemicals of
potential concern. Risks
cannot be quantitatively
characterized for these
compounds and may result in
underestimates of risk.
• For some chemicals or
classes (e.g., PCBs, PAHs),
in the absence of toxicity
values, the cancer slope
factor or RfO of a highly toxic
class member is commonly
adopted. This approach may
overestimate risks.
1
Toxicity Assessment
' Critical toxicity values are
derived from animal studies
using high dose levels.
Exposures in humans occur
at low dose levels.
Assumption of linearity at
low dose may result in
overestimates or under-
estimates of risk.
Inappropriate selection of
detection limit can result in
overestimates or under-
estimates of risk.
Extrapolation of results of
toxicity studies from
animals to humans may
introduce error and
uncertainty, inadequate
consideration of
differences in absorption,
pharmacokinetics, and
target organ systems, and
variability in population
sensitivity.
There is considerable
uncertainty in estimates of
toxicity values. Critical
toxicity values are subject
to change as new evidence
becomes available. This
may result in overestimates
or underestimates of risk.
Use of conservative high to
low dose extrapolation
models may result in
overestimation of risk.
Source: Adapted from EPA 1989a.
21-002-004
-------
Risk assessment can be a simple operation, using only
screening-level data, or can be comprehensive, requiring
a robust data set designed to support statistical analyses.
Exhibit 5 discusses the range of uncertainty of baseline
risk assessment The first column in Exhibit 5 defines
the range of the analysis from a low to a high degree of
uncertainty. The second column describes the associated
data useability and limitations in the risk analysis.
• The first level of analysis in Exhibit 5 is a
quantitative risk assessment based on a sampling
program that can be statistically analyzed. The
assessment explicitly bounds and quantitates the
uncertainty in all estimates. This analysis may
strive to attain an ideal based upon the complexity
of the site. The assessment is "quantitative" in that
numeric estimates are derived for potentially
adverse non-carcinogenic and carcinogenic effects,
and in that the level of certainty is quantitated.
• The second level of analysis in Exhibit 5 is a
quantitative assessment based on a limited number
of samples or on data that cannot be fully
quantitated. The risk characterization may include
numeric estimates of excess lifetime cancer risks
and the calculation of hazard indices. However,
the level of analytical uncertainty for these
measures may be significant but is either not
quantitated or is estimated. Given the limitations
of the analytical data, only a qualitative evaluation
of the analytical uncertainty is feasible. Most
baseline risk assessments fall within this category.
Bias may need to be determined for its effect on
predicted exposures and consequent risk.
The third level of the continuum is a qualitative
assessment of risk. The assessment is qualitative
because no numeric measures can be derived to
indicate the potential for adverse effects, and the
level of certainty cannot be assessed. The risk to
human health is considered only in general terms.
Qualitative assessments are based upon limited
sources of historical information, such as disposal
records, circumstantial evidenceof contamination,
or preliminary site assessment data.
EXHIBIT 5. RANGE OF UNCERTAINTY OF RISK ASSESSMENT
Rang* of Analya
Description/Limitations
Quantitative Assessment of Risk:
Uncertainty minimized, quantified,
and explicitly stated. Resulting or
final uncertainty may be highly
variable (either high or low).
Risk assessment conducted using well-designed,
robust data sets and models directly applicable to site
conditions. Sampling program, based on geostatistical
or random design, will support statistical analysis of
results. Statistical analysis used to characterize
monitoring data. Confidence limits or probability
distributions may be developed for all key input
variables.
Quantitative Assessment of Risk:
Magnitude of uncertainty
unknown. No explicit quantitative
estimates provided. Qualitative,
tabular summary of factors
influencing risk estimates may be
provided for determination of
possible bias in error.
Risk assessment conducted using data set of limited
quality and size. No meaningful statistical analysis can
be conducted. Results of risk assessment may be
quantified but uncertainty surrounding these measures
cannot be quantified. Only a qualitative statement is
possible. The majority of baseline risk assessments
typically fall within this category.
Qualitative Assessment of Risk:
Only qualitative statement of
uncertainty is possible.
Uncertainty is high.
Risks cannot be quantified due to insufficient monitoring
or modeling data. Qualitative statement of risks based
on historical information or circumstantial evidence of
contaminantion is provided. This evaluation must be
considered a preliminary, screening level assessment.
21-002-006
10
-------
f All data can be used in the baseline risk
assessment as long as their uncertainties
are clearly described.
Risk assessments must sometimes be conducted using
data of limited quantity and of differing quality. When
RPMs and other technical experts involved in the RI
understand the quantity and quality of data required in
risk assessments, they are better able to design data
collection programs to meet these requirements.
2.1.1 Data Collection and Evaluation
Overview of methods for data collection and
evaluation. Data collection begins with a statement of
the risk assessment purpose and a conceptual model of
the current understanding of the problems to be addressed
for the site under investigation. The model draws from
all available historical data (EPA 1989a). It is first
created with a best estimate of the types and
concentrations of chemicals, or of key chemicals that
are likely to be present, given the history of the site. Site
records, site maps, the layout of existing structures,
topography, and readily observable soil, water and air
characteristics on and off the site help to estimate
chemicals of potential concern, likely important exposure
pathways, potentially exposed populations, and likely
temporal and spatial variation. All of these elements
comprise the conceptual model (Exhibit 6 and Appendix
IX). Once the conceptual model has been developed
and information has been disseminated to project staff,
the site is scoped to identify data gaps and requirements
for the baseline risk assessment.
Several key issues that are part of the development of
data quality objectives (DQOs) should be addressed at
scoping (Neptune, et. ol. 1990):
• The types of data needed (e.g., environmental,
lexicological),
• How the data will be used (e.g., site character-
ization, extent of plume, etc., what chemicals of
concern will drive the risk-based decision), and
• The desired level of certainty for the conclusions
derived from the analytical data (e.g., what are the
probabilities of false positive and false negative
results as a function of risk and concentration).
Carefully designed sampling and analysis programs
minimize the subsequent need to qualify the
environmental data during the data assessment phase.
The objective of the data collection effort is to produce
data that can be used to assess risks to human health with
a known degree of certainty.
A complete list of chemicals of potential concern is
produced when the analytical data have been collected
and evaluated. This list of analytes is the focus of the
risk assessment. EPA no longer advocates the selection
of "indicator compounds," because this practice may
not accurately reflect the total risk from exposure to
multiple site chemicals of potential concern, nor does it
improve the quality or accuracy of the risk assessment
(EPA 1989a).
Uncertainty in data collection and evaluation. Four
principal decisions must be made during data collection
and evaluation in the risk assessment:
• The presence and levels of contaminants at the site
at a predefined level of detail,
• If the levels of site-related chemicals differ
significantly from their background levels,
• Whethertheanalyticaldataareadequatetoidentify
and examine exposure pathways and exposure
areas, and
• Whether the analytical data are adequate to fully
characterize exposure areas.
These decisions are examined in detail in subsequent
chapters. The discussion in this section introduces basic
concepts.
Determining what contamination is present and at
what level. Once a site is suspected to be contaminated
and chemicals of potential concern have been identified,
the levels of chemical contamination in the affected
environmental media must be quantitated to derive
exposure and intake estimates. Estimates of the site
contamination must be produced, with explicit
descriptions of the degree of certainty associated with
the concentration values.
Variability in observed concentration levels arises from
a combination of variance in sampling characteristics of
the site, in sampling techniques, and in laboratory
analysis. The key issue in optimizing the useability of
data for risk assessment is to understand, quantify, and
minimize these variabilities.
EPA's objective is to protect human health and the
environment. Therefore, the design of RI programs is
intended to minimize two potential errors:
• Not detecting site contamination that is actually
present (i.e., false negative values), and
• Derivingsiteconcentrationsthatdonotaccurately
characterize the magnitude of contamination.
11
-------
UJ
0
o
UJ
O)
UJ
o
o
o
u.
O
H
UJ
2
a.
O
UJ
ui
a
ffi
X
X
UI
12
-------
Determining if site concentrations differ significantly
from background concentrations. A fundamental
decision in baseline risk assessments is whether the site
poses an increased risk to human health and the
environment. The decision depends on the degree of
certainty that the background concentrations are
significantly different from the concentrations of the
chemicals of potential concern at the site. Generally,
this question can be confidently answered only if the
design of the sampling program accommodates the
collection of both site and background samples and if
the selection of analytical methods is appropriate.
The differences between site and background
concentrations is evaluated by comparing observed
levels of chemicals of potential concern at the site with
measured background concentrations of the same
chemicals in the same environmental media.
Statistically, this is a test of the null hypothesis, that the
mean concentration of a chemical at the study area is not
significantly different from the mean concentration of
the chemical at the background location. (Historical on-
site levels or nearby off-site levels may be used to
supplement background data. An example of an off-site
area is the 4-mile radius used for the air exposure
pathway in the Hazard Ranking System.) If data from
background samples are clearly different from the results
of site monitoring (e.g., mean chemical concentrations
differ consistently by twoordersofmagnitude),statistical
analysis of the data may not be necessary. Under such
circumstances, RAGS indicates that the primary issue is
establishing a reliable representation of the extent of the
contaminated area. Determining extentofcontamination
is not discussed in this guidance and involves different
decisions, DQOs, and sampling designs. If the results
of site monitoring are less than two orders of magnitude
above background, the procedures used for sampling
and analysis for risk assessment should follow the
recommendations of Chapter 4.
The null hypothesis is always evaluated and accepted or
rejected with a specified level of certainty. This level of
certainty is defined by the significance, or confidence,
level. A type I error is the probability that the null
hypothesis is rejected when in fact it is true (which
contributes to false positive conclusions). A type II
error is the probability that the null hypothesis is accepted
when it is false (a false negative conclusion). How
sampling and analysis design affects the likelihood of
these two types of errors is described in Chapter 4.
Evaluating whether analytical data are adequate to
identify and examine exposure pathways and their
exposure areas. Identifying and delineating exposure
pathways and their exposure areas are important in
identifying potentially exposed populations and for
developing intake estimates. In the baseline risk
assessment, the risk assessor combines data on
contamination with information on human activity
patterns to identify exposure pathways and to determine
the exposure area. The ability to accomplish this
depends on the adequacy of analytical data.
Sampling should be designed to provide representative
data for exposure areas at a site, to address hot spots, to
evaluate the transport of site-related chemicals of
potential concern, and to facilitate the identification of
all exposure pathways. A well-designed sampling and
analysis program results in data of known quality and
quantification of spatial and temporal variability; it
specifies how to interpret the magnitude of observed
values (such as by comparison with background levels
or some other benchmark). Analytical data should
characterize the extent of contamination at the site in
three dimensions.
Evaluating whether analytical data are adequate to
fully characterize exposure areas. Heterogeneity
should be considered in the environmental medium
under evaluation. Hot spots need to be identified and
characterized. Neptune, et. al. 1990, have proposed the
concept of an "exposure unit" as the area over which
receptors integrate exposure. This concept establishes
a basis for summarizing the results of monitoring and
transportmodeling. Tbesamplingandanalysisprogram
must be designed to enable the risk assessor to refine the
initial characterization of exposure pathways and to
spatially and temporally identify the critical areas of
exposure.
2.1.2 Exposure Assessment
Overview of methods for exposure assessment. The
objectives of the exposure assessment are:
• To identify or define the source of exposure,
• To define exposure pathways along with each of
their components (e.g., source, mechanism of
release, mechanism of transport, medium of
transport, etc.),
• To identify potentially exposed populations
(receptors), and
• To measure or estimate the magnitude, duration,
and frequency of exposure to site contaminants for
each receptor (or receptor group).
Actions at hazardous waste sites are based on an estimate
of the reasonable maximum exposure (RME) expected
to occur under both currentand future conditions of land
use (EPA 1989a). EPA defines the RME as the highest
exposure that is reasonably expected to occur at a site
13
-------
over time. RMEs are estimated for individual pathways
and combined across exposure pathways if appropriate.
Once potentially exposed populations are identified,
environmental concentrations at points of exposure
must be determined or projected. Intake estimates (in
mg/kg-day) are then developed for each chemical of
potential concern using a conservative estimate of the
average concentration to which receptors are exposed
over the exposure period. (RAGS recommends a 95%
upper confidence limit (UCL) on the arithmetic mean.)
The concentration estimate is then combined with other
exposure parameters (e.g., frequency, duration, and
body weight) to calculate intake.
In the risk assessment report, estimates of intake are
accompanied by a full description (including sources)
of the assumptions made in their development This
information may be used subsequently in sensitivity
and uncertainty analyses in the risk characterization.
Uncertainty analysis in exposure assessment.
Exposure assessments can introduce a great deal of
uncertainty into the baseline risk assessment process.
Small measures of uncertainty in each of the input
parameters which comprise an exposure scenario may
result in substantial uncertainty in the final assessment.
The largest measure of uncertainty is associated with
characterizing transport and transformation of chemicals
in the environment, establishing exposure settings, and
deriving estimates of chronic intake. The ultimate
effect of uncertainty in the exposure assessment is an
uncertain estimate of intake.
The following sections discuss the significance of the
uncertainty in the analytical data set on selected aspects
of exposureassessment Foramorecomplete discussion
of the exposure assessment process, the readeris referred
to RAGS, Part A.
Characterizing environmental fate, identifying
exposure pathways, and identifying receptors at
risk. An evaluation of the transport and transformation
of chemicals in the en vironmentisconductedfor several
reasons:
• To understand the behavior of site-related
chemicals of potential concern,
• To project the ultimate disposition of these
chemicals,
• To identify exposure pathways and receptors
potentially at risk, and
• To characterize environmental concentrations at
the point of exposure.
These evaluations cannot be accomplished with any
degree of certainty if the analytical data are inadequate.
Monitoring data are most appropriately used to estimate
current or existing exposure when direct contact with
contaminated environmental media is the primary
concern. Modeling may be required, however, in order
toevaluate the potential for future exposure, or exposure
at a distance from the source of release, or to predict
present concentrations where measurement is too costly.
In each case, success in estimating potential exposures
depends heavily on the adequacy of the analytical data.
Environmental fate and transport assessment often uses
models to estimate concentrations in environmental
media at points distant from the source of release.
Models, of necessity, are simplifications of a real,
physical system. Consequently, it is critical that the
limitations of the model (the way that the model differs
from reality) be understood and considered when
applying the model to a particular site. The degree to
which the model differs from reality (in critical areas of
the analysis) contributes to the uncertainty of the analysis.
Transport models are commonly selected for their utility
in describing or interpreting a set of monitoring data.
Chemical transport models must be carefully selected
for their ability to meaningfully characterize the behavior
of chemicals in the environmental medium for the
specific site under investigation. Models that are
inappropriate for the geophysical conditions at the site
will result in errors in the exposure assessment. For
example, the model may be designed to predict
contaminant movement through sand, while soils at the
site are primarily made up of clay. Additionally, if the
analytical data set is severely limited in size or does not
accurately characterize the nature of contamination at
the site, a transport model cannot be properly selected or
accurately calibrated. This introduces additional
uncertainty.
•• Uncertainty in the analytical data,
compounded by uncertainty caused by the
selection of the transport models, can yield
results that are meaningless or that cannot
be interpreted.
Estimating chemical intake. Uncertainties in all
elements of the exposure assessment come together,
and are compounded, in the estimate of intake. It is here
that the professional judgment of the risk assessor is
particularly important The risk assessor must examine
and interpret a diversity of information:
• Thenature,extentandmagnitucteofcontaminabon,
• Results of environmental transport modeling,
• Identification of exposure pathways and areas.
14
-------
• Identification of receptor groups currently exposed
and potentially exposed in the future, and
• Activity patterns and sensitivities of receptors and
receptor groups.
Based on this information, the risk assessor characterizes
the exposure setting and quantifies all parameters needed
in the equations to estimate intake (EPA 1989a).
Chemical intake is a function of the concentration of the
chemical at the point of contact, the amount of
contaminated medium contacted per unit time or event,
the exposure frequency and duration, body weight, the
ability of the chemical to penetrate the exchange
boundary, and the average time period during which
exposure occurs. Exhibit 7 is the generic form of the
intake equation used in exposure assessment.
The specific form of the intake equation varies depending
upon the exposure pathway under consideration (e.g.,
ingestion, inhalation, dermal contact) (EPA 1989a).
Each of the variables in these equations, including
chemical concentration, is commonly characterized as
a point estimate. However, each intake variable in the
equation has a range of possible values. Site-specific
characteristics determine the selection of the most
appropriate values. In an effort to increase consistency
among Superfundriskassessments, EPA has established
standardized exposure parameters to be used when site-
specific data are unavailable (EPA 1991b). Note that
the combination of all factors selected should result in
an estimate of reasonable maximum exposure for each
chemical in each pathway (EPA 1989a).
For most risk assessments, it may not be possible, nor
necessarily advantageous, to develop a quantitative
uncertainty analysis. In these cases, a summary of
major assumptions and their anticipated effects on final
exposure estimates should be included to provide a
qualitative characterization of the level of certainty in
the intake estimates.
2.1.3 Toxicity Assessment
Overview of methods for toxicity assessment. The
objectives of toxicity assessment are to evaluate the
inherent toxicity of the compounds at the site, and to
identify and select toxicity values to evaluate the
significance of receptor exposure to.these compounds.
Toxicity assessments rely on scientific data available in
the literature on adverse effects on humans and
nonhuman species.
Several values of toxicity are important in human health
risk assessments. Reference doses (RfDs) and reference
concentrations (RfCs) are used for oral and inhalation
exposure, respectively, to evaluate non-carcinogenic
and developmental effects; cancer slope factors and unit
risk estimates are used for the oral and inhalation
pathways for carcinogens.
RfDs and RfCs are values developed by EPA to evaluate
the potential for non-carcinogenic effects in humans.
The RfO is defined as an estimate (with uncertainty
spanning an order of magnitude or more) of a daily
exposure level for human populations, including
sensitive sub-populations, that is likely to be without an
appreciable risk of adverse health effects over the
period of exposure (EPA 1989a). Subchronicor chronic
RfDs may be derived for a chemical for intermediate or
long-term exposure scenarios. These values are typically
derived from the no-observable-adverse-effect level
(NOAEL) or the lowest-observable-adverse-effect level
(LOAEL) and the application of uncertainty and
modifying factors (EPA 1989a). Uncertainty factors
are used to account for the variation in sensitivity of
human sub-populations and the uncertainty inherent in
extrapolating the results of animal studies to humans.
Modifying factors account for additional uncertainties
in the studies used to derive the NOAEL or LOAEL.
Cancer slope factors and unit risk values are defined as
plausible, upper-bound estimates of the probability of
cancer response in an exposed individual, per unit
intake over a lifetime exposure period (EPA 1989a).
EPA commonly develops slope factors for carcinogens
with weight-of-evidence classifications that reflect the
likelihood that the toxicant is ahuman carcinogen (EPA
1989a).
To reduce variability in lexicological values used for
risk assessment, a standardized hierarchy of available
lexicological data is specified for Superfund. The
primary source of information for these data is the
Integrated Risk Information System (IRIS) database
(EPA 1989d). IRIS consists of verified RfDs, RfCs,
cancer slope factors, unit risks, and other health risk and
EPA regulatory information. Data in IRIS are regularly
reviewed and updated by an EPA workgroup. If toxicity
values are not available in IRIS, the EPA Health Effects
Assessment Summary Tables (HEAST) (EPA 1990a)
are used as a secondary current source of information.
Additional sources of toxicity information are provided
in RAGS.
The toxicity assessment is conducted parallel with the
exposure assessment, but may begin as early as the data
collection and evaluation phase. As chemicals of
potential concern are identified at the site, the lexicologist
begins to identify the appropriate toxicity values. A
well-designed sampling andanalysis program facilitates
timely identification of the chemicals that will be the
focus of the risk assessment.
15
-------
EXHIBIT 7. GENERIC EQUATION FOR
CALCULATING CHEMICAL INTAKES
= CX
/CR x EFP\ _L
V BW /* AT
Where'
I = intake; the amount of chemical at the exchange
boundary (mg/kg body weight-day)
Chemical-related variable
C = chemical concentration; the average
concentration contacted over the exposure
period (e.g., mg/liter water)
Variables that describe the exposed population
CR = contact rate; the amount of contaminated
medium contacted per unit time or event (e.g.,
liters/day)
EFD = exposure frequency and duration; describes how
long and how often exposure occurs. Often
calculated using two terms (EF and ED):
EF = exposure frequency (days/year)
ED = exposure duration (years)
BW = body weight; the average body weight over the
exposure period (kg)
Assessment-determined variable
AT * averaging time; period over which exposure is
averaged (days)
Source: RAGS (EPA 1989a).
21402-007
16
-------
Uncertainty analysis and toxicity assessment. The
toxici ty assessment is another contributor to uncertainty
in risk assessment. Limitations in the analytical data
from environmental samples affect the results of the
toxicity assessment, but not to the extent that they affect
other components of the risk assessment process. Data
on physical and chemical parameters that may influence
bioavailability can influence route-to-route and vehicle-
related adjustments to toxicity values. The selection of
appropriate toxicity values is influenced by monitoring
data from environmental samples to the extent that this
information assists in identifying chemicals of potential
concern, exposure pathways, and the time periods over
which exposure may occur. Based on this information,
the lexicologist identifies sub-chronic or chronic RfDs,
RfCs, and cancer slope factors for oral, dermal, and
inhalation exposure pathways.
A list of toxicity values for risk assessment should
include an indicationof the degree of certainty associated
with these values. Weight-of-evidence classifications
provide a qualitative estimate of certainty and should be
included in the discussion of cancer slope factors.
Uncertainty and modifying factors used in deriving
RfDs and RfCs should also be included in the discussion
of non-carcinogenic effects.
2.1.4 Risk Characterization
Overview of methods for risk characterization. The
last step in the baseline risk assessment is risk
characterization. This is the process of integrating the
results of the exposure and toxicity assessments, by
comparing estimates of intake with appropriate
lexicological values to determine the likelihood of
adverse effects in potentially exposed populations. Risk
characterization is considered separately for
carcinogenic and non-carcinogenic effects, because
organisms typically respond differently following
exposure to carcinogenic and non-carcinogenic agents.
For non-carcinogenic effects, lexicologists recognize
the existence of a threshold of exposure below which
there is likely to be no appreciable risk of adverse health
impacts in an exposed individual. It is the current EPA
position that exposure to any level of carcinogenic
compounds is considered to carry a risk of adverse
effect, and that exposure is not characterized by the
existence of a threshold.
EPA's procedure for calculating risk from exposure to
carcinogenic compounds (EPA 1986a, EPA 1989a,
EPA 1989b)usesanon-thresbold, dose-response model.
The model is used to calculate a cancer slope factor
(mathematically, the slope of the dose-response curve)
for each chemical. Generally, the cancer slope factor is
used in conjunction with the chronic daily intake to
derive a probabilistic upperbound estimate of excess
lifetime cancer risk to the individual.
17
The dose-response model most commonly used by EPA
in deriving the cancer slope estimates is linearized and
multistage. The mathematical relationship of the model
assumes that the dose-response relationship is linear in
ihe low-dose portion of the curve (EPA 1989a). Given
this assumption, the slope factor is a constant, and risk
is directly proportional to intake.
The recommended practice for evaluating the potential
for non-carcinogenic effects is to compare the RfD of a
given chemical to the estimated intake of the potentially
exposed population from a given exposure pathway
(EPA 1989a). This ratio (intake/RfD) is termed the
"hazard quotient." It is not a probabilistic estimate of
risk, but simply a measure of concern, or an indicator of
the potential for adverse effects. A more detailed
discussion of risk characterization is presented in RAGS.
Further discussion of methods for risk characterization,
and of specific factors such as metabolic rate factors,
gender differences, and variable effects due to multiple
chemicals of potential concern, is available from many
sources (EPA 1988a, EPA 1989b, EPA 1989c).
Uncertainty analysis in risk characterization. No
risk assessment is certain. Risk assessment is a process
that provides an estimate of potential (present and
future) individual risk, along with the limitations or
uncertainties associated with the estimates. The most
obvious effect of limitations in the analytical data on
risk characterization is the ability to accurately estimate
the potential for adverse effects in potentially exposed
individuals. Clearly, if theavailablemoniioringdatado
not facilitate ameaningful determination of RME values,
the risk estimates will directly reflect this uncertainty.
«• Uncertainties in toxicological measures
and exposure assessment are often
assumed to be greaterthan uncertainties in
environmental analytical data; thus, they
are assumed to have a more significant
effect on the uncertainty of the risk
assessment.
Resource and time constraints often limit the opportunity
to develop a well-designed and comprehensive data set.
Risk assessments must be conducted using the available
information, even when there is no opportunity to
improve the data set However, the results should be
presented with an explititstatementrcgarding limitations
and uncertainty.
If possible, a sensitivity analysis should be conducted to
bound theiesultsofriskassessments. Asimpleapproach
might consist of establishing the range of potential
values (e.g., minimum, most likely, and maximum) for
key input variables and discussing the influence on the
resulting risk estimates. The key variables can then be
ranked with respect to the magnitude of potential effect
on the risk estimates. In certain instances, more
-------
quantitative approaches to uncertainty analysis may be
useful if they can be supported by the available
information. Combining probability distributions using
Monte Carlo techniques is one commonly cited example
(EPA 1988b, EPA 1989a, Fmkel 1990). An overview
of recommended methods for assessment of uncertainty
in risk characterization is presented in RAGS.
Risk* Assistant, a software tool developed for EPA,
provides an uncertainty analysis that determines the
effect on the final risk estimate of using alternative
parameter values, indicates the relative contribution of
each pathway to risks from the contaminated media, and
(for carcinogenic risks) determines the percentage of
total risk from a contaminant in each medium (Thistle
Publishing 1991). A more detailed consideration of
uncertainty analysis in risk assessment may be found in
Methodology for Characterization of Uncertainty in
Exposure Assessment (EPA 1985) and Confronting
Uncertainty inRiskManagement: A Guide for Decision-
Makers (Fiakel 1990).
2.2 ROLES AND RESPONSIBILITIES
OF KEY RISK ASSESSMENT
PERSONNEL
The risk assessor generally enlists the participation of
individuals with specific skills and technical expertise.
The quality and utility of the baseline risk assessment
will ultimately depend on the planning and interaction
s Key participants include
the RPM and the risk assessor, who are primarily
responsible for ensuring that data collected during the
RI are useable for risk assessment activities. Other
participants include hydrogeologists, chemists,
statisticians, quality assurance staff, and other technical
support personnel involved in planning and conducting
the RI. Exhibit 8 summarizes the roles and
responsibilities of the risk assessment participants.
2.2.1 Project Coordination
All data collection activities that support the risk
assessment are coordinated by the RPM. TheRPM's
responsibilities begin upon site listing and continue
through deletion of the site from the National Priorities
List. A network of technical experts, including
representatives of other agencies involved in human
health or environmental/ecological assessments or
related issues, is established at the start of the RI. This
ensures that the potential for adverse effects to human
health and the environment is adequately assessed during
tbeRI. TosuccessfuUyplananddirectthesampIingand
analysis effort, the RPM must facilitate interaction
among key participants.
2.2.2 Gathering Existing Site Data
and Developing the Conceptual
Model
The RPM is responsible for gathering and evaluating all
historical and existing site data. This is an important
element in planning the scope of the risk assessment and
data collection, and in determining additional data needs.
Sources of information especially pertinent for risk
assessment include data from potentially responsible
parties, industrial records identifying chemicals used in
processes, preliminary natural resource studies, Agency
for Toxic Substances and Disease Registry (ATSDR)
health studies, environmental impact statements,
transport manifests, site records, site inspection
documents, and site visits. Aerial photographs and site
maps showing past and present locations of structures
and transportation corridors should also be collected.
The RPM should also consider the application of a
computer-based Geographical Information System
(GIS) as a major tool.
The RPM should ensure that a broad spectrum analysis
was conducted at the site for all media and should
review industry-specific records to minimize the
potential for false negatives. From the inspection of
historical data and broad spectrum analyses, a
preliminary list of the chemicals of potential concern is
prepared to assist in scoping and in developing the
conceptual model of the site. Once all the existing
historical site data have been collected, the RPM works
with the risk assessor to develop a conceptual model.
The conceptual model is a depiction and discussion of
the current understanding of the contamination, the
sources of release to the environment, transport
pathways, exposure pathways, exposure areas and
receptors at risk. Preliminary identification of potential
exposure pathways at the site under investigation is
particularly important for the design of a thorough data
collection effort The conceptual site model should be
provided to all key participants in the RI during the
project scoping and should be included in the workplan.
As work progresses and the site is better characterized,
the RPM and the risk assessor should update the
conceptual model.
2.2.3 Project Scoping
The adequacy of the sampling and analysis effort
determinesthequalityoftheriskassessment. Therefore,
it is imperative that the risk assessor be an active
member of RI planning and continue to be involved
during the entire course of the project
18
-------
EXHIBIT 8. ROLES AND RESPONSIBILITIES OF
RISK ASSESSMENT TEAM MEMBERS
Remedial project manager
• Directs, coordinates and monitors all activities.
• Establishes network with other data users including federal, state and local agencies.
• Creates conceptual model.
• Gathers existing site data.
• Organizes scoping meetings.
• Controls budget and schedule.
• Guides preparation of QA documents.
• Ensures that the risk assessor receives preliminary analytical data.
• Contributes to data assessment.
• Develops preliminary list of chemicals of potential concern.
• Resolves problems affecting Rl objectives, including risk assessment issues (e.g., resampling,
reanalysis).
Risk ••••••or
• Reviews all relevant existing site data.
• Assists the RPM in developing the conceptual model and the preliminary list of chemicals of potential
concern.
• Contributes to recommendations on sampling design, analytical requirements, including chemicals of
potential concern, detection limits and quality control needs during project scoping.
• Helps to refine the conceptual model.
• Communicates frequently with (he RPM, hydrogeologist and chemist to ensure that data collection
meets needs.
• Reviews and contributes to SAP and QA documents.
• Assesses preliminary data as soon as available to verify conceptual site model.
• Specifies additional needs.
• Assesses reviewed data for useability in risk assessment.
• Communicates all site activities with specific groups, such as chemists.
• Prepares risk assessment
Hydrogeotoglat, chemiat and other technical support
Provides technical input to scoping.
Prepares/provides input to SAP and QA documents in support of risk assessment data needs.
Communicates frequently with the RPM and/or risk assessor on status of data collection and issues
affecting data.
Provides preliminary data to the RPM and/or risk assessor for review.
Supports fate and transport modeling for the exposure assessment.
Implements corrective actions to improve data useability.
Quality aMurance specialist
• Responsible for data quality review and technical assistance in preparing QA documents.
• Provides historical performance QA data or recommendations for appropriate QC.
• Ensures adequate QA procedures are in place, including field and analytical audits.
21-002-006
19
-------
*• Analytical data collected solely for other
purposes may not be of optimal use to the
risk assessment.
Data obtained solely with the aim of characterizing the
nature and extent of contamination at a site may not
fully support the needs of the risk assessor in quanti taring
exposure, and therefore the potential for adverse effects
in human and nonhuman receptors. Data on the nature
and extent of contamination may therefore be rejected
by the risk assessor, requiring an additional round of
sampling. For example, data identifying the boundaries
of the site may not be representative of the level of
contamination within an exposure area. Therefore, it is
important to maintain the risk assessment data
requirements as a high priority throughout remedial
investigations.
Sampling and analysis methods discussedduring scoping
should ultimately be based on site-specific data needs.
The RPM, risk assessor, hydrogeologist, statistician,
and project chemist must maintain open communication
during scoping and throughout the RI to ensure that this
occurs. Data review and deliverable requirements should
be determined during the scoping meetings so that these
specifications can be included in the sampling and
analysis plan (SAP) for the RI. The RPM should
prepare a checklist of considerations for the scoping
meetings and provide it to all individuals involved.
Exhibit 9 presents an example checklist of items useful
for risk assessment to be considered by the RPM during
scoping. Chapters 3 and 4 give specific guidance for
planning the data collection efforts to support risk
assessments.
2.2.4 Quality Assurance Document
Preparation and Review
After scoping, the RPM guides the preparation of the
workplan and quality assurance documents. The
workplan, the SAP, and the quality assurance project
plan (QAPjP) should document the combined decisions
of the RPM, risk assessor, and other project staff.
EXHIBIT 9. EXAMPLE RISK ASSESSMENT
CHECKLIST FOR USE IN SCOPING
• Has all historical information been gathered and characterized
and is it appropriate and available for use?
• What sample matrices should be investigated?
• What analytical methods should be used?
• Are the methods appropriate for risk assessment, given
specific contaminants present and their toxicity?
• Will any special quality control requirements be necessary?
• Who will conduct the analysis (e.g., which type of laboratory)?
• What analytical data sources should be used (fixed laboratory
and/or field analysis)?
• What sampling designs are appropriate?
• How many samples will be needed?
• How will the data review be accomplished?
• What types of deliverables will be required? Specify the types of
deliverables required from both laboratory and data validation.
• What budget or other limitations constrain data collection (e.g.,
due date, contractor availability)?
20
-------
Particular emphasis is placed on establishing confidence
limits, acceptable error, and level of quality control
(discussed in Chapter 3). This facilitates cost-effecti ve
design of the sampling and analytical program and
minimizes the collection of data of limited use for risk
assessment.
The risk assessor reviews the workplan and SAP to
ensure that the relevant data quality issues, sampling
design, analytical needs, and data assessment procedures
are adequately addressed for risk assessment. Exhibits
10 and 11 provide checklists to aid the review of the
workplan and SAP.
2.2.5 Budgeting and Scheduling
As the overall site manager, the RPM must address and
balance risk assessment data needs with other data use
needs, such as health and safety, treatability studies,
transport, and the nature and extent of contamination.
The risk assessor is responsible for identifying specific
data requirements for risk assessment and
communicating these needs to the RPM. The RPM is
responsible for developing and implementing the
schedule for acquiring the data. Balancing costs and
services while adhering to the schedule is a major
responsibility of the RPM.
The RPM must coordinate the use of analytical services.
Data from different analytical sources provide the
flexihilily needed to balance cost with smnpling needs
and time constraints. The advantages anddisadvantages
of field analyses and fixed laboratory analyses should
be considered, as described in Chapters 3 and 4. The
risk assessment participants can assist in the development
of field sampling plans and the selection of appropriate
analytical methods that will provide the risk assessor
with a set of useable data, within the budgeting and
scheduling constraints of the RPM.
2.2.6 Iterative Communication
Continuing, open, and frequent communication among
the participants is critical to the success of the RI and
baseline risk assessment A single meeting or discussion
is rarely adequate to ensure that all relevant issues have
been addressed. Development of the risk assessment
within the RI report is an iterative process of action,
feedback, and correction or adjustment.
After review of the workplan, the SAP, and the QAPjP,
the RPM monitors the flow of information. The risk
assessor assists the RPM toensure that the data produced
are in compliance with the requirements of the workplan
and SAP. Key questions they consider once the data
become available are:
• Have correct sampling protocols been followed?
• Have all critical samples been collected?
EXHIBIT 10. CHECKLIST FOR REVIEWING THE WORKPLAN
Does the workplan address the objectives of baseline risk assessment?
Does the workplan document the current understanding of site history and the physical setting?
Have historical data been gathered and assessed?
Has information on probable background concentrations been obtained?
Does the workplan provide a conceptual she model for the baseline risk assessment, including a
summary of the nature and extent of contamination, exposure pathways of potential
concern, and a preliminary assessment of potential risks to human health and the environment?
Does the workplan document the decisions and evaluations made during project scoping,
including specific sampling and analysis requirements for n'sk assessment?
Does the workplan address all data requirements for the baseline risk assessment and explicitly
describe the sampling, analysis and data review tasks?
21-002-010
21
-------
EXHIBIT 11. CHECKLIST FOR REVIEWING THE SAMPLINC3
AND ANALYSIS PLAN
• Do the objectives of the QAPjP and the field sampling plan meet risk assessment needs
established in the scoping meeting?
• Are QA/QC procedures provided for in the SAP adequate for the purposes of the baseline
risk assessment?
• Have the data gaps for risk assessment that were identified in the Rl workplan been
adequately addressed in the SAP?
• Are there sufficient QC samples to measure the likelihood of false negatives and false
positives, and to determine the precision and accuracy of resulting data?
• Have analytical methods been selected that have detection limits adequate to quantitate
contaminants at the concentration of concern?
• Have SOPs been prepared for sampling, analysis and data review?
• Will the sampling and analysis program result in the data needed for the baseline risk
assessment:
- to address each medium, exposure pathway and chemical of potential concern,
- to evaluate background concentrations,
- to provide detail on sample locations, sampling frequency, statistical design and analysis,
- to evaluate temporal as well as spatial variation, and
- to support evaluation of current as well as future resource uses?
• Have the samples been analyzed as requested?
• Are data arriving in a timely fashion?
• Have approbate samplequantitationliinits/detec-
tion limits been achieved?
• Has quality assurance been addressed as stated in
tbeSAPandQAPjP?
• Have the data been reviewed as stated in the SAP?
• Is the quality of the analytical data acceptable for
their intended use?
Based upon these considerations, the RPM, risk assessor
if any corrective actions are needed, such as requesting
additional sampling, using alternative analytical
methods, or reanalyzing samples.
2.2.7 Data Assessment
The RPM and risk assessor work with other participants
to identify a list of chemicals of potential concern and
21-002-011
decide on data review procedures. This information is
developed during project scoping and incorporated into
the workplan and SAP. The RPM, risk assessor, and
project chemist should agree on the type and level of
data review required for both positive and "non-detect"
results. Typically, the RPM assesses the overall data
reviewed by the chemist, and the risk assessor reviews
data relevant to risk assessment, unless other
arrangements have been established and explicitly stated
in the SAP.
The risk assessor may request preliminary data, or
results that have received only a partial review, in order
to expedite the risk assessment to save time and resources.
Preliminary data can be used to validate the conceptual
model or to begin the toxicity assessment Thedatamay
also indicateaneedfor modifying sampling or analytical
procedures. However, preliminary data should not be
used in calculating risk. Once the full analytical data set
is obtained, the RPM and risk assessor should consult
with the project chemist and statistician to assess the
utility of all available information.
22
-------
2.2.8 Assessment and Presentation
of Environmental Analytical
Data
Once environmental data are evaluated in the data
review process, the risk assessor develops a final data
set for use in the baseline riskassessment All chemicals
of potential concern should now be identified. The risk
assessor prepares summary tables containing the
following information:
• Site name and sample locations,
• Number of samples per defined, representative
area of each medium (e.g., do not count background
samples together with other samples),
• Sample-specific results,
• Analyte-specific sample quantitation limits,
• Number of values above the quantitation limit,
• Measures of central tendency (e.g., 95% UCL on
the arithmetic mean of the environmental
concentration),
• Specifications for the treatment of detection or
quantitation limits and treatment of qualified data,
and
• Ranges of concentrations.
All assumptions, qualifications, and limitations should
be explicitly stated in the tables. The risk assessor
provides the final data summary tables to the RPM,
project hydrogeologist, project chemist, and other
appropriate project staff for review. These are the data
that will be used in the baseline risk assessment to
determine the potential risk to human health. It is
essential, therefore, that this information consists of the
best data available and reflects the collective review of
the key participants in the risk assessment An example
of such a set of data is given in Appendix I.
23
-------
Chapter 3
Useability Criteria for Baseline Risk Assessments
This chapter applies data useability criteria to data
collection planning efforts to maximize the useability of
environmental analytical data in baseline risk
assessments. It also addresses preliminary issues in
planning sampling and analysis programs.
The chapter has two sections. Section 3.1 discusses data
useability criteria involved in risk assessment and
suggests ways they can be applied to ensure data are
useable. Section 3.2 presents preliminary sampling and
analysis issues including identification of chemicals of
potential concern, available sampling and analytical
strategies or methods, and probable sources of
uncertainty.
Before scoping the RI, it is critical for successful planning
that the RPM develop a conceptual site model (Exhibit
6) in consultation with the risk assessor and all
appropriate personnel. This chapter provides the
background information necessary to plan for the
acquisition of environmental data for baseline risk
assessments. The quality of a risk assessment is
intimately tied to the adequacy of the sampling and
analysis plan (SAP) developed during the RI.
«• Effective planning improves the useability
of environmental analytical data in the final
risk assessment.
Data needs for baseline risk assessments are not
necessarily met by data the RPM acquires to identify the
nature and extent of contamination at a Superfund site.
For example, a sampling strategy designed to determine
the boundaries of a contaminated area may not provide
data to quantitate concentrations within an exposure
area. The risk assessment may also require more
precision and accuracy, and lower detection limits.
Accordingly, the risk assessor should be an active
member of the team planning the RI and must be
consulted from the start of the planning process.
Four fundamental decisions for risk assessment are to
be made with the data acquired during the RI, as
discussed in Chapter 2.
• If the sampling design is representative, the
question of what contamination is present and at
what concentration is an analytical problem. Key
concerns are the probability of false negatives and
false positives. Theselectionofanalyticalmethods,
laboratory performance, and type and amount of
data review affects these issues for both site and
background samples.
• Assuming that chemicals of potential concern
have been identified, the secondquestion involves
background levels of contamination. Are site
concentrations sufficiently elevated from true
background levels to indicate an increased risk for
human health due to site contamination?
. • All exposure pathways and exposure areas must
be identified and examined. The two decisions
concerning exposure pathways and areas primarily
involve identifying and sampling the media of
concern.
• The final decision involves characterizing exposure
areas. Sampling and analysis must be
representative and satisfy performance objectives
determined during the planning process.
RI planning and implementation of RI plans affect the
certainty of chemical identification and quantitation.
Therefore, the RI needs to collect useable environmental
analytical data to enable the risk assessor to make these
decisions.
Acronyms
AA atomic absorption
CLP Contract Laboratory Program
CRDL contract required detection limit
CRQL contract required quantitation limit
DQI data quality indicator
DQO data quality objective
GC gas chromatography
HRS Hazard Ranking System
ICP inductively coupled plasma
IDL instrument detection limit
LOL limit of linearity
LOQ limit of quantitation
MDL method detection limit
MS mass spectrometry
OVA organic vapor analyzer
PA/SI primary assessment/site inspection
PAH polycyclic aromatic hydrocarbon
PCS polychlorinated bipbenyl
PQL practical quantitation limit
QA quality assurance
QC quality control
QAPjP quality assurance project plan
QTM Quick Turnaround Method
RI remedial investigation
RI/FS remedial investigation/feasibility study
RPM remedial project manager
RRF relative response factor
RRT relative retention time
SAP sampling and analysis plan
SOP standard operating procedure
SQL sample quantitation limit
TIC tentatively identified compound
TRIS Toxic Release Inventory System
XRF X-ray fluorescence
25
-------
3.1 DATA USEABILITY CRITERIA
Exhibit 12 lists the six data useability criteria involved
in planning for the risk assessment, summarizes the
importance of each criterion to risk assessment, and
suggests actions to take during the planning process to
improve the useability of data. The following sections
define each criterion and describe its effect on risk
assessment.
3.1.1 Data Sources
The data sources selected during the RI planning process
depend on the type of data required and their intended
use. Data collected prior to the RI are considered
historical; data collected during the RI are considered
current and are usually specified in the RI planning
process. Data may be analytical or non-analytical. The
same analytical data requirements apply, whether the
data are current or historical. Field screening methods
can be used, and sufficient documentation produced, to
actas an initial source of data. The minimum criteriafor
analytical data are discussed in Chapter 5.
Exhibit 13 identifies available data sources and their
primary uses in the risk assessment process. Historical
and current analytical data sources are briefly discussed
below.
Data sources prior to remedial investigation.
Historical data sources are useful for determining
sampling locations and analytical approaches in the RI.
Early site inspections may locate industrial process
information that suggests chemicals of potential concern.
Historical data indicate industry-specific analytes and
general levels of contamination and trends that are
useful for identifying exposure pathways, for developing
the sampling design, and for selecting analytical methods.
Historical analytical data are often available from the
preliminary assessment/site inspection (PA/SI),
including reports on the physical testing, screening, and
analysis of samples. Other sources of analytical data for
baseline risk assessment include the Hazard Ranking
System (HRS) documentation, site records on removal
and disposal, and industry-specific .systems for chemical
discharge permits. Results from analyses by state or
local governments may also indicate chemicals of
potential concern. Exact locational data for historical
samples should be obtained whenever possible.
«•• Use historical analytical data and a broad
spectrum analysis to initially identify the
chemicals of potential concern or exposure
areas.
The quality of historical data must be determined prior
to their use in the RI. For historical analytical data to be
EXHIBIT 12. IMPORTANCE OF DATA USEABILITY CRITERIA
IN PLANNING FOR BASELINE RISK ASSESSMENT
Data
UMability
Criterion
Importance
Suggested Action
Data Sources
(3.1.1)
Data sources must be comparable if data are combined tor
quantitative use in risk assessment. Plans can be made in
the RI for use of appropriate data sources so that data
compatibility does not become an issue.
Use data from different data sources together to
balance turnaround lime, quality of data, and
cost Consult with a chemist or statistician to
assess compatibility of data sets.
Documentation
(3.1.2)
Deviations from the SAP and SOPs must be documented
so that the risk assessor will be aware of potential
limitations in the data. The risk assessor may need
additional documentation, such as field records on weather
conditions, physical parameters and site-specific geology.
Data useabte for risk assessment must be linked ID a
specific location.
Review the workplan and SAP and, If
appropriate, SOPs. As the data arrive, check
for adherence to the SAP so (hat corrective
action such as resampling may be taken and still
adhere to the project timetable
Stress importance of chain-of-custody for
sample point identification in RI planning
meetings.
Analytical
Methods and
Detection
Limits
(3.1.3)
The method chosen must test for the chemical of potential
concern at a detection limit that will meet the concentration
levels of concern in applicable maticas. Samples may
have to be refinar/zed at a lower detection Nmtt if tie
detection limit is not tow enough to confirm (he presence
and amount of contamination.
Participate with chemist in selecting methods
will appropriate detection Bmfe during RI
planning. Consultation with a chemist is
required when a metxxfs detection limit is at or
above the concentration level of concern.
21-002-012
26
-------
EXHIBIT 12. IMPORTANCE OF DATA USEABILITY CRITERIA
IN PLANNING FOR BASELINE RISK ASSESSMENT
(Cont'd)
Data
Usability
Criterion
Importance
Suggested Action
Data Quality
Indicators
(3.1.4)
Completeness
Comparability
Representa-
tiveness
Precision
Accuracy
Completeness for critical samples must be 100%.
Unforeseen problems during sample collection (as defined
in Chapter 4) and analysis can affect data completeness.
If a sample data set tor risk assessment is not complete,
more samples may have to be analyzed, affecting Rl time
and resource constraints.
The risk levels generated in quantitative risk assessment
may be questionable if incompatible data sets are used
together.
Sample data must accurately reflect the site
characteristics to effectively represent the site's risk to
human health and the environment. Hot spots and
exposure area media must have representative data.
If the reported result is mar the concentration of concern,
it is necessary to be as precise as possible in order to
quantify the likelihood of false negatives and false
positives.
Quantitative accuracy information is critical when results
are reported near the level of concern. Contamination in
the field, during shipping, or in the laboratory may bias the
analytical results. Instruments that are not calbrated or
tuned according to Statement of Work requirements may
also bias results. The use of data that is biased may affect
the interpretation of risk levels.
Define completeness in the SAP for both the
number of samples and quantity of useabte data
needed to meet performance objectives.
Identify critical samples during scoping. The
SAP should be reviewed by the RPM before
initiation of sampling.
Plan to use comparable methods, sufficient
quality control, and common units of measure for
different data sets that will be used together, to
facilitate data compatability. Consult with a
chemist to ensure compatibility of data sets.
Discuss plans for collection of sufficient number
of samples, a sample design that accounts for
exposure area media, and an adequate number
of samples for risk assessment during scoping
and document plans in the SAP. This guidance
may be modified by Region-specific guidelines.
Plan for the use of QC samples (duplicates,
replicates and/or collocated samples) applicable
to risk assessment before sampling activities
begin. Assess confidence limits from the QC
data on the basis of the sampling design or
analytical method used.
Plan and assess QC data (blanks, spikes,
performance evaluation samples) to measure
bias in sampling and analysis. Consult a
chemist to interpret data qualified as
"estimated* that are near a concentration of
concern.
Data Review
(3.1.5)
Use of preliminary data or partially reviewed data can
conserve time and resources by allowing modification of
the sampling plan while the Rl is in process. Critical
analytes and samples used for quantitative risk
assessment require a full data review.
Decisions regardmg level and depth of review wll
conserve time and project resources and should
be made in conjunction with the RPM and
analytical chemist. "Non-detecT results require
a fun review.
Reports
to Risk
Assessor
(3.1.6)
Data reviewers should report data in a format that provides
readability as well as clarifying information. SQLs, a
narrative, and quaifiers that are fuly explained reduce the
time and effort required in interpreting and using the
analytical results. Limitations can be readily identified and
documented in the risk assessment report
Prescribe a report format during scoping, and
include it in the SAP. Communicate with the
potential data reviewer to aid the definition of a
specific report format. Region-specific
guidelines may apply.
27
-------
EXHIBIT 13. DATA SOURCES AND THEIR
USE IN RISK ASSESSMENT
Available Data
Sources
Data Type
Primary Use(s)
PA/SI data
Analytical
' Scoping and planning
1 Identifying data trends
1 Determining historical background levels
MRS
documentation
Site records,
manifests,
PA/SI,
analytical
> Quantitating the risk assessment
> Identifying trends
1 Planning (by identifying the chemicals present)
Site records on
removal and disposal
Administrative
Planning (by identifying the chemicals present)
Toxic Release
Inventory System
(TRIS) (Industry-
Specific)
Chemical
discharge
Planning (by identifying the chemicals present)
Site, source and
media characteristics
as found in PA/SI data
and reference
materials
Physical
parameters
(e.g., meteor-
ological,
geological)
1 Determining fate and transport
' Defining exposure pathways
Field screening
Analytical
> Performing a preliminary assessment
> Characterizing the site
Field analytical
Analytical
1 Quantitating the risk assessment
1 Characterizing the site
Fixed laboratory,* both
CLP and non-CLP
(EPA, state, PRP,
commercial)
Analytical
' Quantitating the risk assessment
' Providing a reference
' Broad screen
1 Confirming screening data
1 Characterizing a site
Mobile laboratories often have the same instrumentation available as fixed laboratories,
with the exception of ICP or MS.
21402-013
useful in the quantitative risk assessment, sampling
design, sampling and analytical techniques, and detection
limits must be documented, and the data must nave been
reviewed.
Historical analytical data of unknown quality may be
used in developing the conceptual model or as a basis
for scoping, but not in determining representative
exposure concentrations. Analytical data from the PA/
SI thatmeetminimum data useability requirements (see
Section 5.1.1) can be combined with data from the RI to
estimate exposure concentrations. Similarly, historical
data of lower quality may be used if the concentrations
are confirmed by subsequent RI analyses.
Data sources for the remedial investigation. It may
be efficient to use a variety of data sources during an RI.
For example, analytical services providing a rapid
turnaround of estimated data can be used to estimate the
three-dimensional extent of contamination or to "chase"
a groundwater pollutant plume. Rapid turnaround
analytical services include field analysis or Quick
28
-------
Turnaround Method (QTM) analyses under the Contract
Laboratory Program (CLP). On the other hand, if an
unexpected situation arises, such as the discovery of
buried drums on the site, it may be appropriate to
procure the analytical services of a local commercial
laboratory. Data requiring a rapid turnaround are
typically produced from streamlined analytical methods,
and a certain percentage should be analyzed using a
confirmatory method, such as CLP analytical services.
The planning process for the RI identifies gaps in the
available analytical data and determines additional data
collection requirements. Three types of analytical data
sources can be used during the RI to acquire analytical
data forarisk assessment. These include fieldscreening,
field analyses, and fixed laboratory analyses.
• Field screens are performed using chemical field
test kits, ion-specific probes, and other monitoring
equipment, but should be confirmed by other
techniques. Field screening is usually performed
to provide a preliminary assessment of the type
and level of concentration of the chemicals of
potential concern.
• Field analyses are performed using instruments
and procedures equivalent to fixed laboratory
analyses; they produce legally defensible data if
QC procedures are implemented. Field analyses
are usually performed as pan of an integrated
sampling and analysis plan to quantitate risk
assessment and site characterization.
• Fixed laboratory analyses are particularly useful
for broad spectrum and confirmation analyses.
They often provide more detailed information
over a wider range of analytes than field analyses.
Fixed laboratory analyses are critical toquantitative
risk assessment and site characterization.
A discussion of issues related to fieldand fixed laboratory
analyses is presented in Section 32.9.
Analytical services constitute a significant portion of
the Superfund budget and should be conserved when
possible. CLP costs do not appear on the remedial
investigation/feasibility study (RI/FS) project budget.
Analyte-specific methods may be used for chemicals
identified after a broad spectrum analysis by CLP or
other fixed laboratory analysis, and may provide more
accurate results. Site samples analyzed by CLP routine
analytical services take an average of 35 days to produce
results and data review will add to the overall turnaround
time. Otherdatasources.suchasamobilelaboratoryor
CLP QTM or special analytical services, can quickly
produce good "first look" results which can be followed
up immediately while on site. Mobile laboratory services
can replace some CLP services if analytical capabilities
are adequately demonstrated by method validation data
and if minimum QC requirements aremet (see p. 59). At
least 10% of sample analyses should be confirmed by
fixed laboratory analysis in all situations.
3.1.2 Documentation
Data collection and analysis procedures must be
accurately documented to substantiate the analysis of
the sample, conclusions derived from the data, and the
reliability of the reported analytical data. Plans should
be prepared during the RI scoping to document data
collection activities. This RI documentation can be
used later to evaluate completeness, comparability,
representativeness, precision, and accuracy of the
analytical data sets. Four major types of documentation
are produced during an RI:
• Thesamplingandanalysisplan,includingaquality
assurance project plan (QAPjP),
• Standard operating procedures (SOPs),
• Field and analytical records, and
• Chain-of-custody records.
Sampling and analysis plan. The scoping meetings
and the SAP must clearly establish the end use
requirements for data The data quality indicators for
assessing results against stated performance objectives
should also be documented in the SAP (see Section
3.1.4). The SAP includes the QAPjP and information
required in toe SOPs, field and analytical records, and
chain-of-custody records (EPA 1989a).
Standard operating procedures and field and
analytical records. SOPs for field and analytical
methods must be written for all field and laboratory
processes. Adherence to SOPs provides consistency in
samplingandanalysisand reduces the level of systematic
errorassociated with data collection andanalysis. Exhibit
14 lists the types of SOPs, field records, and analytical
records that are usuallyassociated with RI data collection
and analyses, and relates the importance of each to the
risk assessment
All deviations from the referenced SOPs should be pre-
approved by the RPM and documented. Samples that
are not collected or analyzed in accordance with
established SOPs may be of limited use because their
quality cannot be determined.
Chain-of-custody. The technical team must decide
during scoping what data may be used for cost recovery
actions, and plan accordingly for the use of full-scale
chain-of-custody or less formal chain-of-custody
procedures. Full-scale chain-of-custody is required for
29
-------
EXHIBIT 14. RELATIVE IMPORTANCE OF
DOCUMENTATION IN PLANNING AND ASSESSMENT
Documentation
Importance
Sampling and Analysis Plan
Selection and identification of sampling points
Sample collection SOP
Analytical procedures or protocols
SOP for data reporting and review
QA project plan
Method-specific QC procedures
QA/QC procedures
Documented procedures for corrective action
SOP for corrective action and maintenance
Sample preservation and shipping SOP
SOPs for sample receipt, custody, tracking and storage
SOP for installation and monitoring of equipment
Critical
High
High
High
High
Medium
Medium
Medium
Medium
Medium
Low
Low
Chain-of-Custody
Documentation records linking data to sample location
Sampling date
Sample tags
Custody seals
Laboratory receipt and tracking
Critical
Critical
High
Low
Low
Reid and Analytical Records
• Reid log records
• Reid information describing weather conditions, physical parameters
or site-specific geology
• Documentation for deviations from SAP and SOPs
* Data from analysis - raw data such as instrument output spectra,
chromatograms and laboratory narrative
• Internal laboratory records
High
High
High
High
Low
KEY Critical - Essential to the useability of data for risk assessment
High • Should be addressed in planning for risk assessment
Medium • Primarily impacts how data are qualified in risk assessment
Low - Usualry has little effect on useaMity of data for risk assessment
21-002-014
cost recovery and enforcement actions, but does not
affect a quantitative determination of risk. Full-scale
chain-of-custody includes sample labels and formal
documentation that prove the sample was not tampered
with or lost in the data collection and analysis process.
Sample identity must be verifiable from the collector's
notebook and laboratory data sheets, as well as from a
formal chain-of-custody.
3.1.3 Analytical Methods and
Detection Limits
The choice of analytical methods is important in RI
planning. Appropriateanalyticalmethodshave detection
limits that meet risk assessment requirements for
chemicals of potential concern and have sufficient QC
measures to quantitate target compound identification
and measurement. The detection limit of the method
directly affects the useability of data because chemicals
reported near the detection limit haveagreaterpossibility
of false negatives and false positives. The risk assessor
or RPM must consultachemistfor assistance in choosing
an analytical method when those available have detection
umitsneartherequiredactionlevel. Wheneverpossible,
methods should not be used if the detection limits ate
above the relevant concentrations of concern.
30
-------
3.1.4 Data Quality Indicators
Data quality indicators (DQIs) are identified during the
development of data quality objectives (DQOs), to
provide quantitative measures of the achievement of
quality objectives. This section discusses each of five
DQIs as they relate to the assessment of sampling and
analysis.
• Completeness
• Comparability
• Representativeness
• Precision
• Accuracy
These indicators are evaluated through the review of
sampling and analytical data and accompanying
documentation. The risk assessor may need to
communicate with a chemist or statistician after the data
collection process has been completed to evaluate DQIs.
Therefore, the SAP, field and analytical records, and
SOPs should be accessible. Exhibits 15 and 16
summarize the importance of DQIs to sampling and
analysis in risk assessmentand suggest planning actions.
Each DQI is defined in this section. Note that the
specific use of the indicators to measure data useability
is different for sampling and analysis. For example,
completeness as applied to sampling refers to the number
of samples to be collected. Completeness as applied to
analytical performance primarily refers to the number
of data points that indicate an analytical result for each
chemical of interest (e.g., 10 samples analyzed for 25
chemicals will produce a total of 250 data points, 10
data points for each chemical).
EXHIBIT 15. RELEVANCE OF SAMPUNG DATA QUALITY INDICATORS
Data Quality
Indicator*
Importance
Suggwtad Planning Action
Completeness
Complete materials enable assessment
of sample representativeness for
identification of false negatives and
estimation of average concentration.
Stipulate SOPs for sample
collection and handling in
the SAP to specify requirements for
completeness.
Comparability
Comparable data give the ability to
combine analytical results across
sampling episodes and time periods.
Use the same sample design across
sampling episodes and similar time
periods.
Representativeness
Representative data avoid false negatives
and false positives (field sampling
contamination).
Non-representative data may result in
bias of concentration estimates.
Use an unbiased sample design.
Collect additional samples as
required.
Prepare detailed SOPs for handling
field equipment
Precision
Variability in concentration estimates may
increase uncertainty.
Increase number of samples.
Use appropriate sample designs.
Use QC results for monitoring.
Accuracy
Contamination during sampling process,
loss of sample from improper collection or
handling (loss of volatile*) may result in
bias, false negatives, or false positives
and inaccurate estimates of
concentration.
Use SOPs for sample collection,
handling, and decontamination.
Use QC results for monitoring.
31
-------
EXHIBIT 16. RELEVANCE OF ANALYTICAL DATA
QUALITY INDICATORS
Data Quality
Indicators
Completeness
Comparability
Importance
Poor data quality or lost samples
reduces the size of the data set
and decreases confidence in
supporting information.
Comparable data allow the ability
to combine analytical results
acquired from various sources
using different methods for
samples taken over the period of
investigation.
Suggested Planning Action
Prepare SOPs to support sample
tracking and analytical procedures,
review, and reporting aspects
of laboratory operations.
Reference anatyto-spedfic method
performance characteristics.
Reference applicable fate and transport
documentation.
Anticipate field and laboratory
variability.
Representativeness
Non-representative data or
non-homogeneity of sample
increases the potential for false
negatives or false positives.
Potential for change in sample
before analysis may decrease
representativeness.
Include requirement for broad spectrum
analyses across site area.
Ensure sample is mixed and adequately
represents the environment (not
applicable to volatiles).
Include provision for blank (transport,
storage and analytical) QC monitoring.
Use field methods when applicable,
since they have an advantage in
minimizing variability from transport and
storage.
Precision
Monitoring can indicate the level
off precision.
Precision provides the level of
confidence to distinguish
between site and background
levels of contamination. It is of
primary importance when the
conuenbabon of concern
approaches the detection limit.
Method QC component and site-specific QC
samples that use external reference are the
best monitoring techniques.
Consider in method selection whether
anticipated site levels are near the MDL and |
above action limits.
Accuracy
Accuracy also provides the level
of confidence to distinguish
between site and background
levels of contamination. As
concentration of concern
approaches the detection limit,
the differentiation includes
confidence in determining
presence or absence of chemical
of potential concern.
Broad spectrum screening methods may
have significant negative bias for chemicals
of potential concern. Consider method
accuracy and detection limits if site levels
approach concentrations of concern.
21-002-016
32
-------
Completeness. Completeness is a measure of the
amount of useable data resulting from a data collection
activity. The required level of completeness should be
defined in the QAPjP for the numberof samples required
in the sampling design and for the quantity of useable
data for chemical-specific data points needed to meet
performance objectives. All required data items must
be obtained for critical samples and chemicals, which
are identified in the QAPjP. Incompleteness in any data
item may bias results as well as reduce the amount of
useable data.
Problems that occur during data collection and analysis
affect the completeness of a data set. Fewer samples
may be collected and analyzed than originally planned
because of site access problems. Laboratory performance
may be affected if capacity is exceeded, causing data to
be rejected. Some samples may not be analyzed due to
matrix problems. Samples that are invalid due to
holding time violations may have to be re-collected or
the data set may be determined as useable only to a
limited extent. Therefore, both advance planning in
identifying critical samples and the use of alternative
sampling procedures are necessary to ensure
completeness of a data set for the baseline risk
assessment.
Comparability. Comparability expresses the
confidence with which data are considered to be
equivalent Combined data sets are used regularly to
develop quantitative estimates of risk. The ability to
compare data sets is particularly critical when a set of
data for a specific parameter is applied to a particular
concentration of concern.
Comparability for sampling primarily involves sampling
designs and time periods. Typical questions to consider
in determining sampling comparability include:
• Was the same approach to sampling taken in two
sampling designs?
• Was the sampling performed at the same time of
year and under similar physical conditions in the
individual events?
• Were samples filtered or unfiltered?
• Were samples preserved?
Typical questions to consider in determining analytical
comparability include:
• Were different analytical methodologies used?
• Were detection limits the same or at least similar?
• Were different laboratories used?
• Were the units of measure the same?
• Were sample preparation procedures the same?
Use routine available methods and consistent units of
measure when data collection will span several different
sampling events and laboratories, to increase the
likelihood that analytical results will be comparable.
For field analyses confirmed by laboratory analyses,
careful attention must be taken to ensure that the data
from field and fixed laboratories are comparable or
equivalent (see Section 3.2.9). When precision and
accuracy are known, the data sets can be compared with
confidence. Planning ahead for comparable sampling
designs, methods, quality control, and documentation
will aid the risk assessor in combining data sets for each
exposure pathway.
Representativeness. For risk assessment,
representativeness is the extent to which data define the
true risk to human health and the environment. Samples
must be collected to reflect the site's characteristics and
sample analyses must represent the properties of the
field sample. The homogeneity of the sample, use of
appropriate handling, storage, preservation procedures,
and the detection of any artifacts of laboratory analyses,
such as blank contamination, are particularly important.
For risk assessment, sampling and analyses must
adequately represent each exposure area or the definition
of an exposure boundary.
Representativeness can be maximized by ensuring that
sampling locations are selected properly, potential hot
spots are addressed, and a sufficient number of samples
are collected over a specified time span. The SAP
should describe sampling techniques and the rationale
used to select sampling locations.
Precision. Precision is a quantitative measure of
variability, comparing results for site samples to the
mean, and is usually reported as a coefficient of variation
or a standard deviation of the arithmetic mean. Results
of QC samples are used to calculate the precision of the
analytical or sampling process. Measurement error is a
combination of sample collection and analytical factors.
Field duplicate samples help to clarify the distinction
between uncertainty from sampling techniques and
uncertainty from analytical variability. Analytical
variability can be measured through the analysis of .
laboratory duplicates or through multiple analyses of
performance evaluation samples. If analytical results
arereportednearaconcentration of concern, the standard
deviation or coefficient of variation can be incorporated
in standard statistical evaluations to determine the
confidence level of the reported data. A statistician or
33
-------
a chemist should be consulted to make this deteimination.
Total variability must be evaluated to assess the precision
of data used to define parameters in risk assessment.
Accuracy. Accuracy is a measure of the closeness of a
reported concentration to the true value. This measure
is usually expressed as bias (high or low) and determined
by calculating percent recovery from spiked samples.
The risk assessor should know the required level of
certainty for the end use of the data, expressed as DQOs,
when reviewing accuracy information. When results
are reported at or near a concentration of concern,
accuracy information is critical.
Accuracy of identification may be affected by sample
contamination introduced in the field, during shipping,
or at the laboratory. Held and trip blanks should be used
during the RI to identify contamination and the associated
bias related to sample collection or shipment. Method
blanks, audit samples, and calibration check standards
should be used to monitor laboratory contamination.
Accuracy information may be of less importance if the
precision (bias) is known.
3.1.5 Data Review
This section discusses the importance of alternative
levels of data review to the risk assessment. The two
major effects of data review on data useability are:
• The timeliness of the data review and
• The level and depth of review (e.g., entire site,
specific sample focus, specific analyte focus,
amount of QC data assessed).
A tiered approach involving combinations of data re view
alternatives is recommended so that the risk assessor
can use preliminary data before extensive review. The
RPM, in conjunction with the risk assessor and the
project chemist, must reach a consensus on the level and
depth of data review to be performed for each data
source, to balance useability of data and resource
constraints. Exhibit 17 summarizes the characteristics
and uses of different levels of data review.
Timing of review. Plans for die timing of the data
review should be made prior to data collection and
analysis. The risk assessor uses preliminary data in a
qualitative manner to identify compounds for toxicity
studies and, initially, to ascertain trends in concentrations
and distributions of the analytes of concern, to plan for
additional sampling, and to request additional analyses.
Using data as they become available will usually reduce
the time needed to complete the risk assessment.
However, all data must receive a minimum level of
review before use in the quantitative aspects of risk
assessment. Iterations on data review is resource
intensive; if they are used, they should be planned
carefully as part of a structured process.
EXHIBIT 17. ALTERNATIVE LEVELS OF REVIEW OF ANALYTICAL DATA
Laval of
Reviaw
None
Full
Partial
Automated
Samples
Initial
Initial samples
analyzed for broad
spectrum components
Analyte*
All
All
Critical samples for all analytes
or
All samples for critical analytes
All
All
Parameters
Analytical results
All analytical results,
QC, and raw data
Selected analytical
results, QC, or raw
data
Parameters available
to the automated
system. No raw data
are evaluated.
Potential Uses
Qualitatively identify risk
assessment analytes.
Modify SAP.
Quantitatively perform risk
assessment. Modify SAP.
Modify review process.
Improve timeliness,
overall efficiency,
save resources.
Focus on chemicals
of potential concern.
Improve timeliness,
consistency, cost
effectiveness. If data are
electronically transferred to
a database, eliminates
transcription errors.
21402417
34
-------
«• To expedite the risk assessment,
preliminary data should be provided to the
risk assessor as soon as they are available.
Level and depth of review. The RPM may select
different levels of data review, in consultation with the
risk assessor or other data users and the project chemist.
All data must have a minimum level of review. Data
review levels can range from all site samples with all
reported data to specific key analytes and samples and
may be specified in EPA Regional policies. Careful
consideration is required in selecting a level of review
that is consistent with data quality requirements.
A full data review minimizes false positives, false
negatives, calculation errors, and transcription errors.
"Non-detect" results must be reviewed to avoid "false
negative" conclusions. Partial review should be utilized
only after broad spectrum analysis results have
undergone full review; it may be useful after chemicals
of potential concern have been identified. A flexible
approach to data review alternatives allows the RPM to
balance time and resource constraints.
Depth of data review refers to which evaluation criteria
are selected, ranging from generalized criteria that may
affect an entire data set (e.g., holding time) to analyte-
specific criteria that may affect only a portion of results
from one sample (e.g., recovery of a surrogate spike for
organics or analyte spike recovery for inorganics). The
RPM decides the depth of review for each data source,
to provide a balance between useability of data and
resource constraints. Chemicals of potential concern in
the quantitative risk assessment should not be eliminated
from concern without a full data review.
Automated data review systems. Automated data
review systems can be used to assess all samples and
analytes for which there are computer-readable data in
the format required by the automated system. The depth
of review depends on both the data and the assessment
system. The primary advantages of automated data
review systems for the risk assessor are timeliness, the
elimination of transcription errors that can be introduced
during manual review processes, and computer-readable
output which usually includes results and qualifiers.
This information can be transferred to computer-assisted
riskassessmentandexposuremodeling systems. Exhibit
18 provides a list of software that aid data review and
evaluation.
EXHIBIT 18. AUTOMATED SYSTEMS*
TO SUPPORT DATA REVIEW
System
CADRE
Computer Assisted Data
Review and Evaluation
eOATA
Electronic Data Transfer
and Validation System
EPA Contact
Gary Robertson
Quality Assurance Div.
USEPA, EMSL-LV
(702)798-2215
William CoaMey
USEPA, Emergency
Response Team
(908) 906-6921
Description
An automated evaluation system
that accepts files from CLP format
disk delivery or mainframe transfer
and assesses data based on
National Functional Guidelines for
Organic (or Inorganic) Data Review
(EPA 1991a. EPA 1988e) (default
criteria). System accepts manual
entry of other data sets, and rides for
evaluation can be user-defined to
reflect specific information needs.
(Inorganic system is in development.)
An automated review system
developed to assist in rapid
evaluation of data in emergency
response. May be applicable for both
CLP and non-CLP data. System
combines OQOs, pre-established
site specifications, QC criteria, and
sample collection data with laboratory
results to determine useability.
Both systems operate on an IBM-compatible PC AT with a minimum of 640K RAM. 1
A fixed disk is recommended. 1
35
-------
3.1.6 Reports from Sampling and
Analysis to the Risk Assessor
Preliminary data reports assist the risk assessor in
identifying sampling or analytical problems early enough
so that corrective actions can be taken during data
collection, before sampling or analysis resources are
exhausted. The risk assessor should request preliminary
data during RI planning and formalize the request in the
SAP. The use of such information may reduce the
overall time required for the risk assessment and increase
the quality of a quantitative risk assessment.
Exhibit 19 lists the final data and documentation needed
to support risk assessment, and rates the importance of
each item. Data are most useable when reported in a
readable format and accompanied by additional,
clarifying information. Regional policy usually defines
report structures which specify the format for manual
summaries, for machine-readable data (where required),
and for summary tables from data review. The RPM can
request the data reviewers to provide a data summary
table listing sample results, sample quantitation limits,
and qualifiers on diskette for downloading into Risk*
Assistant (an automated tool to support risk assessment),
spreadsheets, or other software programs that the risk
EXHIBIT 19. DATA AND DOCUMENTATION NEEDED
FOR RISK ASSESSMENT
Data and Documentation
• Site description with a detailed map indicating site location, showing
the site relative to surrounding structures, terrain features, population or
receptors, indicating air and water flow, and describing the operative industrial
process if appropriate.
• Site map with sample locations (including soil depths) identified.
• Description of sampling design and procedures including rationale.
• Description of analytical method used and detection limits including
SQLs and detection limits for non-detect data.
• Results given on a per-sample basis, qualified for analytical limitations
and error, and accompanied by SQLs. Estimated quantities of
compounds/tentatively identified compounds.
• Field conditions and physical parameter data as appropriate for the media
involved in the exposure assessment.
• Narrative explanation of qualified data on an analyte and sample basis,
indicating direction of bias.
• QC data results for audits, blanks, replicates and spikes from the field and
laboratory.
• Definitions and descriptions of flagged data.
• Hardcopy or diskette results.
• Raw data (instrument output, chromatograms, spectra).
• Definitions of technical jargon used in narratives.
Importance
Critical
Critical
Critical
Critical
Critical
Critical
High
High
High
Medium
High
Low
KEY Critical = Essential to tne useability of data for risk assessment
High = Should be addressed in planning tor risk assessment
Medium = Primarily impacts how data are qualified in risk assessment
Low = Has little effect on useability of data for risk assessment
36
-------
assessor may use. An example of a recommended
report format for tabular results appears in Appendix I.
The data reviewer should provide a narrative summary,
which is comprehensible to a nonchemist, describing
specific sampling or analytical problems, data
qualification flags, detection limit definitions, and
interpretation of QC data. This summary must always
be followed and supported by a detailed commentary
that explicitly addresses each item from the narrative on
a technical basis. The explanation for data qualification
in the commentary facilitates data use. If a nontechnical
narrative is unavailable, the risk assessor must (at a
minimum) be provided with explanations of qualification
flags, detection limits, and interpretation of QC data
(see Appendices I, V and VI for examples). A chemist
familiar with the site can be requested to interpret the
analytical review with site-specific information, such as
physical site conditions that affect sample results.
3.2 PRELIMINARY SAMPLING AND
ANALYTICAL ISSUES
This guidance cannot encompass sampling design in the
assessment of environmental sampling and analysis
procedures; however, this section does sketch a
framework for these activities. It discusses key issues
for determining the potential impact of sampling and
analysis procedures on data useability for risk assessment
and for identifying situations that require statistical or
methodological support The sampling discussion
primarily focuses on soil issues, bu t some generalizations
can be made to other media such as sediment or
groundwater. Rules of thumb, reference tables, statistical
formats and checklists support the statistical
understanding and sophistication of RPMs and risk
assessors. A Sampling Design Selection Worksheet, a
Soil Depth Sampling Worksheet, and aMethod Selection
Worksheet are tools, presented with step-by-step
instructions in Chapter 4, to focus planning efforts.
Sampling issues. Resolving statistical and non-
statistical sampling issues provides the risk assessor,
project chemist, and QA personnel with a basis for
identifying sampling design and data collection
problems, interpreting the significance of analytical
error, and selecting methods based on the expected
contribution of sampling and analytical components to
total measurement error. Comprehensive discussions
of environmental sampling procedures are given in
Principles of Environmental Sampling (Keith 1987),
Environmental Sampling and Analysis (Keith 1990a),
Methods for Evaluating the Attainment of Cleanup
Standards (EPA. 1989e), and the Soil Sampling Quality
Assurance User's Guide (EPA 1989f).
Several assumptions concerning sampling and associated
statistical procedures have been made to simplify the
discussion in this section:
• The RPM and risk assessor are familiar with basic
environmental sampling and statistical terms and
logic and have access to a statistician.
• Sampling designs are mainly based on stratified
random or systematic random sampling (grid), or
variations thereof. Systematic sampling requires
special variance calculations for estimating
statistical performance parameters such as power
and confidence level; these calculations are not
provided in this guidance.
• Statisticians are consulted for any significant
problems or issues not covered in this guidance.
• Superfund contaminant concentrations for a site
generally fit a log-normal distribution.
Measurements of variability are generally given
in log-transformed units. Overviews of statistical
methodology include Gilbert (1987) and Koch
and Link (1971). Parametric tests in transformed
units (Aitchison and Brown 1957) have logarithmic
forms (Seichel 1956). Graphical methods of
determining re-transformed means and their 95%
confidence levels are available (Krige 1978).
• Quality assurance procedures for sampling and
analysis are not separate, even though the
discussion addresses them separately.
Exhibit 20 summarizes the importance of each of the
preliminary sampling planning issues to the risk
assessment, proposes planning actions to reduce or
eliminate their effect on data useability, and refers the
reader to further discussion in the text Information
relevant to preliminary sampling planning can be
obtained by collecting site maps, photographs and other
historical and current documents which depict
production, buildings, sewage andstorm drains, transport
corridors, dump sites, loading zones, and storage areas.
Areliableandcurrent base map is particularly important.
Data adequacy. All data users should clearly state the
level of data adequacy they desire. These statements,
and the resources that will be committed, should be
incorporated into the sampling plan objectives. If an
appropriate level of uncertainty cannot be determined at
this stage, an initial goal should be agreed on for the
final level of reliability, which may be revised during
the iterative sampling process. Since each site is unique,
it may be extremely difficult to attain a given level of
data adequacy. An iterative sampling program may
37
-------
EXHIBIT 20. IMPORTANCE OF SAMPLING ISSUES IN RISK ASSESSMENT
Issue
Chemicals of Potential
Concern
(3.2.1)
Sampling and
Analytical Variability
versus Measurement
Error (32.5)
Media Variability
(3.2.5)
Sample Preparation
and Sample
Preservation
(3.2.G)
Identification of
Exposure Pathways
(3.2.7)
Use of Judgmental or
Purposive Sampling
Design
(3.2.8)
Importance
Chemicals have different rates of
occurrence and coefficients of variation.
This impacts the probability of false
negatives and reduces confidence limits for
estimates of concentration.
Samping variability can exceed
measurement error by a factor of three to
four (EPA1989C).
Sampling variability increases uncertainty
or variability; measurement error
increases bias.
Samping problems vary widely by media as
do variability and bias.
Contamination can be introduced during
sample preparation, producing false
posiives. Fitering may remove
contaminants sorbed on particles.
Not aR samples taken in a site
characterization are useful for risk
assessment. Often only a few samples have
been taken in the area of interest.
Statistical sampling designs may be costly
and do not take advantage of known areas
of contamination.
Suggested Action L
Increase the number of samples for 1
chemicals with low occurrence and/or
high coefficients of variation.
Reduce sampling variability by taking
more samples (using less expensive
methods). This allows more samples
to be analyzed.
Use QC samples to estimate and
control bias. Prepare SOPs for
handling all field equipment.
Design media-specific sampling
approaches.
Use blanks al sources of potential
contamination. Collect filtered and
unfiKered samples.
Specifically address exposure
pathways in sampling designs. Risk
assessors should participate in
scoping meeting.
Use judgmental sampling to examine
known contaminated areas, then use
an unbiased method to characterize
exposure. •
allow a realistic appraisal of the variability present at the
site; a phased investigation may be warranted, with an
increase in data adequacy at each phase.
Natural variation. It is important to realize that natural
variation (environmental heterogeneity) in both soil
and water systems may be so great that variation due to
field sampling is significantly greater than that due to
laboratory analysis. For example, laboratory sample-
sample precision is commonly of the order of less than
1%, whereas soil sample-sample precision is commonly
between 30% to 40%. Sampling variation is influenced
by the homogeneity of material being sampled, the
number of samples, collection procedures, and the size
of individual samples.
Uncertainty in sampling measurements is additive.
Exhibit 21 lists the components of sampling variability
and measurement error. The final error associated with
an estimate is the sum of the errors associated with
natural variation (intrinsic randomness, microstructure,
macrostructure), plus sampling error, plus laboratory
measurement error. Poor sampling techniques can
swamp the natural phenomenon that is being evaluated.
Therefore, sampling options must be fully reviewed and
the probable uncertainty from sampling must be
acceptable.
Initial survey sampling plan. A preliminary sampling
plan should be chosen that provides a basis for evaluation
of overall sampling goals, sampling techniques,
feasibility, and statistical analysis techniques. General
categories of sampling plans include simple random,
stratified random, systematic, judgmental/purposive,
and spatial systematic. The features of these different
plans are discussed in more detail in Chapter 4.
Statistical analysis of the survey data allows evaluation
of how well the sampling program is doing. Depending
on the contaminant, current technology may allow on-
site "laboratory" analysis of the samples using portable
microcomputers and telecommunications. On-site
statistical analysis is also possible. On-site analysis
reduces project completion time and costs. In a truly
38
-------
EXHIBIT 21. SAMPLING
VARIABILITY AND
MEASUREMENT ERROR
Sampling variability: The variation
between true sample values that is a
function of the spatial variation in the
pollutant concentrations.
Measurement error The variation
resulting from differences between
true sample values and reported
values. Measurement error is a
function of uncertainty due to the
following:
• Sample collection variation
• Sample preparation/handling/
preservation/storage variation
• Analytical variation
• Data processing variation
iterative sampling campaign, on-site statistical analysis
can guide the sampling teams, maximizing information
capture and minimizing time-related costs.
Analytical issues. The following assumptions
concerning analytical procedures have been made in
this section:
• The RPM and the risk assessor are familiar with
standard analytical chemical procedures.
Reference books on environmental issues in
analytical chemistry are available and can be
consulted (ASTM 1979. Mananan 1975, Dragun
1988, Baudo, el. ai, eds. 1990, Taylor 1987).
• Chemists are available and will be consulted for
any significant problems or situations not covered
in this guidance.
• Analytical QA procedures are used in conjunction
with and affect sampling QA procedures, even
though the discussion treats these procedures
separately.
Exhibit 22 summarizes the importance of each analytical
issue to risk assessment, lists suggested actions during
the planning process, and refers the reader to further
discussion in the text Each issue is discussed in terms
of its effect on data quality for risk assessment, and bow
to anticipate and plan for potential problems. The RPM
should also consult the project chemist to determine the
appropriate sample volumes or weights required for
different types of analysis.
Biota sampling and analytical issues. The type of
assessment (e.g., human health or ecological) determines
the type of samples to be collected. An ecological
assessment may require analysis of the whole body or of
a specific organ system of a target species (because
organic, and some inorganic, chemicals of concern are
often concentrated in tissues with high lipid contents).
Human health risk assessment usually concentrates on
edible portions.
Typical sampling considerations for biota include
specifying the species to be sampled, sampling locations,
tissue to be analyzed, number of individuals to be
sampled, and the method of analysis of the chemical of
concern. Biota analyses should include a method
validation that incorporates tissues or plant analyte
spikes, and any available performance evaluation
materials. The purpose of spiking is to determine
whether tins analytes are recoverable from the matrix or
clean-up steps hinder detection of the analyte.
Spiking and duplicate information can be used to assess
method precision and accuracy. The primary source of
performanceevaluation materials is the National Bureau
of Standards repository. Samples and performance
evaluation materials should be matched by matrix
(species and whole/edible portions).
Volatile analytes are very difficult to measure in biota.
Samples should be stored on dry ice immediately after
collection. Fat and cholesterol can also block columns
and impede chromatography for base/neutral/acid
extractable tissue analysis. Gel permeation
chromatography procedures may only be marginally
effective in clean up, and the lipids present may retain
analytes of concern, thereby reducing recoveries. Plant
matrices are often difficult to digest, and a variety of
digestion procedures using hydrogen peroxide or
phosphoric acid may be warranted. Tissues for organic
analysis should be wrapped in aluminum foil for
shipment to the laboratory, and tissues formetals analysis
should be wrapped in plastic film. All tissues should be
sent frozen on dry ice.
Air sampling and analysis issues. Air sampling
procedures should account for wind speed and direction
as well as seasonal and daily fluctuations; they should
also account for the influence of these factors on the
exposed population (e.g., the largest population may be
potentially exposed in the evening when the wind speed
may be least). The definition of detection limits is very
important for air analyses. For example, the same
concentration will appear very different if expressed on
a weight/volume basis than on a volume/volume basis.
Sampling strategies may need to distinguish between
paniculate and gaseous forms of chemicals of concern.
It is important to collect media blanks to determine the
type and amount of contamination that may be found.
Blanks should also be provided to the laboratory for
spiking to determine analytical precision and accuracy.
39
-------
EXHIBIT 22. IMPORTANCE OF ANALYTICAL ISSUES
IN RISK ASSESSMENT
Analytical Issue
Chemicals of
Potential Concern
(3.2.1)
Tentatively Identified
Compounds
(3.2.2)
Identification and
Quantitation
(3.2.3)
Detection Limits
(3.2.4)
Media Variability
(3.2.S)
Sample Preparation
(3.2.6)
Field Analyses versus
Fixed Laboratory Analyses
(3.2.9)
Laboratory Performance
Problems
(3.2.10)
Importance
Chemicals of potential
lexicological significance may be
omitted.
Identification and quantitation do
not have high confidence.
False negatives may occur when
analytes are present near the
MDL
Significant nsk may result at
concentrations lower than
measurable.
Variability and bias may be
introduced to analytical
measurements.
Variability and bias may be
introduced to analytical
measurements.
Tradeoffs required with regard to
speed, precision, accuracy,
personnel requirements,
identification, quantitation and
detection limits.
Quality of data may be
compromised.
b
Suggested Action
Examine existing data and site history
for industry-specific wastes to
determine analytes for maasurement.
Perform broad spectrum analysis.
Be prepared to request further
analyses if potentially toxic
compounds are discovered during
screening. Compare results from
multiple samplings or historical data. I
Use technique with definitive 1
identification (e.g., GC-MS). •
Alternatively, use technique with I
definitive identification first, followed 1
by another technique (e.g., GC) to
achieve lower quantitation limits.
Review available methods for
appropriate detection limit.
Use environmental samples as QC
samples to determine recovery and
reproducibility in the sample media.
Select analytical methods based on
sample medium and strengths of the
sample preparation technique.
Consider options and set priorities.
Select experienced laboratory and
maintain communication.
The sample medium should be checked to ensure that
recovery rates are documented.
3.2.1 Chemicals of Potential Concern
Chemicals of potential concern are chemicals that may
be hazardous to human health or the environment and
are identified at thesite, initially firom historical sources.
Chemicals identified at Superfund sites have varying
rates of occurrence, average concentrations, and
coefficients of variation. Thesedifferencesarearunction
of fate and transport properties, occurrence in different
media, and interactions with other chemicals, in addition
to use and disposal practices. Information on frequency
of occurrence and coefficient of variation determines
the number of samples required to adequately
characterize exposure pathways and is essential in
designing sampling plans. Low frequencies of
occurrence and high coefficients of variation mean that
more samples will be required to characterize the
exposure pathways of interest Potential false negatives
40
-------
occur as variability increases and occurrence rates
decrease. From an ecological standpoint, chemicals of
potential concern may be different from those for human
health concerns. For example, copper is an analyte of
high concern from an ecological perspective, but of low
concern from a human health perspective. In addition,
if water quality criteria are used as lexicological
thresholds, it should be determined whether the criteria
are based on ecological or human health effects.
*• To protect human health, place a higher
priority on preventing false negatives in
sampling and analysis than on preventing
false positives.
Data are available for volatiles, extractable organics,
pesticides/PCBs, tentatively identified organic
compounds, and metals (see Appendix II), for aqueous
and soil/sediment matrices, and releases from industries
known to produce waste commonly found at Superfund
sites. Data from CLP Superfund sites are also available
for calculating site-specific coefficients of variation.
Exhibit 23 indicates the occurrence rates and coefficients
of variation for selected chemicals of potential concern
to risk assessors. Many other chemicals (which are not
of concern) may be present without affecting the level
of risk to the exposed population.
w Use preliminary data to identify chemicals
of potential concern and to determine any
need to modify the sampling or analytical
design.
The need for risk assessment indicates that there is
already some knowledge of contamination at the site.
Based on available lexicological and site data, the risk
assessor can recommend target chemicals (or chemical
classes) for analysis and desired detection limits. For
example, explosive chemicals are likely to be present at
a former munitions site. Exhibit 24 presents data on
munitions compounds, such as feasible detection limits
and health advisory limits.
Information on industry-specific analytes is summarized
in Exhibit 25 and detailed in Appendix n. If historical
data are incomplete, a broad spectrum analysis should
be performed on selected samples from each sampling
location to provide necessary scoping information.
The RPM or risk assessor should inform the planning
team about chemicals of potential concern at the site,
exposure path ways, if known, concentrationsof concern,
and other pertinent information, particularly any
requirement to distinguish specific states of the chemicals
of potential concern. Some oxidation states of metals
(e.g., chromium) are more easily absorbed or are more
toxic than others, and organically substituted metals
such as mercury are more toxic than their elemental
states. If these concerns are imporlant, analyses thai
determine metal specification rather than elemental
analyses should be performed, if available. Similarly,
for organic compounds, such as tetrachloroethane,
degradation products or metabolites may be more toxic
than the parent compounds. In this case, sampling
procedures and analytical methods should include the
parentcompound, degradation products, and metabolites
of chemicals of potential concern.
3.2.2 Tentatively Identified
Compounds
Gas chromatography-mass spectrometry (GC-MS)
analyses categorize organic compounds in two ways.
Target compounds are those compounds for which the
GC-MS instrument has been specifically calibrated
using authentic chemical standards. A target compound
in an environmental sample is identified by matching its
mass spectrum and relative retention time (RRT) to
those obtained for the authentic standard during
calibration. Quanutation of a target compound is
achieved by comparison of its chromatographic peak
area to that of an internal standard compound, normalized
to the relative response factor (RRF) which is the ratio
of the peak areas of the authentic chemical standard and
the internal standard measured during calibration.
<*• Specific analysis for compounds ident-
ified during library search can be requested.
Tentatively Identified Compounds (TICs) are any other
compounds which are reported in the sample analysis,
but for which the GC-MS instrument was not specifically
calibrated. A TIC is identified by taking its mass
spectrum from the environmental sample, and comparing
it to a computerized library of mass spectra.
Computerized comparison routines score the various
library spectra for their similarity to the TIC and rank
the spectra most similar to the TICs spectrum. If the
TIC is reported as a specific compound, it is usually
reported to be one of the compounds whose spectra
were retrieved in the library search. Quanutation of a
TIC is less accurate than for target compounds, because
the true RRF is not known (since no calibration for this
specific compound was performed). The RRF is assumed
to be 1.0; whereas, measured RRFs below 0.05 and
above 10.0 are known.
Confidence in the identification of aTIC can be increased
in several ways. The main steps in identifying and
quantitating TIC data are summarized in Exhibit 26.
An analytical chemist trained in the interpretation of
mass spectra and chromatograms can review TIC data
41
-------
EXHIBIT 23. MEDIAN COEFFICIENT OF VARIATION FOR
CHEMICALS OF POTENTIAL CONCERN ^
Chemical of
Potential Concern
Chlorome thane
Trichloromethan a/Chloroform
Tetrachloromethane/Carfaon tetrachlonde
1,2-Dichloroethane
Tetrachloroe thane
Vinyl chloride
Tetrachloroe the ne
Dichloropropane
Isophorone
Bis (2-chloroethyl) ether
1,4-Dtehlorobenzene
Bis (2-ethylhexyl) phthalate
Benzo(a) pyrene
Styrene
N-nitrosodiphenylamine
DOE
DDT
Dieldrin
Heptachlor
Gamma-BHC (lindane)
PCB1260
Arsenic
Beryllium
Cadmium
Chromium
Mercury
Lead (Pb)
Soil/Sediment
Median %Cv2
16.7
53.9
15.4
17.6
17.0
11.0
24.5
19.0
0.7
0.5
0.9
0.7
0.5
16.9
0.5
4.5
2.9
4.4
4.8
6.3
0.21
40.3
271.3
134.6
11.9
1032.3
10.9
Number of Sites
at Which Chemical
was detected3
61
392
38
64
56
55
392
29
74
10
120
1197
1058
117
142
329
521
274
249
142
251
1098
1091
1096
1096
1098
1098
Water
Median %CV2
50.0
45.2
9.3
24.7
17.4
15.7
33.3
13.3
18.4
20.1
17.3
29.5
10.8
33.3
30.5
813.0
5882
3.3
351.9
454.1
41.7
58.0
100.0
33.7
23.0
500.0
97.3
Number of Sites
at Which Chemical
was detected3
134
519
90
158
101
197
367
79
72
34
119
782
76
69
96
40
125
101
151
134
23
940
931
945
948
948
939
1 List of chemicals of potential concern is derived from health-based levels and frequency of occurence at Superfund
sites listed in the CLP Statistical Database. (Number of sites for which data exist totals 8.900.)
2 Median percent coefficient of variation of analyte concentrations.
3 November 1988 to present
I
42
-------
EXHIBIT 24. MUNITIONS COMPOUNDS AND THEIR
DETECTION LIMITS
Health
Advisory
Acronym
Compound Name
Detection Limit'
(PRb)
HMX Octahydro-1,3,5.7-tetranitro-1,3,5,7-tetrazocine
RDX Hexahydro-1,3,5-trinitro-1,3,5-triazine
— Nitrobenzene
TNB 1,3,5-Trinitrobenzene
DNB 1,3-Dinitrobenzene
Tetryl Methyl-2,4,6-trinitrophenylnitramine
TNT 2,4,6-Trinitrotoluene
2,4 DNT 2,4-Dinitrotoluene
TAX Hexahydro-1 -(N)-acetyl-3,5-dinitro-1,3,5-triazine
SEX Octahydro-1-(N)-acetyl-3,5,7-trinitro-1,3,5,7-tetrazocine
2,6 DNT 2,6-Dinitrotoluene
2,4,5 TNT 2,4,5-Trinitrotoluene
2 Am DNT 2-Amino-4,6-dinitrotoluene
4 Am DNT 4-Amino-2,6-dinitrotoluene
2,4 DAmNT 2,4-Diamino-6-nitrotoluene
2,6 DAmNT 2,6-Diamino-4-nitrotoluene
DIMP Disopropyl-methylphosphonate
TNG Gylcerol trinitrate (Nitroglycerin)
— Nitrocellulose
DMMP Dimethyl methylphosphonate
NG Nrtroguanadine
Health advisory complete.
Health advisory in preparation (1990).
5.1
Depending upon matrix and instrument conditions, these compounds may be chromatographable
and may be tentatively identified as indicators of the presence of munitions during GC-MS library
search procedures.
2
Detection limits are provided where available. Specific compounds with complete health advisories
are designated as target analytes with defined detection limits specified in a high performance liquid
chromatographic method developed and provided by the U.S. Army Toxic and Hazardous
Materials Agency.
21-002-024
43
-------
EXHIBIT 25. SUMMARY OF MOST FREQUENTLY OCCURRING
CHEMICALS OF POTENTIAL CONCERN BY INDUSTRY*
Compound
Acetone
Aluminum
Ammonia
Ammonium Nitrate
Ammonium Sulfale
Anthracene
Arsenic
Benzene
Biphenyf
Chlorine
Chlorobenzene
Chromium
Copper
Cydohenne
Dibenzofuran
Dichlcfomeihane
Formaldehyde
Freon
GlycolBhera
Hydrochlonc Acid
Lead
Manganese
Methanol
Methyl BhylKetone
Naphthalene
Nickel
Nitric Acid
_
Propytane
Sodium SuKate
Sodium Hydrando
SurtuncAod
f • 1.1 _»fc.
Toluene
Titanium Tetrachkxide
Xytene
1,1,1-trichkxoethane
Industry Lj
1
X
X
X
X
X
X
X
X
X
X
2
X
X
X
X
X
X
X
X
X
X
3
X
X
X
X
X
X
X
X
X
X
4
X
X
X
X
X
X
X
X
X
X
s
X
X
X
X
X
X
X
X
X
X
c
X
X
X
X
X
X
X
X
X
X
7
X
X
X
X
X
X
X
X
X
X
KEY 4 'Electroplating
1 > Battery Recycling 5 • Wood Preservatives
2 -Munitions/Explosives 6 • Leather Tanning
3 « Pesticide Manufacturing 7 » Petroleum Refining
'Summarized from Appendix II.
44
-------
EXHIBIT 26. STEPS IN THE
ASSESSMENT OF TENTATIVELY
IDENTIFIED COMPOUNDS
Identification • GC-MS analysis indicates the
presence of a tentatively
identified compound.
• Incorporate retention
time/retention index matching
and use physical
characteristics (boiling point (
or vapor pressure) to
determine if identification is
reasonable.
• Examine historical data and
industry-specific compound
lists.
• Reanalyze sample with an
authentic standard.
Quantitation • Assess known analytical
response characteristics for
similar compounds or similar
compound classes.
• Determine response
characteristics by analysis of
an authentic standard.
mass spectra and cbromatograms can review TIC data
and eliminate many false positive identifications. The
use of retention indices or relative retention times can
confirm TICs identified by the GC-MS computer (Eckel,
et. al. 1989). Examination of historical data, industry-
specific compound lists, compound identifications from
iterative sampling episodes, and analyses performed by
different laboratories may also increase confidence in
the identification of a TIC. The final identification step
is to reanalyze the sample after calibrating the GC-MS
instrument with an authentic standard of the compound
that the TIC is believed to be.
If toxic compounds are identified as TICs by this type of
broad spectrum analysis, the RPM or risk assessor
should request further analyses to positively identify the
compound and to accurately quantitate it. The risk
assessor or RPM should discuss data requirements with
an analytical chemist to determine the appropriate
analytical method.
Many compounds that appear as TICs during broad
spectrum analyses belong to compound classes.
Examples of compound classes are saturated aliphatic
hydrocarbons and polycyclic aromatic hydrocarbons
(PAHs). The risk assessor may be able to make a
preliminary judgment of toxicity at the compound class
level without a definitive identification of each
compound present. For example, in a sample
contaminated by gasoline, organics analysis would
indicate a series of TICs as aliphatic hydrocarbons of
increasing size. These may not be carcinogenic, and
more precise identification may not be required. If a
similar sample were contaminated with coal tar, larger
hydrocarbons and a series of PAHs would be found
during the analysis. The aliphatic hydrocarbons are not
especially toxic, but the PAH compound class contains
carcinogens and are of greater concern.
3.2.3 Identification and Quantitation
A risk assessor first confirms chemical identification,
and then determines the level of contamination. This
section summarizes the effects of detection limits and
sample contamination considerations on the confidence
inanalyteidentificationandquantitation. Requirements
for confidence are specified in Exhibit 27. When
analytes have concentrations of concern approaching
method detection limits, the confidence in both
identification and quantitation is low. This case is
illustrated in Exhibit 28. In addition, confidence in
identifying and quantitating as representative of site
EXHIBIT 27. REQUIREMENTS FOR
CONFIDENT IDENTIFICATION AND
QUANTITATION
Identification • Analyte present above the IDL
• Organic - Retention time and/or
mass spectra matches authentic
standards.
• Inorganic - Spectral absorptions
compared to authentic
standards.
• Knowledge of blank
contamination (if any).
Quantitation • Instrument response known
from analysis of an authentic
standard.
• Detected concentration above
the limit of quantitation and
within the limit of linearity
(instrument response not
saturated).
45
-------
EXHIBIT 28. RELATIVE IMPACTS OF DETECTION LIMIT
AND CONCENTRATION OF CONCERN: DATA PLANNING
Relative Position of Method
Detection Limit (MDL) and
Concentration of Concern (COC)
Consequence
Confidence MDL
Limits
Confidence
Limits
Non-Detects and
Detects Useable
Concentration
Possibility of
False Positives and
False Negatives
COC
MDL
Concentration
Non-Detects Not
Useable
Detects Useable
Possibility of False
Negatives
conditions is potentially diminished if the chemicals of
potential concern are present as contaminants from
laboratory or field procedures. This section identifies
analy tes and cites situations in which this is most likely
to occur.
The first requirement of analysis is confidence in the
identification of chemicals of potential concern.
Identification means that the chemical was present in
the environmental sample above the detection limit.
Chemicals can be correctly identified at lower
concentrations than are suitable for accurate quantitation.
If lower quantitation limits are required for risk
assessment purposes, a larger initial sample size may be
processed, or the sample extract may be concentrated to
a smaller final volume. However, concentration of an
extract to a smaller volume, or increasing the sample
size, may saturate the instrument in the presence of
matrix interferences. The RPM should discuss these
issues with an analytical chemist to determine the best
approach. A further discussion of limits of quantitation
is presented in Section 3.2.4. and Appendix III.
To ensure maTimmn confidence in the identification of
an organic chemical contaminant, an instrumental
technique, such as mass spectrometry, that provides
definitive results is necessary. Although alternative
techniques are available, GC-MS determination is the
best available procedure for confident identification or
confirmation of volatile and extractable organic
chemicals of potential concern. The application of this
technique minimizes the risk of error in qualitative
identification and measures chemicals of potential
concern at environmental levels above the detection or
quantitation limits listed in Appendix HI. In cases
where the target detection limit b; too low to allow
46
-------
but more definitive, instrumental techniques can be
used.
The identification of inorganic chemicals is more certain.
A reported concentration determined by atomic
absorption (AA) spectroscopy or inductively coupled
plasma (ICP) atomic emission spectroscopy is generally
considered evidence of presence at the designated level
reported, provided there is no interference. If
interferences exist, the laboratory should try to
characterize the type of interferences (background,
spectral or chemical) and take the necessary steps to
correct them.
3.2.4 Detection and Quantitation
Limits and Range of Linearity
The following discussion is intended to provide the
RPM and risk assessor with an understanding of the
various ways that detection or quantitation limits can be
reported. The term "detection limit" is frequently used
without qualification. However, there are several
methods for calculating detection limits. The RPM
should consult with the project chemist and the risk
assessor whenever analytical methods are to be selected,
Common Detection and Quantitation Limits
Instrument detection limit The DDL includes
only the instrument portion of detection, not
sample preparation, concentration/dilution
factors, or method-specific parameters.
Method detection limit The MDL is the
minimum amount of an analyte that can be
routinely identified using a specific method.
The MDL can be calculated from the IDL by
using sample size and concentration factors
and assuming 100% analyte recovery.
Sample quantitation limit The SQL is the
MDL adjusted to reflect sample-specific action
such as dilution or use of a smaller sample
aliquot for analysis due to matrix effects or the
high concentration of some analytes.
Contract required quantitation (detection)
limit The CRQL for organics and CRDL for
inorganics are related to the SQL that has been
shown through laboratory validation to be the
lower limit for confident quantitation and to be
routinely within the defined linear ranges of
the required calibration procedures.
Practical quantitation limit The PQL,
defined in S W846 methods, is the lowest level
that can be reliably achieved within specified
limits of precision and accuracy during routine
laboratory operating conditions.
and specify the nature of the detection limits that must
be reported; itisthelaboratory'sresponsibilitytoadhere
to this requirement. If no requirement has been specified,
then the laboratory should be requested to explicitly
describe the types of the detection limits it reports.
Detection limits can be calculated for the instrument
used for measurement, for the analytical method, or as
a sample-specific quantitation limit. The risk assessor
should request that the sample quantitation limit (SQL)
be reported whenever possible. The term "detection
limit" should be considered generic unless the specific
type is defined. Exhibit 29 illustrates the relationship
between instrument response and the quantity of analyte
presented to the analytical system (i.e., a calibration
curve).
»• The closer the concentration of concern
is to the detection limit, the greater the
possibility of false negatives and false
positives.
*• The wide range of chemical concen-
trations in the environment may require
multiple analyses or dilutions to obtain
useable data. Request results from all
analyses.
The definitions that follow are intended to provide the
RPM and risk assessor with an understanding of the
various methods for calculating detection limits, the
terms used to describe specific detection limits, and the
limitations associated with identification and
quantitation of chemicals of potential concern at
concentrations near specified detection limits.
Understanding the different terms used to describe
detection limits helps avoid reporting problems. Exhibit
30 provides examples of calculations of the three most
commonly reported types of detection limits.
*• Define the type of detection or quanti-
tation limit for reporting purposes; request
the sample quantitation limit for risk
assessment.
Instrument detection limit The instrument detection
limit (IDL) includes only the instrument portion of
detection, not sample preparation, concentration/dilution
factors, or method-specific parameters. The IDL is
operationally defined as three times the standard
deviation of seven replicate analyses at the lowest
concentration that is statistically different from a blank.
This represents 99% confidence that the signal identified
is the result of the presence of the analyte, not random
noise. The IDL is not the same as the method detection
limit. Use of the DDL should be avoided for risk
assessment.
Method detection limit The method detection limit
47
-------
EXHIBIT 29. THE RELATIONSHIP OF INSTRUMENT
CALIBRATION CURVE AND ANALYTE DETECTION
Region of
Unknown Identification and
Quantitation
Region of Known
Quantitation
of Less
Certain
Quantitation
Region of
Less Certain
Quantitation
g Region
g of Less
£ Certain
IOL • Instrument Detection Limit
MDL - Method Detection Limit
LOG » Limit of Quantitation
LOL • Limit of Linearity
Concentration
IDL MDL LOQ LOL
Method detection limit The method detection limit
(MDL) is the minimum amount of an analyte that can be
routinely identified using a specific method. The MDL
can be calculated from the IDL by using sample size and
concentration factors and assuming 100% analyte
recovery. Tbisestimateofdetectionlimitmaybebiased
low because recovery is frequently less than 100%.
MDLs are operationally determined as three times the
standard deviation of seven replicate spiked samples
run according to the complete method. Since this
estimate includes sample preparation effects, the
procedure is more accurate than reported IDLs.
However, the evaluation is routinely completed on
reagent water. As aresult, potentially significant matrix
interferences that decrease analyte recoveries are not
addressed.
The impact of an MDL on risk assessment is illustrated
in Exhibit 28. When planning to obtain analytical data,
the risk assessor knows the concentration of concern or
preliminary remediation goal. When the concentration
of concern of an analyte is greater than the MDL, to the
extent that the confidence limits of both the MDL and
concentration of concern do not overlap, then both
"non-detect" and "detect" results can be used with
confidence. There will be a possibility of false positives
and false negatives if the confidence limits of the MDL
and concentration of concern overlap. When the
concentration of concern is sufficiently less than the
MDL that the confidence limits do not overlap, then
there is a strong possibility of false negatives and only
"detect" results are useable.
48
-------
EXHIBIT 30. EXAMPLE OF DETECTION LIMIT CALCULATION
IDL = 3 x SD* of replicate injections
Example: 100 ppb pentachlorophenol standard
If: SD = 5 ppb
Then: IDL = 3 x 5 ppb = 15 ppb
MDL = 3 x SD of replicate analyses (extraction and injection)
Example: 100 ppb pentachlorophenol spiked in sample producing average measured
concentration of 50 ppb (not all analyte is recovered or measured)
SD = 18ppb
MDL = 3x 18ppb = 54ppb
Incorporate calculation of MDL from IDL
SQL = MDL corrected for sample parameters
100 ppb pentachlorophenol with MDL of 57 ppb
Dilution factor = 10 (sample is diluted due to matrix interference or high
concentrations of other analytes)
SQL = 10 x 57 ppb = 570 ppb
SD = Standard Deviation
Sample quantitation limit. The SQL is the MDL
adjusted to reflect sample-specificaction such as dilution
or use of smaller aliquot sizes than prescribed in the
method. These adjustments may be due to matrix
effects or the high concentration of some analytes. The
SQL is the most useful limit for the risk assessor and
should always be requested.
For the same chemical, the SQL in one sample may be
higher than, lower than, or equal to SQL values for other
samples. In addition, preparation or analytical
adjustments, such as dilution of the sample for
quantitation of an extremely high level of one chemical,
could result in non-detects for other chemicals included
hi the analysis, even though these chemicals may have
been present at trace quantities in the undiluted sample.
The risk assessor should request results of both original
and dilution analyses in this case. Since the reported
SQLs take into account sample characteristics, sample
preparation, and analytical adjustments, they are the
most relevant quantitation limits for evaluating non-
detected chemicals.
Contract required quantitation (detection) limit.
The CLP specifies a contract required quantitation limit
21-002-030
(CRQL) for organics and a contract required detection
limit (CRDL) for inorganics. Each of these quantities is
related to the SQL that has been shown through laboratory
validation to be the lower limit for confident quantitation
and to be routinely within the defined linear ranges of
the required calibration procedures.
The use of CRQLs and CRDLs attempts to maintain the
analytical requirements within performance limits
(which are based upon laboratory variability using a
variety of instruments). CRQLs are typically two to five
times the reported MDLs and they generally correspond
to the limit of quantitation.
Practical quantitation limit Thepracticalquantitation
limit (PQL), defined in SW846 methods, is the lowest
level that can be reliably achieved within specified
h^tsof precision andaccuracyduringroutine laboratory
operating conditions. It is important to note that the
SQL and PQL are not equivalent Use of PQL values as
measures of quantitation limits should be avoided
wherever possible in risk assessment.
Other quantitation measurements. The limit of
quantitation (LOQ) is the level above which quantitative
49
-------
results may be obtained with a specified degree of
confidence. At analyte concentrations close to, but
above the MDL, the uncertainty in quantitation is
relatively high. Although the presence of the analyte is
accepted at 99% confidence, the reported quantity may
be in the range of ±30%. Ten times the standard
deviation measured for instrument detection is
recommended to demonstrate a level at which confidence
is maximized (Borgman 1988).
The limit of linearity (LOL) is the point at or above the
upper end of the calibration curve at which the
relationship between the quantity present and the
instrument response ceases to be linear (Taylor 1987).
Instrument response usually decreases at the LOL, and
the concentration reported is less than the amount
actually present in the sample because of instrument
saturation. Dilution is necessary to analyze samples in
which analyte concentrations are above the LOQ.
However, dilutions correspondingly increase SQLs.
Data shouldbe requested from both diluted and undiluted
analyses.
3.2.5 Sampling and Analytical
Variability Versus
Measurement Error
Sampling and analytical variability and measurement
error are two key concepts in data collection. Each is
discussed in the context of evaluating strategies for the
collection and analysis of both site and background
samples.
Exhibit21 defines sampling variability andmeasurement
error. Most SAPs are a necessary compromise between
cost and confidence level Basically, two types of
decisions must be made in planning:
• What statistical performance is necessary to
produce the quality of data appropriate to meet the
risk assessor's sampling variability performance
objectives and
• What types and numbers of QC samples are
required to detect and estimate measurementenor.
«• When contaminant levels in a medium
varywidely,increasethenumberofsamples
or stratify the medium to reduce variability.
Sampling plans attempt to estimate and minimize both
sampling variability and measurement error. Sampling
variability affects the degree of confidence and power
theriskassessorcanexpectfromtheresults. Confidence
is the ability to detect a false positive hypothesis, and
power is the ability to detect a false negative. Power is
more important for risk assessment An estimate of the
sampling variability that is a function of the spatial
variation in the concentrations of chemicals of potential
concern is obtained by calculating the coefficient of
variation for each chemical. When the coefficient of
variation is less than 20% and a substantial quantity of
data are available, the effect of spatial and temporal
variation on concentrations of chemicals of potential
concern is minimal, and the power and certainty of
statistical tests is high (EPA 1989c).
Spatial variability can be analyzed after an initial
sampling effort through simple statistical summation or
through the use of variogram analysis, a part of the
geostatistics. EPA has developed software to assist a
risk assessor in this analysis: Geostatistical
Environmental Assessment Software (GEOEAS) (EPA
1988c) and Geostatistics for Waste Management
(GEOPACK) (EPA 1990b).
Measurement error is estimated using the results of QC
samples and represents the difference between the true
sample value and the reported value. This difference
has five basic sources: the contaminant being measured,
sample collection procedures, sample handling
procedures, analytical procedures, and data production
procedures. Measurement error due to analytical
procedures is discussed in Section 3.2 under analytical
issues. Measurement error due to sampling is estimated
by examining the precision of results from field
duplicates. The minimum recommended number of
field duplicates is 1 for every 20 environmental samples
(5%). A minimum of one set of duplicates should be
taken per medium sampled unless many strata are
involved; five sets are recommended. Exhibit 31
summarizes the types and uses of QC samples in defining
variation and bias in measurement.
<*• Sampling variability typically contributes
much more to total error than analytical
variability.
In summarizing the discussion of sampling variability
and measurement error, one finding puts the concepts in
perspective: "An analysis of the components of total
error from soils data from an NPL site sampled for PCB s
indicated mat 92% of the total variation came from the
location of the sample and 8% from the measurement
process" (EPA 1989f). Of the 8%, less than 1% could
be attributed to the analytical process. The rest of the
8%is attributable to sample collection, samplehandling,
data processing and pollutant characteristics. Sampling
variability is often three to four times that introduced by
measurement error. Exceptions to this observation on
the components of variation or sources of error occur in
instances of poor method performance for specific
analytes.
50
-------
EXHIBIT 31. MEASUREMENT OF VARIATION AND BIAS
USING FIELD QUALITY CONTROL SAMPLES
Quality Control
Sample Types
Variation or Bias Measured
Field duplicate
Field blank
Field rinsate
Trip blank
Provides data required to estimate the sum of
subsampling and analytical variances.
Provides data required to estimate the bias due to
contamination introduced during field sampling or
cleaning procedures. Also measures contamination at
laboratory. Compare with laboratory method blank
to determine source of contamination.
Provides data required to estimate the sum of the bias
caused by contamination at the time of sampling from
sampling equipment and by analysis and data handling.
Indicates cross-contamination and potential contamination
due to sampling devices.
Provides data required to estimate the bias due to
contamination from migration of volatile organics into the
sample during sample shipping from the field and sample
storage at the laboratory.
Source: EPA 1990c.
Media or matrix variability. Appropriate samples
must be collected from each medium of concern and, for
heterogeneous media, from designated strata.
Stratification reduces variability in results from
individual strata, which can be differentlayers or surface
areas. Media to be sampled should include those
currently uncontaminated but of concern, as well as
those currently contaminated. For media of a
heterogeneous nature (e.g., soil, surface water, or
hazardous waste), strata should be established and
samples specified by stratum to reduce variability, the
coefficient of variation and the required number of
samples.
Sampling considerations vary according to media. The
sampling concern may involve contaminant occurrence,
temporal variation, spatial variation, sample collection,
or sample preservation. Exhibit 32 indicates potential
sampling problem areas for each medium. Problem
areas are classified relative to other media. RPMscan
use this exhibit to plan for possible sampling problems
in the data collection design. Sampling designs must be
structured to identify and characterize hot spots.
Information needed for fate and transport modeling
should be obtained during a site sampling investigation.
This information also differs by the medium of concern
(EPA 1989a).
The type of medium in which a chemical is present
affects the potential sensitivity, precision, and accuracy
of the measurement. Sharp distinctions occur in applying
a single method to media such as water, oil, sludge, soil,
or tissue. Medium or matrix problems are indicated by
the presence of analytical interferences, poor recovery
of analytes from the matrix, physical problems such as
viscosity (flow parameters), and paniculate content that
affect sample processing. Exhibit 33 shows the sources
of uncertainty across media. Spiked environmental
samples monitor the effect of these sourcesof uncertainty
on the accuracy of recovery of target compounds from
the matrix. Duplicates quantify the effect of these
parameters on precision. The method must be chosen
carefully if a difficult medium such as oily waste or soil
is to be analyzed. Routine methods usually specify the
medium or media for which they are applicable.
Method detection and general confidence in analytical
determinations are also often affected by specific media
types and by analytical interference. The impact of
matrix interference on detection limits, identification,
51
-------
EXHIBIT 32. SAMPLING ISSUES AFFECTING CONFIDENCE
IN ANALYTICAL RESULTS
Major
Sampling
Issues Soil
Contaminant VV
Migration
Temporal
Variation
Spatial VV
Variation
Topographic/ VV
Geological
Properties
Hot Spots VV
Sample V
Collection
Sample VV
Preparation/
Handing
Sample
Storage
Sample
Preservation
Problem Likelihood by Medium
Ground Surface
Water Water Air Biota
VV
vv
V
VV
VV
V V
VV V
- VV V V
V
VV VV
VV VV V
VV VV VV
VV VV
Key: VV - Likely source of significant sampling problem.
V • Potential source of sampling problem.
Source: Modified from Keith 1990b.
Hazardous I
Waste I
VV I
VV
VV
V
V
and quantitation is illustrated by the following
discussions (which arc not meant to be comprehensive).
• Oil and hydrocarbons affecting GC-MS analyses,
• Phthalates and non-pesticide chlorinated
compounds that can interfere with pesticide
analyses, and
• Iron spectral interference affecting ICP sample
results.
Oil and hydrocarbons. The presence of appreciable
concentrations of oil and other hydrocarbons may
interfere with the extraction or concentration process.
Also, even at low concentrations, oil in a sample usually
produces a large series of chromatograpbic peaks that
interfere with the detection of other chemicalsof potential
concern during gas chrornatography. Any chemicals of
potential concern that may elute concurrently from the
GC column are obscured by the hydrocarbon response
and may not present a distinct spectrum. Also,
hydrocarbons that are present hi significant quantity are
often identified as TICs, potentially adding a large
number of compounds for consideration by the risk
assessor.
During RI planning, the risk assessor should determine
if there is a potential for hydrocarbon contamination,
through knowledgeof historical site use and examination
of historical data. The laboratory can be instructed to
add cleanup protocols to the analysis, or to use a
supplemental analysis for which the hydrocarbons are
not interferences (e.g., electron capture detection for
halogenated compounds).
Phthalates and non-pesticide chlorinated
compounds. Phthalates interfere with pesticideanalyses
by providing a detector response similar to that for
chlorinated compounds. Phthalates and non-pesticide
52
-------
EXHIBIT 33. SOURCES OF UNCERTAINTY THAT FREQUENTLY
AFFECT CONFIDENCE IN ANALYTICAL RESULTS
Degree of Significance by Medium
Source of
Uncertainty Soil
SAMPLING
Design VV
Contamination VV
Collection V
Preparation VV
Storage
Preservation
LABORATORY
Storage VV
Preparation VV/V
Analysis VV
Reporting
ANALYTE-SPECIFIC
Volatility VV
Photodegradation
Chemical Degradation V
Microbial Degradation VV
Contamination VV
Water
V
V
VV
VV
VV
VV
VV/V
V
V
VV
VV
Air Biota
VV
V
VV
VV
VV
VV
V V
V VV
V
V
VV
Hazardous
Waste
V
V
VV
VV
VV
KEY:
VV » LJkely source of significant error or uncertainty.
V » Potentially source of significant error or uncertainty.
VV/V = Magnitude of effect determined by examination of data.
chlorinated compounds are often present in greater
concentrations than the pesticides of concern. Pesticide
data are often required at low detection limits and,
therefore, GC-MS analyses are not used for quancitation.
In these cases, a gas chromatographic analysis using
electron capture detection is more sensitive, providing
a wider useful range of detection. The phthalates and
chlorinated compounds can coelute with chemicals of
potential concern, thereby obscuring the detection of
target analytes and raising the analyte-specific
quantitation limit. Phthalates and chlorinated
compounds also produce additional peaks on the
chromatogram that can be interpreted as false positive
responses to pesticides. A second analysis using a
different column provides an extrameasureof confidence
in identification. Alternatively, sample extracts from
positive analyses can be further concentrated for
confirmation by GC-MS if concentrations of analytes
are sufficient
Iron. Large quantities of iron in a sample affect the
detection and quantitation of other metallic elements
analyzed by ICP atomic emission spectroscopy at
wavelengths near the iron signals. The strong iron
response overlaps nearby signals, thereby obscuring the
results of potentially toxic elements present at much
lower concentrations. An interference check sample for
ICP analyses monitors the effect of such elements. High
concentrations of iron are analyzed with low
concentrations of other metals in these samples to
indicate whether iron interfered with metal detection at
lower concentrations. If spectral interferences are
observed, data may be qualified as overestimated. The
risk assessor or RPM should consult the project chemist
to determine if a particular method requires a
performance check.
53
-------
3.2.6 Sample Preparation and
Sample Preservation
Some samples require preparation in the field to ensure
that the results of analyses reflect the true characteristics
of the sample. Sample filtration and compositing
procedures are discussed in this section. Exhibit 34
summarizes the issues which the various sample
preparation methods address. Exhibit 35 outlines the
primary information gained with the various sampling
techniques.
EXHIBIT 34. SAMPLE
PREPARATION ISSUES
EXHIBIT 35. INFORMATION
AVAILABLE FROM DIFFERENT
SAMPLING TECHNIQUES
Issue
Sample
Integrity
Source of
Analyte
Media
Analyte
Special! on
Large
Number of
Samples to
be Analyzed
Action
Preservation — acids, biocides
(may be applicable to vdatiles
or metals).
Unfiltered samples - measure
total analytes
Rltered samples — discriminate
sorbed and unsorbed analytes
Choice of sample preparation
protocols affects analyte
speciation
Composite samples
(However, this raises the
effective detection limit in
proportion to the number of
samples composited.) 1
Filtration. If the risk assessor needs to discriminate
between the amount of analyte present in true solution
in a sample and that amount sorbed to solid particles,
then the sample must be filtered and analyses should be
performed for both filtered and unfiltered compounds.
Some samples, such as tap water, are never filtered
because there is no paniculate content Filtration should
be performed in the field as soon as possible after the
sample has been taken and before any preservative has
been added to the sample. Filtration often does not
proceed smoothly. It is common practice only to filter
a small proportion of all samples taken, and to perform
analyses for the total content of the analyte in the
majority of samples. Filtered samples generally provide
a good indication of the fraction of contaminant likely
to be transported over large distances horizontally in a
plume. However, in the immediate vicinity of a source
orpoint of exposure, unfiltered samples may be valuable
in providing an indication of suspended material that
Sample
Type
Filtered
Unfiltered
Grab
Composite
Information
Can differentiate sorbed
and unsorbed analytes.
Total amount of analyte
in sample is measured.
Can be used to locate
hot spots.
Can provide average
concentrations over an
area at reduced cost.
21-002435
may act as a source or sink of dissolved contaminants
and may therefore modify overall transport.
Compositing. Reducing the number of samples by
compositing is also a form of sample preparation.
Compositing may be performed to reduce analytical
costs, or in situations where the risk assessor has
determined that an average value will best characterize
an exposure pathway. Compositing cannot be used to
identify hot spots, but can be effective when averaging
across the exposure area. Caution should be exercised
when compositing since low level detects can be
averaged out and become non-detects.
Preservation. Sample characteristics can be disturbed
by post-sampling biological activity or by irreversible
sorption of analytes of concern onto the walls of the
sample container. A variety of acids and biocides used
for preservation are discussed in standard works such as
Standard Methods for the Examination of Water and
Wastewater (Clesceri, et. al., eds. 1989). Samples are
also usually shipped with ice toreduct; biological activity.
Preparation. Several factors in sample preparation
affect analytical data. These factors include sample
matrix, desired detection limit, extraction solvent,
extraction efficiency, sample preparation technique,
and whether the analysis is performed in the field or in
a fixed laboratory. In addition, parameters such as
turnaround time may preclude the use of some sample
preparation alternatives.
An extraction method must be able to release the
chemicals of concern from the sample matrix. For
example, organic solvents will extract non-polar organic
compounds from water. Polar and ionic compounds
54
-------
(such as unsymmetrically halogen-substituted
compounds, phenols, and carboxylic acids) may require
additional techniques for extraction from water. The
choice of solvent is also critical to the extraction
efficiency. Methanol would be expected to extract a
larger quantity of volatile organic material from soils or
sediments than from water. For inorganic analyses, the
matrix may require additional acidification to dissolve
metal salts that have precipitated from the solution.
Sample preparation procedures for organic analy tes are
applied based on volatility. Volatile organics are
analyzed using head-space orpurge and trap techniques.
Extraction alternatives for the analysis of less volatile
(extractable) organic chemicals include separatory
funnels, Soxhlet extraction apparatus, continuous liquid-
liquid extractors, and solid phase cartridges. Details of
these extraction options can be obtained from the project
chemist. Strengths and weaknesses of each of these
preparation procedures are described in Exhibit 36.
For inorganic analyses, the sample matrix is usually
digested in concentrated acid. The released metals are
introduced into the instrument, then analyzed by flame
AA or ICP atomic emission spectrophotometry. The
selection of the acid fordigestion influences the detection
limit because different acids have different digestion
abilities.
• If digestion is not used, the sample measurement
corresponds to a determination of soluble metals
rather than total metals. If soluble metals have a
greater lexicological significance, this difference
may be important to the risk assessment.
• If the sample is filtered in the field or the laboratory
before digestion, any metals associated with
particulates are removed before analysis. If
particulates are an exposure pathway in the risk
assessment, sample filteration would
underestimate risk.
The analytical requestmust specify if the sample is to be
filtered and whether or not it is to be digested (to
measure soluble metals). Unless otherwise specified,
samples are usually digested but not filtered.
3.2.7 Identification of Exposure
Pathways
Exposure pathways and their components, such as
source, mechanism of release, etc., should be designated
prior to the design of the sampling procedures. For the
risk assessment, at least one broad spectrum analytical
sample is required and two or three are recommended
for each medium and potential source in an exposure
pathway. If the site sampling design fails to consider all
exposure pathways and media, additional samples will
be required.
Current and future exposure pathways may be limited to
particular areas of a site. If sampling activity can be
concentrated in these areas, the precision and accuracy
of the datasupporting risk assessments can be unproved.
Risk assessment requires characterization of each
exposure area for the site. Samples not falling within
the areas of potential concern are not used in the
identification of chemicals of potential concern nor in
the calculation of reasonable maximum exposure
concentration. Depending on exposure pathways, the
risk assessor may utilize only asmall number of samples
that were collected at a site. Exhibit 37 shows why the
identification of exposure pathways is critical to the
sampling design in order to maximize the number of
samples that are useable in the risk assessment.
3.2.8 Use of Judgmental or
Purposive Sampling Design
Judgmental or purposive designs that specify sampling
points based on existing site knowledge may be
appropriate for the initial phase of site sampling or when
the risk assessment is performed using few samples. In
such instances, non-statistical approaches may be more
effective in accomplishing the purpose of the risk
assessment for human health, than statistical designs
with unacceptably large sampling variability.
Judgmental samples can be incorporated intoastatistical
design if the samples designate the area of suspected
contamination as an exposure area or stratum. The
judgmental samples are then selected randomly or within
a grid in the area of known contamination. Under the
procedures described, the initial judgmental samples
are not considered biased for the exposure area. Exhibit
38 summarizes some strengths and weaknesses of biased
and unbiased sampling designs.
Resource constraints sometimes restrict the number of
samples for the risk assessment and therefore potentially
increase the variability associated with the results. When
the number of samples that can be taken is restricted,
judgmental sampling may identify the chemicals of
potential concern, but cannot estimate the uncertainty
of chemical quantities. The reasonable maximum
exposure or upper confidence limit cannot be calculated
from results of a judgmental design. Bias can be
avoided with the procedures described in the previous
paragraph.
55
-------
EXHIBIT 36. COMPARISON OF SAMPLE PREPARATION OPTIONS
Fraction
* Matrix
Prop* ration
Strengths
Weaknesses
Volatile
Soil/Water
Head-space
Purge and Trap
ExtractaMe
Organic*
in Water
Separately
Funnel
Continuous
Extraction
SoidPhase
Extraction
Rapid, simple, potentially automated and
minimal interferences if standards are
prepared using sample media to minimize
the effects of ionic strength variability
between samples and standards.
General/ recommended lor this analysis
(comparabilities): can be automated;
broady apoicabte and allows concentration
(actor good recoveries across analyte 1st.
High precision and recoveries lor waters.
Relatively rapid processing and low set-up
costs; relatively high PAH recovery.
Minimal matrix probtoms; generally higher
analytical precision and high phenol
recoveries; overal high extraction
efficiency (accuracy).
Very rapid, simple technique; sample* can
be extracted in the field for laboratory
analysis; potentialy tow MOL in a dean
Qualitative identification; comparison of
concentration possfcle but quantitative
standardization is difficult, especially true
for complex matrix (e.g., particulales and
day in sol); no mechanism for
concentration; appicatkxi and sensitivity
are very afiatyte-specinc.
Sacrifice of either highly votatle analytes or
inadequate purge of low volatility analytes;
dependent on purge and trap parameters.
Sots have variable response dependent on
soil characteristics. Efficiency of soil purge
is not monitored.
Genera*/ low recovery ol target anaiytes;
high potential for matrix problems; poor
method precision.
Lower recovery of PAH and phthalates
(especially higher molecular weight);
time-consuming procedure and high initial
set-up costs; more potential for
contamination.
Procedure has limited available performance
Presence of interference and matrix
n affect extraction efficiency
ExtractaWe
Organic* in
Soil
Inorganics
Sonication
SoxhM
Extraction
Acid Digestion
0.45 urn
I lamli»aii- —
ivwrmrww
Firation
Direct Aspiration
Rapid sample preparation; relatively tow
solvent requirement good efficiency of
analyte recovery/matrix exposure to
solvent
Relatively routine requirement for dract
analytical support: relatively good
exposure of sample to solvent if sample
texture appropriate; relatively tow initial
cost
and data quaMy. Each batch of extraction
medwm must be tested for efficiency by
recovery of standards, preferably in the
same matrix BreaUhrough (loss) occurs at
high sa/rple concentrations.
Labor intensive; constant attention to
procedure; relatively high initial cost.
Methytone chloride/acetone solvent mixture
result! in many condensation products and
often in method blank contamination.
Relatively high operating cost-replacement
apparatus; solvent: for some matrices may
not provide efficient sample/solvent contact
(e.g., enameling, very slow sample output).
Dta
• provides resukc for
Somt
total metals.
Isolates Dissolved metals species.
No preparation required; provides
for Dissolved metals.
»mpounds are ackll insoluble;
Digestion may promote interference effects.
FMratbn problems in nekt does not provide
a total metass assay; is an extra step in
sample eofection.
Partfcuta
affect)
56
-------
EXHIBIT 37. IDENTIFICATION OF EXPOSURE PATHWAYS PRIOR TO
SAMPLING DESIGN IS CRITICAL TO RISK ASSESSMENT
E)
camples of sampling design missing exposure areas ol concern:
Systematic Gnd:
x x.-' .-'x x
t •
x x.-' .-'x x x
« t
* 9
X.-'' .-'X X X
» *
f *
X.-'' ,-'X XXX
# »
(A)
Random:
' V *
/ * x
* *
• •
X .•' / X X
/x/
0 *
* *
• *
• «
« *
* *
* •
* *
/X/' X
« _#
(A)
No samples
tor exposure
pathway A
and
five for B
(B)
No samples
for exposure
pathways
and
three for A
(B)
3.2.9 Field Analyses Versus Fixed
Laboratory Analyses
Held analyses are typically used to gather preliminary
information to reduce errors associated with spatial
heterogeneity, or to prepare preliminary maps to guide
further sampling. Held analyses are often conducted
during the RI to provide data to determine worker
protection levels, the extent of contamination, well
screen casing depths, and the presence of underground
contamination, and to locate hot spots. For many sites,
field analyses can often provide useful data for risk
assessment The analyses provide semi-quantitative
results, often free of significantmatnx interference, that
can be used quantitatively if confirmed by a quantitative
analysis from fixed laboratories.
Held instruments are usually divided into three classes:
field portable instruments that can be carried by a single
person, field transportable instruments that can be moved
and used in the field or in a mobile laboratory, and
mobile laboratory instruments that are installed in a
trailer for transport to a site. Instrumentation used may
be GC, X-ray fluorescence (XRF), or organic vapor
analyzer (OVA). Examples and applications of these
instruments might include on-site GC analysis of soil
gas to indicate the presence of underground
contamination, XRF for soil lead analyses, and the
OVA to detect volatile organics, reported in benzene
equivalents rather than in standard units of concentration.
Analytical methods that have traditionally been restricted
to off-site laboratories can now be employed in the field.
In addition, the quality of field.instrumentation has
unproved steadily, allowing for better measurements at
the site. Rugged versions of fixed laboratory
instrumentation, such as XRF and GCs, can often be
performed in trailers if adequate ventilation and power
supplies are available. With field analyses, greater
numbers of samples can be analyzed with immediate, or
very short, holding times with no shipping and storage
requirements. 'At least 10% of field analyses should be
confirmed by fixed laboratory analyses to ensure
comparability.
*• Field methods can produce legally
defensible data if appropriate method QC is
available and if documentation is adequate.
57
-------
EXHIBIT 38. STRENGTHS AND
WEAKNESSES OF BIASED AND UNBIASED
SAMPLING DESIGNS
Sampling
Design
Biased
(judgmental,
purposive)
Unbiased
(random,
systematic
grid,
geostatislical)
Strengths
• Uses knowledge of
location
• Fewer resources
• Timeliness
• Focuses sampling
effort
• Abiity to calculate
uncertainty
• Ability to determine
upper confidence
limit
• Representativeness
• Reduces probability
of false negative
Weaknesses
• Inability to calculate
uncertainty
• Inability to determine
upper confidence
limit
• Decreases
representativeness
• Increases
probability of false
negatives
• Resource intensive
• May require
statistician
• Timeliness
• More samples
required
Significant QA oversight of field analyses is
recommended to enable the data to be widely used.
Fieldanalysis performance data are often notavailable—
in part becauseof the variety of equipmentandoperating
environments, variety of sample matrices, and relative
"newness" of certain technologies. Therefore, an in-
field method validation program is recommended.
Spikes and performance evaluation materials should be
incorporated, if available in addition to other standard
QC measures such as blanks, calibration standards, and
duplicates.
The precision and accuracy of individual measurements
may be lower in the field than at fixed laboratories, but
the quicker turnaround and the possibility of analyzing
a larger number of samples may compensate for this
factor. A final consideration is the qualifications of
operators in the field. The RPM, in consultation with
chemists and quality assurance personnel, should set
proficiency levels required for each instrument class
and decide whether proposed instrument operators
comply with these specifications.
Fixed laboratory analyses are particularly useful for
conducting broad spectrum analyses for target
compounds, to avoid the possibility of false negatives.
They generally provide more information for a wider
range of analyies than field analyses, and are generally
more reliable than field screening or field analytical
techniques.
<•" To minimize the potential for false neg-
atives, obtain data from a broad spectrum
analysis from each medium and exposure
pathway.
Fixed laboratory analysis commonly uses mass
spectrometry for organic analyses, which provides
gready enhanced abilities for compound identification.
For inorganics, AA spectroscopy or ICP atomic emission
spectroscopy should be used for reliable identification
of target analytes. Once the broad spectrum analysis
and contaminant identification Idas occurred, other
methods may be employed that offer lower detection
limits, better quanutate specific iinalytes of concern,
and that may be less expensive.
<*• The CLPor otherfixedlaboratory sources
are most appropriate for broad spectrum
analysis or for confirmatory analysis.
Characteristics such as turnaround time, detection and
identification ability of the instruments, precision and
accuracy requirements of the measurements, and
operator qualifications should be considered when
selecting field or fixed laboratory instrumentation.
Exhibit 39 compares the characteristics of field and
fixed laboratory analyses. The risk assessor and RPM
should consult the project chemist to consider the
available options and make a choice of analysis based
on method parameters, turnaround time, and cost, as
well as other data requirements pertinent to risk
assessment needs (e.g., legal defensibility). Exhibit 40
compares the strengths and weaknesses of field and
fixed laboratory analyses.
3.2.10 Laboratory Performance
Problems
The RPM should be aware of problems that occur
during laboratory analyses, even though the resolution
of such problems are usually handled by the project
chemist. This section discusses common performance
problems and explains how to differentiate laboratory
performance problems from method performance
problems.
<•*• Solicit the advice of the chemist to en-
sure proper laboratory selection and to
minimize laboratory and/or methods
performance problems that occur in sample
analysis.
58
-------
EXHIBIT 39. CHARACTERISTICS OF FIELD AND
FIXED LABORATORY ANALYSES
Characteristic
Prevention of
false negatives
Prevention of
false positives
Analytical
Turnaround Time
Sample
Preparation
Field Analysis
Immediate analysis
means volatiles not lost
due to shipment and
storage.
No sample to sample
contamination during
shipment and storage.
Data available
immediately or in up to
24 to 48 hours
(additional time
necessary for data
review).
Limited ability to prepare
samples prior to
analysis.
Fixed Laboratory
Analysis
More extensive sample
preparation available to
increase recovery of
analytes.
Contamination by
laboratory solvents
minimized by storage
away from analytical
system.
Data available in 7 to 35
days unless quick
turnaround time
requested (at increased
cost).
Samples can be
extracted or digested,
thereby increasing the
range of analyses
available.
Laboratory performance problems may occur for routine
or non-routine analytical services and can happen with
the most technically experienced and responsive
laboratories. Laboratory problems include instrument
problems and down-time, personnel inexperience or
insufficient training, and overload of samples. Issues
that may appear to be laboratory problems, although
they are actually planning problems, include inadequate
access to standards, unclear requirements in the analytical
specifications, difficulty in implementing non-routine
methods, and some sample-related problems. Another
problem for the RPM may be a lack of laboratories with
appropriate experience or available capacity to meet
analytical needs. These problems can usually be a verted
by "up-front" planning and by a detailed description of
required analytical specifications.
• Instrumentproblemscanberevealedwithaunique
identifier for each instrument in the laboratory that
is reported with the analyses. Calibration and
21-002-039
performance standards, such as calibration check
standards, internal standards, or system monitoring
compounds, should be specified in the analytical
method to monitor performance of each instrument.
In addition, the use of instrument blanks should be
specified (to avoid the possibility of carry-over
during the analysis).
Some degradation in data quality may appear
when new personnel are operating or when the
sample load for a laboratory is high. The contrib-
uting personnel for each analysis should be
identifiedcleariyin laboratory records and reports,
disqualifications ofpersonnel required in contracts
should be documented.
Sample and method problems can often be
distinguished from laboratory problems if they are
not associated with aspecific instrument or analyst.
A review of method QC data should distinguish
between laboratory and sample problems.
59
-------
EXHIBIT 40. STRENGTHS AND WEAKNESSES OF FIELD
AND FIXED LABORATORY ANALYSES
Analysis*
Strengths
Weakness**
Field -Portable XRF
(Metals)
Extremely high volume sampling and analysis;
compatible with sophisticated sampling and
data handling software. Detection limit may be
above laboratory instrument values but
applicable to specific site levels of interest
Confirmation technique recommended.
Comparability may require external
standardization of calibration because
quantitation is based on soil surface area
versus a soil volume. Results often lower
than from AA analyses.
Field GC
Rapid analysis supporting high volume sampling
for variety of volatile and extractabte organic
target compounds (includes pesticides/RGBs).
Minimization of sample handling variability and
data quality indicators comparable to fixed
laboratory methods.
Requires prior site knowledge to ensure
applicability to specific conditions (e.g.,
soil-gas may not be appropriate for
investigation in sandy area). Confidence
in identification is matrix- and site-specific
and highly variable depending on sample
complexity. Confirmation technique
recommended.
Mobile Laboratory
XRF. AA (Metals)
Combines the high volume sample capacity of
field analyses with the detection limits, data
quality and confidence associated with
laboratory analyses.
Requires significant resources, time,
and personnel to transport, maintain
and operate; generally most appropriate at
high volume sites, especially remote.
Mobile Laboratory
Luminescence
Rapid survey of analytes that routinely
require sample preparation (e.g., PAHs and PCBs).
Detection limits can be adjusted within limits to
site-specific concentrations of concern.
Technique has had minimal use in EPA
site investigation. Comparability may
be an issue and require extensive
confirmatory analyses.
Mobile Laboratory
GC, GC-MS
Combines high volume capacity of field
analyses with increased confidence in
identification (GC-MS) or improved data
quality (GC). GC methods may be identical
to laboratory procedures but quality is
intermediate due to site conditions (e.g.,
temperature, humidity and power requirements).
Same weaknesses as for mobile
laboratory inorganics. An additional
weakness is the increased! training
requirements and decreased availability
of experienced GC-MS operators for
totally independent system operation.
Possibiity of site contamination and
cross-contamination.
Fixed Laboratory
XRF, AA, ICP
(Metals - Available
Routine Methods)
Highest comparability and representativeness.
Data quality, including detection limits,
generally predictable. Efficient match of analyses
required to instrument (e.g., multiple analyses
run simultaneously by ICP).
Slow delivery of data; increased
documentation requirement due to
the number of participants -relatively
high sample cost
Fixed Laboratory
GC & GC-MS
(Organfcs - Available
Routine Methods)
Highest comparability and representativeness.
Necessary confirmation of qualitative
identification. Data quality and detection
limits generally predictable. In depth
analysis and sample archives for follow-up
testing.
Same weaknesses as for fixed
laboratory metals; analyte-specific
performance.
ICP = Inductively Coupled Plasma Spectroscopy. Graphite AA = Graphite Furnace (electrothermal) Atomic Absorption
Spectroscopy. Flame AA = Flame Atomic Absorption Spectroscopy. ICP-MS - Inductively Coupled Plasma-Mass
Spectroscopy. XRF = X-Ray Fluorescence. GC = Gas Chromatography. GC-MS * Gas Chromatography-Mass
Spectrometry. AA = Atomic Absorption Spectroscopy.
n-ora-oto
60
-------
EXHIBIT 40. STRENGTHS AND WEAKNESSES OF FIELD
AND FIXED LABORATORY ANALYSES
(Cont'd)
Analysis*
Strengths
Weaknesses
ICP
Simple, automated, extremely rapid; can assay
metals simultaneously; can detect ppb levels.
Subject to salt or iron interferences; lacks
detection capability at low levels; not
suitable for less than 20 ppb Arsenic, Lead,
Selenium, Thallium, Cadmium, Antimony;
requires background and interelement
correction.
Graphite AA
Simple, automated; can assay most metals; can
assay low level metals; can detect ppb levels.
Lower precision and accuracy result unless
methods of standard additions used.
Method is time-consuming; requires
background correction; requires matrix
modifiers; subject to spectral interferences.
Graphite tube requires replacement
frequently.
Flame AA
Simple, rapid, very suitable for high concentration
sodium and potassium assays; commonly used and
rugged.
Not as sensitive as graphite AA; salts can
interfere; limited by lamp capabilities;
detects ppm levels.
ICP-MS
Rapid; can detect low levels; accurate.
Method is subject to isobaric molecular and
ion interferences. Nebulization, transport
process, and memory physical
interferences occur. Method is relatively
new and is expensive. Specialized training
is required.
ICP-Hydride
Rapid; can detect low levels of Antimony, Arsenic,
Selenium; Hydride formation eliminates spectral
interferences.
Dependent on analyte oxidation state;
especially sensitive to copper interference.
Method is relatively new. Specialized
training is required.
ICP m Inductively Coupled Plasma Spectroscopy. Graphite AA * Graphite Furnace (electrothermal) Atomic Absorption
Spectroscopy. Flame AA • Flame Atomic Absorption Spectroscopy. ICP-MS • Inductively Coupled Plasma-Mass
Spectroscopy. XRF * X-Ray Fluorescence. GC = Gas Chromatography. GC-MS = Gas Chromatography-Mass
Spectrometry. AA * Atomic Absorption Spectroscopy.
21-002-04O-01
61
-------
Chapter 4
Steps for Planning for the Acquisition of Useable
Environmental Data in Baseline Risk Assessments
This chapter provides planning guidance to the RPM
and risk assessor for designing an effective sampling
plan and selecting suitable analytical methods to collect
environmental analytical data for use in baseline risk
assessments. It is important to understand that the
variances inherent in both sampling and analytical
designs combine to contribute to the overall level of
uncertainty. 11>e chapter also provides a number of
charts and worksheets that should be useful in planning.
It is important to remember that these are provided for
guidance only. Each Region, or the staff at an individual
site, may modify these for their use or develop their own
materials.
The chapter has two sections. The first section of the
chapter describes the process of selecting a sampling
design strategy and developing a sampling plan to
resolve the four fundamental risk assessment decisions
presented in Chapter 2:
• What contamination is present and at what levels?
• Are site concentrations sufficiently different from
background?
• Are all exposure pathways and exposure areas
identified and examined?
• Are all exposure areas fully characterized?
A Sampling Design Selection Worksheet and a Soil
Depth Sampling Worksheet are used as data collection
and decision-making tools in this process. Guidance for
evaluating alternative sampling strategies and designing
statistical sampling plans is included.
The second section of the chapter provides guidance on
selecting the methods for analyzing samples collected
during the RI. A Method Selection Worksheet is used
to compile the list of chemicals of potential concern and
todetermineanalytical priorities so that the mostsuitable
combination of methods is selected.
The risk assessor or RPM, in consultation with other
technical experts, will probably complete several
worksheets, representing different media, exposure
pathways, potential sampling strategies, chemicals of
potential concern, and analytical priorities. This is done
to compile sufficient information to communicate basic
risk assessment requirements to the RPM, and to ensure
that these requirements are addressed in the sampling
and analysis plan (SAP).
The selection of sampling plans and analytical methods
should be based on the performance measures discussed
in this chapter. These measures are assessed by data
quality indicators that quantify attainment of the data
quality objectives (DQOs) developed by the RPM for
the total data collection and evaluation effort.
4.1 STRATEGIES FOR DESIGNING
SAMPLING PLANS
This section provides guidance for evaluating alternative
sampling strategies. Risk assessment may involve
sampling many media at a site: groundwater, surface
water, soil, sediment, industrial sludge, mine tailings, or
air. The strategies for sampling different media often
vary. For example, random stratified sampling may be
the appropriate method for examination of soils at a site,
but the positioning of groundwater monitoring wells is
seldom done on a random basis. Sampling designs for
soils and sediments are usually created to examine
spatial distribution and heterogeneity of chemicals of
concern. Groundwater sampling plans examine the
Acronyms
AA atomic absorption
BNA base/neutral/acid
CAS Chemical Abstracts Service
CLP Contract Laboratory Program
CV coefficient of variation
CVAA cold vapor atomic absorption
DQO data quality objective
EMMI Environmental Monitoring Methods Index
EMSL-LV Environmental Monitoring Systems
Laboratory - Las Vegas
EPA U.S. Environmental Protection Agency
GC gas chromatography
GFAA graphite furnace atomic absorption
GIS Geographic Information System
GPC gel permeation chromatography
ICP inductively coupled plasma
MDL method detection limit
MORD minimum detectable relative difference
MS mass spectrometry
PA/SI primary assessment/site inspection
PCB polychlorinated bipbenyl
QA quality assurance
QC quality control
RAS routine analytical services
RI remedial investigation
RME reasonable maximum exposure
RPM remedial project manager
SAP sampling and analysis plan
VOA volatile organics
XRF X-ray fluorescence
63
-------
extent of a plume containing the chemical of concern,
and also often examine seasonal or temporal variability
in chemical concentrations. Exhibit 41 summarizes the
relative variation in spatial and temporal properties for
different types of measurement.
The terms stratum and strata are used frequently in this
section. A stratum is usually a physically defined layer
or area; it can also be a conceptual grouping of data or
site characteristics that is used in statistical analysis.
Sampling guidance in this section is focused on
determining the spatial extent and variability of the
concentration of chemicals of potential concern.
Therefore, itapplies most directly to soils and sediments.
Some EPA Regions have developed sampling guidances
for groundwater, and the RPM and risk assessor should
consult these whenever available.
Examples of common sampling designs are given in
Exhibit 42, and their overall applicability is shown in
Exhibit 43. Schematic examples of some of the designs
are illustrated in Exhibit 44.
The objective of the sampling plan is to determine a
strategy that collects data representative of site
conditions. The data must have acceptable levels of
precision and accuracy, obtain minimum required levels
of detection for chemicals of potential concern, and
have acceptable probabilities of false positives and false
negatives. Meeting these objectives involves optimizing
the confidence in concentration estimates and the ability
to detect differences between site and background levels.
To accomplish these objectives, the RPM can optimize
the number of samples, the sampling design, or the
efficiency of statistical estimators (e.g., mean, standard
deviation, and standard error).
Increasing the number of samples may increase initial
costs, depending on whether fixed or field analytical
methods are used for analysis, but it is necessary in
EXHIBIT 41. EXAMPLES OF SPATIALLY AND
TEMPORALLY DEPENDENT VARIABLES
Relative Variation in Measurements
Attributable to:
Measurement
Geophysical Measurements
Soil-Gas Measurements
Weather/Air Quality
Surface Water Quality
Usually Small
Usually Large
Physical Soil Properties
Soil Moisture
Aquifer Properties
Groundwater Flow
Usually Large
Usually Small
Concentration of Groundwater
Contaminants
21-002-0*1
64
-------
EXHIBIT 42. EXAMPLES OF
SAMPLING DESIGNS
Design
Judgmental/
Purposive
Classical Random
Classical Stratified:
Random
Systematic
Cluster
Composite
Systematic:
Random
Grid
Search
Surrogate
Phased
Geo statistical
Examples of Application
Monitoring Wells
Hot Spots
Background Soil
Drums at Surface
Waste Piles
Soil from Boreholes
Soil from Test Pits
Determine Concentrations of
Chemicals of Potential
Concern in Soil
Concentrations of Chemicals
of Potential Concern. Surface
Soil Characteristics
Contaminant Hot Spots
Gas Detector Measurements
Extent of Contamination
Distribution of Contamination
certain situations (see Section 4.1.2). The sampling
design can often be improved by stratifying within a
medium to reduce variability, or by selecting a different
sampling approach, such as a geostatistical procedure
termed "kriging." Improving the efficiency of the
statistical estimators involves specifying the type of
data distribution if parametric procedures are being
used, or switching from nonparametric to parametric
procedures if distributional assumptions can be made.
Exhibit 45 is a Sampling Design Selection Worksheet,
structured to assist design selection for the most complex
environmental situation, which is usually soil sampling.
The worksheet contains the elements needed to support
the decisions for RI sampling design to meet data
requirements for risk assessment The RPM and risk
assessor may use this worksheet or use it as a model to
create one specifically suited to their needs. The final
site sampling plan must meet the data useability
requirements of risk assessment The final procedure
for sampling design should be selected based on the
specific reason for sampling (e.g., defining a boundary
or obtaining an average over some surface or volume).
The worksheet should be completed for each medium
and exposure pathway at the site. Once completed, this
initial set of worksheets can be modified to assess
alternative sampling strategies. Completion of a set of
worksheets (i.e., a worksheet for each medium and
exposure pathway at a site, based on a single sampling
strategy) specifies the total number of samples to be
taken for an exposure pathway, and sample breakdown
according to type (i.e., field samples, quality control
samples, and background samples).
The remainder of this section is a step-by-step guide to
completing the Sampling Design Selection Worksheet.
Chemicals of potential concern listed on the Sampling
Design Selection Worksheet should be the same as
those used for the Method Selection Worksheet (Exhibit
52).
4.1.1 Completing the Sampling
Design Selection Worksheet
«•• Use of the Sampling Design Selection
Worksheet will help the RPM or statistician
determine an appropriate sampling design.
Pathway, medium and design alternatives. Sampling
procedures used in environmental sampling are either
unbiased or biased. Classical and geostatistical models
are unbiased in terms of sample evaluation and
hypothesis testing. The classical model is based on
random, or stratified random procedures, and the
geostatistical model on optimizing co-variance.
Systematic grid sampling can be utilized by either the
classical or geostatisticalmodel. Biased, or judgmental/
purposive, designrequires the use of different approaches
to planning and evaluation.
<•• While other designs may be appropriate
in many cases, stratified random or
systematic sampling designs are always
acceptable.
• Classical model: The classical model uses either
a random or stratified random sampling design. It
is appropriate for use in sampling any medium to
define the representative concentration value over
the exposure area. It is not subject to judgmental
biases, and produces known estimates and
recognized statistical measures and guidelines. A
stratified random design provides the RPM and
risk assessor with great flexibility. If the nature
and extent of the exposure areas are not yet well
defined, a pilot random study can be conducted
and the results included in the final design. The
data can be averaged for any exposure area. The
classical model is the basis for calculating
65
-------
EXHIBIT 43. APPLICABILITY OF SAMPLING DESIGNS
Objective of Sampling
Estimate
Chemical
Concentration
Distribution
Evaluate
Trends
Identify
Hot Spots
Judgmental/
Purposive
Classical Random
Classical Stratified
Random
Systematic
Maybe
Maybe
Systematic:
Random
Grid
Geostatistical
confidence levels, power, andminimum detectable
relative differences (MDRDs).
Geostatistical model: Geostatistical techniques
are good for identifying hot spots and can be used
for calculating reasonable maximum exposure
(RME). These techniques require complex
judgmental or purposive calculation procedures.
Even with die use of available computer programs,
a statistician should be consulted because different
21-002-043
approaches to estimating key parameters can
produce different estimates.
Systematic grid sampling: Systematic grid
sampling procedures are good for identifying
unknown hot spots and also provide unbiased
estimates of chemical occurrence andconcentration
(Gilbert 1987) useful in calculating the RME.
Systematic sampling can be used in geostaustical
or classical estimation models. Variance
66
-------
EXHIBIT 44. COMMON SAMPLING DESIGNS
Simple Random
Sampling
Cluster
Sampling
Clusters
Stratified Random
Sampling
Strata
Stratified Systematic
Sampling
Strata
Systematic Grid
Sampling
Systematic Random
Sampling
21-002-044
67
-------
g
0)
UJ
Q
O
o
CD
0)
Exposure
U)
,« co
Q. 0)
3 .£
uV
Offi
UJZ
OC 03
r
3
Csj
8
OUJ
CD
Q.
£
g
I
UJ
O
UJ
if
X
UJ
DC
UJ
in
CD
I
X
UJ
68
-------
EXHIBIT 45. PARTI: MEDIUM SAMPLING SUMMARY
SAMPLING DESIGN SELECTION WORKSHEET
(Cont'd)
A. Site Name
C. Medium: Groundwater, Soil, Sediment. Surface Water, Air
Other (Specify)
D. Comments:
B. Base Map Code.
E. Medium/
Pathway
Code
Exposure Pathway/
Exposure Area Name
Column Totals:
F. Number of Samples from Part II
Judgmental/
Purposive
Back-
ground
Statistical
Design
Geo-
metrical
orGeo-
statistical
Design
QC
G: Grand Total:
Row
Total
21-OONMSO1
69
-------
EXHIBIT 45. PART II: EXPOSURE PATHWAY SUMMARY
SAMPLING DESIGN SELECTION WORKSHEET
(Cont'd)
H.
Chemical of Potential Concern
and CAS Number
1.
Frequency
ol
Occurrence
J. Estimation
Arithmetic
Mean
Maximum
K.
CV
L.
Background
M. Code (CAS Number) of Chemical of Potential Concern Selected as Proxy
N. Reason for Defining New Stratum or Domain (Circle one)
1. Heterogeneous Chemical Distribution
2. Geological Stratum Controls
3. Historical Information Indicates Difference
4. Field Screening Indicates Difference
5. Exposure Variations
6. Other (specify)
O. Stratum or Exposure Area
Name and Code
P.
Reason
Q. Number of Samples from Part III
Judgmental/
Purposive
R. Total (Part I, Step F):
Back-
ground
Statistical
Design
Geo-
metrical
orGeo-
statistical
Design
QC
Row
Total
21-002-OtXa
70
-------
EXHIBIT 45. PART III: EXPOSURE AREA SUMMARY
SAMPLING DESIGN SELECTION WORKSHEET
(Cont'd)
O. Stratum or Exposure Area
E. MediunVPathway Code
S. Judgmental or Purposive Sampling
Comments:
Domain Code _
Pathway Code.
Use prior site information to place samples, or determine location and extent of contamination. Judgmental or
purposive samples generally cannot be used to replace statistically located samples.
An exposure area and stratum MUST be sampled by at least TWO samples.
Number of Samples
T. Background Samples
Background samples must be taken for each medium relevant to each stratum/area. Zero background samples
are not acceptable. See the discussion on page pp. 74-75.
Number of Background Samples
U. Statistical Samples
CV of proxy or chemical of potential concern
Minimum Detectable Relative Difference (MDRD)
Confidence Level (>80%) Power of Test
Number of Samples
(See formula in Appendix IV)
(<40% if no other information exists)
V. Geometrical Samples
Hot spot radius
. (Enter distance units).
Probability of hot spot prior to investigation
Probability that NO hot spot exists after investigation
(see formula in Appendix IV)
W. Geostatistical Samples
Required number of samples to complete grid +
Number of short range samples
.(0 to 100%)
(enter only if >75%)
X. Quality Control Samples
Number of Duplicates
Number of Blanks
Y. Sample Total for Stratum
(Part II, Step U)
(Minimum 1:20 environmental samples)
(Minimum 1 per medium per day or 1 per sampling
process, whichever is greater)
Judgmental/
Purposive
Back-
ground
Statis-
tical
Design
Geo-
metrical
or Geo-
statistical
QC
Row
Total
21-002-045-03
71
-------
calculations required toestimate confidence limits
on the average concentration are available (Caulcutt
1983). Systematic sampling is powerful for
complete site or exposure area characterization
when the exposure area is known to be
heterogeneous.
Determining number of samples. Four factors need to
be considered in determining the total number of samples
required (see Exhibit 46):
• Exposure areas,
• Statistical performance objectives (based on site
environmental samples),
• Quality assurance objectives (based on QC
samples), and
• Background samples (based on MDRD).
variation for a chemical of potential concern and
measures of performance is the basis for determining
die number of samples necessary to provide useable
data for risk assessment.
<•• If the natural variability of the chemicals
of potential concern is large (e.g., greater
than 30%), the major planning effort should
be to collect more environmental samples.
The number of samples can be calculated given a
coefficient of variation, a required confidence level or
certainty, a required statistical power, and an MDRD.
Exhibit 47 illustrates the relationships between the
number of samples required given typical values for the
coefficient of variation and statistical performance
objectives. Calculation formulas in Appendix IV
facilitate the examination of effects beyond the examples
cited.
EXHIBIT 46. FACTORS IN DETERMINING
TOTAL NUMBER OF SAMPLES COLLECTED
Number of Exposure Ar
(P. 74)
i That will be Sampled
• Media within exposure area
• Strata within exposure area medium
Number of Sample* for Each Exposure Area
Grouping Given Required Statistical Performance
• Confidence(1-a),wriereaistheprababiityola
type I error
• Power (1-P). where pis the probability of a type II error
• Minimum detectable relative difference
Number of Quality Control Samples (p. 76)
• Field duplicate (cotocaled)
• ReW duplicate (spit)
• Blank (trip, field, and equipment (rinsate))
• Reid evaluation
Number of Background Samples (p. 74)
• Number of site samples collected
• Minimum detectable relative difference
The number of environmental site samples is ultimately
controlled by performance requirements, given the
statistical sampling design. The relationship between
number of samples andmeasures of performance depends
upon the variability of the chemicals of potential concern,
which is measured by the coefficient of variation. In
other words, the relationship between the coefficient of
4.1.2 Guidance for Completing the
Sampling Design Selection
Worksheet
This section provides step-by-step instructions for
completing the Sampling Design Selection Worksheet
shown in Exhibit 45.
Part I: Medium Sampling Summary
A. Enter the Superfund site name.
B. Enter a code that uniquely identifies a base map of
the site or the exposure unit.
All sampling events should be identified on a map
or in a database such as a Geographical Information
System (CIS).
C. Identify the medium to be sampled (e.g., soil,
groundwater, industrial sludge, mine tailings,
smelter slag, etc.).
D. Enter any comments required to describe the
exposure area, and other information such as the
RPM's name.
E. Enter a medium/pathway code that has been
assigned for the risk investigation.
F. Specify the exposure pathway (e.g., ingestion of
soil).
Leave this entry blank for now, then enter the
number of samples for each category that have
been selected from Part II (Step R) of the worksheet
when completed.
72
-------
EXHIBIT 47. RELATIONSHIPS BETWEEN MEASURES OF STATISTICAL
PERFORMANCE AND NUMBER
OF SAMPLES REQUIRED
Samples Required to Meet
Minimum Detectable
Coefficient
of Variation (%) Power (%)
10
15
20
25
30
35
Note:
95
95
95
95
95
95
Number of samples required
minimum detectable relative
Confidence Relatlve Difference
Level (%)
90
90
90
90
90
90
in a one-sided
5%
36
78
138
216
310
421
10%
10
21
36
55
78
106
20%
3
6
10
15
21
28
one-sample t-test to achieve a
difference at confidence level and
power. CV
based
on geometric mean for transformed data.
Source:
EPA1989C.
Sample types are broken out by sample type:
• Judgmental/Purposive,
• Background,
• Statistical design (e.g., stratified random
sampling),
• Geometrical or geostatistical design (including
hot spot sampling), and
• Quality control samples.
»• At least one broad spectrum analytical
sample is required for risk assessment, and
a minimum of two or three are
recommended for each medium in an
exposure pathway.
G. Enter the grand total of all samples within a specific
medium.
21-002447
Partll: Exposure Pathway Summary
H. List the chemicals of potential concern and their
CAS numbers.
List the known or suspected chemicals of potential
concern based on historical data. This will generally
be from the PA/SI.
I. List the frequency of occurrence (%).
The frequency of occurrence is the percent of
samples in which the chemical of potential concern
has been identified. This may be obtained from
site-specific data or calculated from historial (PA/
SI) data or fate and transport modeling.
J. Enter an estimate of the average (arithmetic mean)
and maximum concentration of the chemical of
potential concern.
Historical data or data from similar sites can be
used to derive these values. More sampling will
usually be necessary to determine statistically
73
-------
significant differences if these values are close to
background levels or to the levels of detection.
K. Estimate the coefficient of variation.
The coefficient of variation (CV) can be estimated
from site-specific data or from data from similar
sites. The number of samples necessary to produce
useable data will generally increase as the CV
increases. The definition of separate strata or
domains should be investigated if a CV is above
50%. Exhibit 23 contains a listing of historical
values for CVs that may be used as an estimate in
the absence of site-specific data.
L. Estimate background concentration.
Background concentration estimates should be for
each medium relevant to each strata/area. Site-
specific data are preferred, but data from similar
sites can be utilized.
M. Select a proxy chemical of potential concern.
Choose a proxy from the list of chemicals of
potential concern to develop sampling plans. Note
that a proxy that has the highest CV, lowest
frequency of occurrence, or whose concentration at
the site is closest to background levels will require
the most samples.
N. Developthereasonfordefiningnewstrataorareas.
• Heterogeneous Chemical Distribution: If a
chemical can be shown to have dissimilar
distributions of concentration in different
areas, then the areas should be subdivided.
For example, hot spots may be considered
separately.
• Geological Stratum Controls: Knowledge of
local geologic conditions can be used to
produce separate areas where similar statistical
distributions are likely to exist In particular,
different "stratigraphic" layers may produce
distinct strata.
• Historical Information: Historical information
on production, discharge or storage of
chemicals of potential concern can be used to
identify separate areas.
• Field Screening: Field analytical results can
be used to locate sub-populations that are
mapped into exposure areas.
• Exposure Variations: Information or
variations in behavior patterns, land use or
receptor groups can be used to identify separate
areas.
• Other reasons can be used to produce separate
sampling areas, such as observed stress on
vegetation, oily appearance of soils, or the
existence of refuse, etc.
O. List the stratum or area name and code.
The stratum or area identifies sub-areas on the site
base-map.
P. Annotate reason from Step N.
Q. List the number of samples estimated after
completing Part III of this worksheet.
R. List the number of samples estimated after
completing Part II and Pan in of this worksheet.
Part HI: Exposure Area Summary
S. Enter judgmental/purposive sampling comments.
A minimum of three to fivejudgmentalorpurposive
samples must be used to sample a stratum or
exposure area. Historical or prior site information
can be used to locate sampling positions to determine
the extent and magnitude of contamination.
Chemical field screening, geophysics, vegetation
stress, remote sensing, geology, etc. can also be
used to guide judgmental sampling. Judgmental or
purposive samples are not recommended for
estimating average and maximum values within a
stratum or domain area, but they can be used in
geostatistical kriging estimations and can be
included in calculating risk.
T. Identify background samples.
For statistical purposes, a sufficient number of
background samples must be taken to determine
the validity of the null hypothesis that there is no
difference between mean values of concentration
in the site and the background samples at the
desired level of confidence. Early sampling and
analysis of background samples will indicate the
ease with which background levels can be
discriminated, and allow modifications to be made
to the SAP if necessary.
Background samples must be taken for each
exposure pathway. As with QC samples, results
from the background sample should be assessed
early to see if background levels will severely
impact the sampling design. The number of
necessary background samples increases as the
variability of the background values increases.
Background samples should not be used in the
estimation of average or maximum values within a
stratum or exposure area, but they can be used in
74
-------
kriging estimations. In those instances where
background levels are close to on-site contamination
levels, it may be necessary to collect as many
background samples as site samples. Small numbers
of background samples increase the probability of
a type II, false negative error (i.e., that no difference
exists between site and" background when a
difference does, in fact, exist). However, rigorous
statistical analyses involving background samples
may be unnecessary if site and non-site related
contamination clearly differ.
•• Collect and analyze background samples
prior to the final determination of the
sampling design since the number of
samples is significantly reduced if little
background contamination is present.
Background levels of contaminants vary by medium
and the type of contamination. If a detectable
background level of a contaminant occurs
infrequently, the number of background samples
analyzed might be kept small. Metals often nave
high rates of detection in background samples.
Some pesticides, such as DDT, are anthropogenic
and also have high rates of detection in particular
matrices. Anthropogenic background levels are
also found in sites near industries and urban areas.
It is important to distinguish detection, or lack of
detection, in a single sample from a false positive
orfalsenegativeresulL Results from single samples
are different estimators than those from statistical
parameters from pooled samples. Background
sampling must be increased in the following
situations:
• Contamination exists in more than one
medium,
• Expectedcoefficientsof variation in chemicals
of concern are high and confirmed by actual
data,
• Relative differences between site and
background levels are small, and
• Site concentrations and concentrations of
concern are low.
U. Identify statistical samples.
Samples should be systematically or randomly
located. The number of samples can be calculated
using the CV of the proxy variable, the required
MDRD, the required confidence level and power of
the test, and the appropriate statistical formula and
appropriate charts.
For example, using the equation in Appendix IV:
Where Za and Z, are obtained from the normal
distribution tables for significance levels a
and 6 respectively; a is the probability of the
false positive error rate, and B is the probability
of the false negative error rate.
Then, if a is 0.2 (20%) and the confidence
level is 80% then Za is 0.842. If B is 0.05 (5%)
then the power is 95% and Z, is 1.648.
If the MDRD is 20% and the CV is 30%, then
D = MDRD which equals 0.666
CV
and n>15 samples are required.
V. Identify samples from geometrical design.
*• Systematic sampling supplemented by
judgmental sampling is the best strategy
for identifying hot spots.
For example, using the equation in Appendix IV:
Where R = 20m
and A = 37,160m2
and X = 0.3 Probability that a hot spot is in the
exposure area from "historical
records" or from field screening or
geophysical tests.
and C = 0.2 The acceptable "walk away"
probability that a hot spot exists
after a sampling grid has been
done.
then:
D = 2.7, R = 54.8m, and
n = 27,160/54.82 =12.37
Therefore 12 samples are required.
Note that the requirements for 15 samples from a
statistical sampling approach can be met in this
example if the hot spot search is augmented by
randomly locating two additional samples. The
results for number of samples from U and V are not
additive.
W. Identify samples from geostatistical design.
A geostatistical sampling pattern should be designed
at the early stage of planning. A statistician should
be consulted to develop the design.
75
-------
X. Quality Control Samples
Generally, duplicates should be taken at a minimum
of 1 duplicate for every 20 environmental samples
(EPA 19890. However, this frequency may be
modified based on site conditions. For example,
the number of duplicates and other QC samples
may be set high for the beginning of site sampling,
evaluated after several duplicates to determine
routine measurement error, and subsequently
adjusted according to observed performance. The
information in Exhibit 48 shows that confidence in
measurement error increases sharply when four or
more pairs of duplicate samples are taken per
medium. Critical samples are recommended for
designation as duplicates in the Q A sampling design.
EXHIBIT 48. NUMBER OF SAMPLES REQUIRED
TO ACHIEVE GIVEN LEVELS OF CONFIDENCE,
POWER, AND MDRD 1
Con*dMK»(1-a)
80%
•0%,
90%
80%
80%,
80%
PoMrp-e) MORO No. of Sample*
90%
oo%2
00%
80%
80%,
80%
1 VakiM tar nucnbtr of unpta «r» bn*d
•n-. oafMunat (80%) «nd PCTMT (80%).
Source EPA IBMc.
10%
20%
20%
10%
20%
40%
on«CVof25%.
42
12
8
10
5
3
CAMBMnWlt
Y.
Blanks provide an estimate of bias due to
contamination introduced by sampling,
transportation, carryover during field filtration,
preservation, or storage. At least one field blank
per medium should be collected each day, and at
least one blank must be collected for each sampling
process (EPA 19890-
Examine results from duplicate and blank samples
as early as possible in the sampling operation to
ascertain if presumed sampling characteristics are
accurate and discover areas where the sampling
strategy requires modification. For a more detailed
discussion of the types and use of QC samples see
A Rationale for the Assessment of Errors in the
Sampling of Soils (EPA 1990c).
Calculate the sample total for stratum or exposure
area (enter in Pan n, Step U).
4.1.3 Specific Sampling Issues
Selection of performance measures. Quantitative
data quality indicators based on performance objectives
should be proposed for completeness, comparability,
representativeness, precision, and accuracy during
planning. Performance measures are specified as
minimum limits for each stratum. Based on the
coefficients of variation of the analyte concentrations,
these limits will determine the numbers of samples
required. The actual valuesorobjectivesaredetermined
by the level of acceptable uncertainty, which includes
that associated with hot spot identification.
Recommended minimum criteria are: specified in Exhibit
48 for statistical performance measures associated with
the uncertainty in risk assessment: confidence level,
power, and MDRD. Recommended minimum criteria
for measurement error and completeness for critical
samples are discussed in the following sections.
Setting minimum acceptable limits for confidence
level, power, and minimum detectable relative
difference. Confidence level, power, and MDRD are
three measures of sampling design precision. These
measures are ultimately determined by the coefficient
of variation of chemical concentration and the number
of samples. Each measure is briefly defined as follows:
• Confidence level: The confidence level is 100
minus a, where a is the percent probability of
taking action when no action is required (false
positive).
• Power Power is 100 minus fi, where B is the
percent probability of not taking action when
action is required (false negative).
• Minimum detectable relative difference: MDRD
is the percent difference required between site and
background concentration levels before the
difference can be detected statistically.
The power and ability to detect differences between site
concentration levels compared to background levels are
critical for risk assessment. Given a CV, the required
levels of confidence, power, and MDRD significantly
affect the number of samples. Exhibit 48 illustrates the
effect when the CV is equal to 25%
It is important to note that the number of samples
required to meet confidence and power requirements
will be low if the acceptable MDRD is large; that is, if
site contamination is easily discriminated from
background levels.
Determining required precision of measurement
error. Held duplicates and blanks are the major field
QC samples of importance to the precision of
measurement error. Duplicates provide an estimate of
76
-------
total measurement error variance, including variance
due to sample collection, preparation, analysis, and data
processing. They do not discriminate between-batch
error variance. If the duplicate is collocated, contaminant
sample variation caused by a heterogeneous medium is
also included in the measure. The precision of the
measurement error estimate is subject to the number of
duplicates on which the estimate is based. Exhibit 49
gives the estimated precision of the measurement error
based on the number of duplicate pairs. With three
duplicates, the true measurement error variance could
be as much as 13.89 times the observed variance, if a
95% level of confidence is required. The resources
needed for the collection and analysis of duplicates
depend on the magnitude and variability of the
concentration of concern for the chemicals of potential
concern.
• Little room for measurement error exists if the
level of concentration of concern is near the method
detection limit, and the precision of the estimate of
• measurement error is critical.
• If the natural variability of the chemicals of
potential concern is relatively large, the major
planning effort will be to collect more samples
from the exposure areas, rather than collecting
more QC samples. More detailed discussions of
the use of QC measures and selection of the
appropriate number of QC samples may be found
in A Rationale for the Assessment of Errors in the
Sampling of Soils (EPA 1990c).
Planning for 100% completeness for critical samples.
Certain samples in a sampling plan may be designated
by the RPM or risk assessor as critical in determining
the potential risk for an exposure area. For example, if
only one background sample is taken fora gi venmedium
and exposure area, then that sample would be considered
EXHIBIT 49. CONFIDENCE LEVELS FOR THE
ASSESSMENT OF MEASUREMENT VARIABILITY
Number of Interval for 95% Confidence that Measurement Error is Within Limits ^
Duplicate
Pair Samples Observed
Variance (s2)
2
3
4
5
6
7
8
9
10
15
20
25
50
100
27
.32
.36
.39
.42
.44
.46
.47
.49
.54
.56
.62
.70
.77
£
S.
£
£
£
£
£
£
£
£
£
£
£
£
True
Variance
a2
a2
a2
2
0
a2
2
o
2
a
2
a
2
a
a2
a2
2
0
a2
a2
£
£
£
£
£
£
£
£
£
£
£
£
£
£
Observed
Variance (s2)
39.21
13.89
8.26
6.02
4.84
4.14
3.67
3.33
3.08
2.40
2.08
1.91
1.61
1.35
1
1
2
s « Observed variance (precision of an estimate).
2
o* * True variance (population variance).
Note:
Source:
Assumes data are or have been transformed to normal distribution.
EPA 1990c.
.
77
-------
"critical." All data associated with such a sample must
he complete. The only acceptable level of completeness
for critical samples is 100%.
"• Focus planning efforts on maximizing
the collection of useable data from critical
samples.
Hot spots and the probability of missing a hot spot.
Hot spots are primarily an issue in soil sampling. The
RPM and risk assessor must determine whether hot
spots exist in the exposure area and the probable size of
the hot spot. This information can often be deduced
from historical data and assisted by judgmental sampling,
although judgmental sampling alone cannot produce
estimates of the probability that a hot spot has been
missed. Procedures for determining the probability of
missing a hot spot are not as effective in random designs
as in systematic and geostatistical designs. However, a
search strategy which stratifies the area based on grids
and then randomly samples within each grid can be used
within the classical technique. Systematic and
geostatistical design approaches provide the best
approach to unknown hot spot identification.
Appendix IV describes numerical procedures and
assumptions to determine the probability that a given
systematic design will detect a hot spot and provides a
calculation formula based on a geometrical approach.
To employ this formula, the distance between grid
points and the estimated size of the hot spot as a radius
must be specified.
Historical data comparability. The RPM may wish to
assess historical data along with current results or may
anticipate that the current data will need to be compared
with results from future sampling activities. Consult a
statistician in either of these cases to determine if the
current sampling design will allow the production of
dataofknown comparability. Factorsother than statistics
may need to be considered when attempting to combine
data from different sampling episodes. Physical
properties of the site such as weather patterns, rainfall
and geologic characteristics of different exposure areas
may need to be considered. Temporal effects, such as
the seasonality or time period of sampling, or seasonal
heightofawatertable,inayalsobe important. Analytical
methods have been modified over time and many
required detection limits have been revised.
f The ability to combine data from different
sampling episodes or different sampling
procedures is a very important consideration
in selecting a sampling design but should
be done with caution.
4.1.4 Soil Depth Issues
The appropriate depth or depths to take soil samples can
be a major issue in determining a sampling design.
Exhibit 50 is a worksheet designed to help the RPM and
risk assessor to determine an appropriate soil sampling
depth. The conceptual site model (Exhibit 6) provides
the basis for completing this worksheet. The nature and
depth of soil horizons at the site should be established
wherever possible. Features such as porosity, humic
content, clay content, pH, and aerobic status often affect
the movement or fate of chemicals of potential concern
through a soil. As with other worksheets provided in
this guidance, this worksheet is intended as a guide or
basis for development. RPMs, in consultation with the
risk assessor and other staff, can revise or modify this
worksheet as appropriate to the site. Consider both
current and future land use scenarios in soil exposure
areas because of the sorpti ve and retentive properties of
soils.
Completing the Soil Depth Sampling Worksheet
1. Land Use Alternatives
A. Identify current or future land use.
B. Identify exposure scenario.
The exposure scenario should be identified for
currentor future land use. Identify the scenario
according to Role of Baseline Risk Assessment
in SuperfundRemedy Selection Decision (EPA
l99lc)andHumanHealthEvaluationManual
Supplemental Guidance: Standard Default
Exposure Factors(EPAl991d). Aresidential
exposure scenario should be used whenever
there are, or may be, occupied residences on or
adjacent to the site. Unoccupied sites should
be assumed to be residential in the future
unless residential land use is unreasonable.
Sites that ate surroundedby operating industrial
facilities can be assumed to remain as industrial
areas unless there is an indication that this
assumption is not appropriate. Other potential
land uses, such as recreation and agricultural,
may be used if appropriate.
2. Chemicals of Potential Concern
A. Specify class of chemical.
Circle the classes of chemicals of potential
concern (e.g., volatile organics (VOAs),
semi volatile organics (semi- VOAs), inorganics
or metals, or special class) that apply.
78
-------
Ill
UJ
X
O)
*
oc
i
o
a.
<
Q.
UJ
Q
J
O
0)
o
to
m
x
UJ
*
|
5
8
e
I
to
•
•
•
^
a
S
to
Zs
*3
S3
86
E <
1 1
•5
B (check one)
. Residential
. CommercliMndui
Other (Specify)
T- 1 1 1
CO
f £
C *0
O -'
a e
S |
(0 =
81
f
i
^l°i 1
sifl 5
9-3 — ,,
>2§§H -
«ii5l 1
f|£! 8
co Q. a «
1 =
" CO
•
c
o
i
j
8-
Q
|
to
&
!-Ji» II -I
JiiiiM 1 1
|o|S|og|^|$ U |
§ 8 1 -1 1 « I S."S * • J > t o
fi S «•££ co jj i=Oa(Sso *• i a
a< a. f<«oo a a
CO CO CO CO
CO
•g
a
c
o
a
(A
3
^
£
O
J1"
T3 ^.
s§
S «
Is
c»
> TD
II
£ «
32
fl
.2S
tJ-5
2S
ets are necessary
it areas of a site, [
9 5
11
8=5
I5
1-®
II
s 'C
II
of a site determ
idential and con
*i
|o
o d>
fi
«
79
-------
B. Record physical properties.
Circle the physical properties of the chemicals
of potential concern that apply. These
properties can be estimated from factors such
as the octanol/water partition coefficient,
Henry's law constant, and water solubility
appropriate to each chemical.
3. Soil Characteristics
A. Record the taxonomic designation of the soil,
if known.
B. Record the organic matter content of the soil.
C. Record the most common particle size of the
soil.
D. Identify any concern for migration of the
chemicals of potential concern to other media
(e.g., air, sediment, surface water, and
groundwater).
4. Vegetative Cover
Circle whether the vegetative cover of the site is
heavy, sparse or intermittent.
5. Other Factors
List other factors or considerations that influence
the desired depth of soil sampling. For example,
geological factors (e.g., depth to groundwater or
bedrock) could influence soil sampling.
6. Expected Depth of Contamination by Chemicals
of Potential Concern
Enter expected depth (and units) of contamination
by chemicals of potential concern, given the
chemicals, soil characteristics and vegetative cover.
Depth can be influenced by disposal practices or
deposition patterns, soil characteristics, vegetative
cover, and physical and chemical properties of the
chemicals of potential concern.
7. Exposure Pathways
Enter exposure pathways by chemicals of potential
concern, soil characteristics and vegetative cover.
Physical and chemical properties of the chemicals
of potential concern will influence their activity in
theexposure path way (e.g., VOAs and theinhalauon
pathway). Soil characteristics and vegetativecover
will also influence the exposure pathway (e.g.,
groundwater and water ingestion pathway).
8. Representative Sample Depths
Record representative sample depths (including
units) indicated by the data completed in Steps 2
through 7.
Basic Soil Depth Definitions
Surface dust is the top 0 to 2 indies of soil that can
be carried by the wind and tracked into houses.
Surface soil is the top 0 to 6 inches of soil. If the
surface is grass covered, surface soil is considered
the 2 inches below the grass layer.
Subsurface soil can typically range from 6 inches
to 6 or more feet in soil depth. For example, at sites
with potential soil moving activity, soil depths
greater than 6 feet could be of concern in risk
assessment.
Other Performance Measures. Other performance
measures may be designated to facilitate the monitoring
and assessment of sampling. For example, field spikes
and field evaluation or audit samples can be used to
assess the accuracy and comparability of results. Field
matrix spikes are routine samples spiked with the
contaminant of interest in the field and do not increase
the number of field samples. Field evaluation samples
are of known concentration, which are introduced in the
field at the earliest stage possible and subject to the same
manipulation as routine samples. Field evaluation
samples will increase the total number of samples
collected. Performance measures for field spikes and
evaluation samples are expressed in terms of percent
recovery. Difficulties associated ivith field spiking,
especially in soil, have resulted in limited use of this
practice (EPA 1989f).
4.1.5 Balancing Issues for Decision-
Making
Completing a number of Sampling Design Selection
Worksheets (Exhibit 45) for different exposure areas,
media, and sampling design alternatives will enable the
RPMandrisk assessor to compare and evaluate sampling
design options and consequences and select the
appropriate sampling design for each medium and
exposure pathway. Practical tradeoffs between response
time, analytical costs, number of samples, sampling
costs, and level of uncertainty can then be weighed. For
example, perhaps more samples can be collected if less
expensive analyses are used. Or, if the risk assessment
is based on a point source, collection of additional
samples to estimate chemical concentrations and
distribution can be avoided.
80
-------
Computer programs are useful tools in developing and
evaluating sampling strategies, especially in trading off
costs against uncertainty, and identifying situations
when additional samples will not significantly affect the
useability of the data (i.e., the point of diminishing
returns). Each automated system has specific data
requirements and is based on specific site assumptions.
The major systems that support environmental sampling
decisions are listed, contacts for information given, and
brief descriptions provided in Exhibit 51.
4.1.6 Documenting Sampling Design
Decisions
It is important to document the primary issues considered
in balancing tradeoff to accommodate resource concerns
and their impact on data useability. Fully document all
final sampling design decisions, including the rationale
for each decision. During the course of the RI, continue
to document pertinent issues that arise and any sampling
plan modifications which are implemented.
4.2 STRATEGY FOR SELECTING
ANALYTICAL METHODS
This section describes how to use the Method Selection
Worksheet shown in Exhibit 52 as a data collection and
decision-making tool to guide the selection of analytical
methods that meet the needs of the risk assessment and
to select the most appropriate method for each analyte.
The RPM and risk assessor should consult the project
chemist and use this worksheet in method selection.
Alternatively, it can be a model to create a worksheet
specifically suited to their needs. Methods selected in
this process may be routine or non-routine.
EXHIBIT 51. AUTOMATED SYSTEMS* TO SUPPORT
ENVIRONMENTAL SAMPLING
System
Data Quality Objective
(Training) - Expert
System
ESES
Environmental Samping
(Plan Design) -Expert
System
GEOEAS
Geostatistical
Environmental
Assessment Software
SCOUT
Multivariate Statistical
Analysis Package
ASSESS
EPA Contact
Dean Neptune
USEPA
Quality Assurance
Management Staff
(202)260-9464
JertVanEe
Exposure Assessment Div.
USEPA. EMSL-LV
(702) 798-2367
Evan Engtund
Exposure Assessment Div.
USEPA, EMSL-LV
(702) 798-2248
Jeff Van Ee
Exposure Assessment Div.
USEPA. EMSL-LV
(702) 79S-2367
Jeff Van Ee
. Exposure Assessment Div.
USEPA, EMSL-LV
(702) 798-2367
recommended.
Description
Training system designed to assist in
planning of environmental
Expert system designed to assist in
planning sample cotaction. Includes
models that address statistical design.
QC. sampling procedures, sample
handling, budget, and documentation.
Current system addresses metal
contaminants in a sol matrix. (Expanded
EMSL-LV.)
Collection of software tools for
of spatially OBtrfcuted data points.
Programs include fie management,
contour mapping, kriging, and variogram
analysis.
A collection of statistical programs that
accept GEOEAS files for muMvariate
analysis.
System designed to assist in
assessment of error in sampling of soils.
Estimates measurement error variance
components. Presents scatter plots of
QC data and error plots to assist in
determining the appropiiate amount of
QC samples.
nimumof640KRAM. Afixeddfekis
81
-------
I
I
I
I
ce
Ul
IU
x
(O
^
cc
o
o
u
ui
UJ
u>
o
o
UJ
CM
in
CO
X
X
UI
I
eg
i o e
o _ « r
d g 5 g
•o
i
1
i
1
=s
o>
e
10
I
•o
CD
O
if
s = -^
ill
82
-------
«• Ensure that critical requirements and
priorities are specified on the Method
Selection Worksheet so that the most
appropriate methods can be considered.
• Routine methods are issued by an organization
with appropriate responsibility (e.g., state or
federal agency with regulatory responsibility,
professional organization), are validated,
documented, and published, and contain
information on minimum performance
characteristics such as detection limit, precision
and accuracy, and useful range.
• Non-routine methods address situations with
unusual or problematic matrices, low detection
limits or new parameters, procedures or
techniques; they often contain adjustments to
routine methods.
*• Use routine methods wherever possible
since method development is time-
consuming and may result in problems with
laboratory implementation.
4.2.1 Completing the Method
Selection Worksheet
1. Identify analytes.
List the chemicals of potential concern to risk
assessment for the site on the Method Selection
Worksheet Use the same list of chemicals that
appears on the Sampling Design Selection
Worksheets. Under Column IB, indicate whether
theconcentrationforeachanaryte should be reported
separately, or the total for the compound class
reported.
2. Identify medium for analysis.
Specify the analysis medium (e.g., soil, sediment,
groundwater, surface water, air, biota).
3. Decide on critical parameters.
Specify the required data turnaround time (IIIA) as
the number of hours or days from the time of
sample collection. Indicate whether chemical
identification alone is desired or identification plus
quantitation (HIE). Specify the concentration of
concern (LUC) and required detection or quantitation
limit (IHD).
4. Identify routine available methods.
Use the final worksheet column, in consultation
wim the projea chemist, to Ust the methods available
that satisfy the requirements in the preceding steps.
Reference sources and software are available to
assist in identifying routine analytical methods
applicable for en vironmental samples (Exhibit 53).
The most common routine methods for organics
and inorganics analyses for risk assessment are
listed in Appendix HI. The methods in the appendix
are from the following sources:
• Contract Laboratory Program (CLP)
Statements of Work for Routine Analytical
Services (EPA 1990d, EPA 1990e),
Test Methods for Evaluating Solid Waste
(SW846): Physical/Chemical Methods (EPA
1986b),
• Standard Methods for the Examination of
Water and Wastewater (Clesceri, et. al., eds.
1989), and
• EPA Series 200, 300, 500, 600 and 1600
Methods (EPA 1983, EPA 1984, EPA 1988d,
and EPA 1989g).
Other sources of methods are:
• Field Analytical Support Project (FASP) (EPA
1989h),
Field Screening Methods Catalog (EPA
1987b),
• Field Analytical Methods Catalog,
• ERTStandard Operating Guidelines,
• Close Support Analytical Methods,
• A CompendiumofSuperfundField Operations
Methods (EPA 1987c),
• Association of Official Analytical Chemists
(AOAC), and
• American Society for Testing and Materials
(ASTM).
Several computer-assisted search and artificial
intelligence-based tools are available, including the
Environmental Monitoring Methods Index (EMMI),
the Smart Methods Index, andacomputerized reference
book on analytical methods. Some of these systems are
designed as teaching tools, as well as informational
compendia. All offer the ability to rapidly search and
compare lists of chemicals and method characteristics
from accepted reference sources. Exhibit 53 lists
software products that aid method selection, identifies
contacts for information, and gives a short description
of the product.
83
-------
EXHIBIT 53. AUTOMATED SYSTEMS*
TO SUPPORT METHOD SELECTION
System
Environmental
Monitoring
Methods Index
(EMMI)
Smart Method!
Index
Geophysical
Techniques
Expert System
EPASwnptng
and Analysis
DataBase
Contsct
W. A. TelHerd
USEPA
Olios of Water
(202)260-7120
John Nooerino
Quality Assurance Oiv.
USEPA. EMSL-LV
(702) 798-21 10
Aldo Maooels
Advanced Monitoring
Ofv.
USEPA. EMSL-LV
(702)796-2254
Lewis Pubishera
1-WXM72-7737
Description
An automated sorting and
selection software package that
currently contains over 900
methods and over 2600
analytes trom more than 80
regulating and non-fegulatng
lists. These are cross-
referenced to facilitate selection
based on required needs (e.g.,
anatyle detection in*,
Instrument).
Natural language expert system
prototype that provides
interactive queries of databases
crass-referenced by method,
snalyle, and performance
features.
An sxpert system mat suggests
and ranks gsophysKal
techniques, including soil-gas, for
appkcabillly of use based on
site-specific characteristics.
A three-volume set of diskettes
and a printed manual provides
a search of sampling and
from a menu-driven program of
ISO EPA-appreved methods.
The database can be searched
by method, analyte. matrix, and
various QA considerations.
*AI systems mil mn on any IOM compatible PC AT with • minimum of S40K RAM.
A fixed drak Is recommended.
4.2.2 Evaluating the Appropriate-
ness of Routine Methods
•r Analyte-specific methods that provide
better quantitation can be considered for
use once chemicals of potential concern
have been identified by a broad spectrum
analysis.
Choice of the proper method is critical to the acquisition
of useable data. See Section 32 for a more detailed
discussion. Routine methods provide data of known
quality for the analysis of chemicals and sample types
described in the method. Data quality issues (precision,
accuracy, and interferences) are usually described in the
method. Consult the project chemist and examine
available methods with respect to the criteria defined on
the Method Selection Worksheet. It may be helpful to
divide the analyte list into categories based on the types
of analysis. For example, a requirement for chromium,
cadmium, and arsenic data couldnot be generated by the
same analysis as data for chlorinated hydrocarbons
because of sample extraction and treatment procedures.
It may be possible to use several methods independently
and combine the data sets for risk assessment purposes.
This is done routinely by the CLP, where inorganics
(elemental analysis), volatiles, extractable organics,
and pesticides are analyzed by different methods. In
some cases, no routine method or series of mernods will
be able to satisfy all criteria and compromises must be
considered. The RPM, with the advice of tie risk
assessor, must men determine which criteria are of
highest priority and which can be modified. Forexample,
if a low detection limit is of high priority, turnaround
time and cost of analysis will likely increase.
Alternatively, low detection limit and precision
requirements may need to be modified if an initial broad
spectrum analysis is ofhigh priority toquicklydetermine
the largest number of chemicals present at the site.
Turnaround time. Turnaround time is determined by
the available instrumentation, sample capacity, and
methods requirements. Turnaround times for field
analyses can be as short as a few hours, while those for
fixed laboratory analyses include transport time and
range from several days to several weeks. Field
instruments can provide the quickest results, especially
if the data do not go through a formal review process.
However, the confidence in chemical identification,
and particularly quantitation, may not be as high. In
general, methods with quick turnaround times may be
less precise and have higher detection limits. If data are
needed quickly, a field method can be used for initial
results and a fixed laboratory method used to produce
more detailed results (or confirm the earlier results),
thereby increasing the confidence in Held analyses.
Sample quantitation limits. Risk assessment often
requires a sample quantitation limit at or below the
detection limit for routine methods for many chemicals
of lexicological concern (see Section 3.2.4). The sample
quantitation limits vary according to the size, treatment,
and analysis of each individual sample. The quantitation
limits for chemicals in water samples are often far lower
than for the same chemicals in soils because of co-
extractable components in the soil. Interferences known
for the method may hinder acquisition of data of
acceptable quality and are more pronounced near the
method detection limit Compare documented method
interferences with site conditions to identify potential
method problems. Some common sources of interference
in organic and inorganic analyses are summarized in
Exhibits 54 and 55. If needed sample quantitation
limits cannot be met by available methods, consult the
project chemist for the feasibility of detection at the
desired level in the required sample type. The chemist
can help determine if method adaptation can resolve the
problem, or if a non-routine method of analysis can be
used.
Usefulrangc. Theusefulrangeofametbodistherange
of concentration of chemicals for which precise and
accurate results can be generated. This range is analyte-
specific. The lower end of the useful range is the
method detection limit, often generically referred to as
84
-------
EXHIBIT 54. COMMON LABORATORY CONTAMINANTS AND
INTERFERENCES BY ORGANIC ANALYTE
Contamination
or
Interference
Fat/Oil
Sulfur
Phthalate
Esters
Laboratory
Solvents
Fraction
Extractable
organics,
pesticides, and
PCBs
Extractable organics,
chlorinated and
phosphorus-
containing pesticides
Chlorinated
pesticides, PCBs,
and extractable
organics
Volatile organics
(methylene chloride,
acetone, and
2-butanone)
Matrix
Tissue,
waste,
soils
Sediment,
waste,
soils
All
All
Effects on
Analysis
Increased
detection limit,
decreased
precision/
accuracy
Presence/
absence,
detection limits,
precision/
accuracy
False positive
identification
(pesticides and
extractable
organics) or
positive bias
(pesticides and
extractable
organics)
False positive
identification or
positive bias
Removal /
Action
GPC (all groups), florisil
(pesticides), acid
digestion (PCBs only)
GPC, copper,
mercury, tetrabutyl I
ammonium sulfate 1
Florisil. GC-MS I
confirmation of identity I
(pesticides, PCBs),
evaluation of reagents
and method blanks for
contamination
Confidence in data use
based on interpretation
of blank data
'Source: EPA 1986a.
the "detection limit." If a lower detection limit is
required, use of a larger sample or smaller final extract
volume can sometimes compensate. However, any
interfering chemicals are also concentrated, thereby
producing greater interference effects. Above the useful
range, the response may not be linear and may affect
quantitation. This causes inaccurate and/or imprecise
measurements. Reducing the sample size for analysis
or diluting the extracted material may bring the
concentration within the useful range. With individual
environmental samples, some chemicals are sometimes
present at the low end of the useful range of the method,
while others are above the useful range. In this situation,
two analyses, at different effective dilutions, are
necessary to produce accurate and precise data on all
chemicals. If detailed criteria for performing and
21-002464
reporting such actions are not already part of the
analytical S tatementof Work, then the laboratory should
be instructed to notify the RPM if this situation occurs,
to allow for sufficient time for reanalysis within the
specified holding time. All relevant analyses should be
reported to maximize the useability of both detected and
non-detected analytes.
*• All results should be reported for samples
analyzed at more than one dilution.
Precisionandaccuracy. Routine methods of ten specify
precision and accuracy with respect to specific analytes
(chemicals) and matrices (sample media). However.be
aware that environmental samples are often difficult to
analyze because of the complexity of the matrix or the
85
-------
EXHIBIT 55. COMMON LABORATORY CONTAMINANTS AND
INTERFERENCES BY INORGANIC ANALYTE
Analyte
Arsenic
Beryllium
Cadmium
Chromium
Lead
Mercury
Selenium
Cyanide
Technique
GFAA
ICP
ICP
GFAA
ICP
GFAA
ICP
GFAA
ICP
CVAA
GFAA
ICP
CcJorimetric/
spectrophotometric
Interference
Iron, Aluminum
Aluminum
Titanium, Vanadium
None except possible
sample matrix effects
Iron
Calcium
Iron, Manganese
Sulfate
Aluminum
Sulfide, High Chloride
Iron, Aluminum
Aluminum
Acids, Sulfide,
Chlorine oxidizing
agents
Removal/
Action
Background correction
(not deuterium) (Zeeman).
If above tOOppm,
correction factor utilized.
If above lOOppm,
correction factor utilized,
Background correction
for matrix effects.
If above 100 ppm,
correction factor utilized.
Add calcium, standardize
suppression, background
correction.
If above 100 ppm,
correction factor utilized.
Lanthanum nitrate
addition as matrix
modifier, background
correction.
If above 100 ppm,
correction factor utilized.
Remove interferences with
cadmium carbonate
(removes sulfide),
potassium permanganate
(removes chloride), excess
hydroxylamine sulfate
(removes free chlorine).
Alternate wavelength for
analysis, background
correction (not deuterium]!
(Zeeman).
Above 100 ppm,
correction factor utilized.
Increase pH to > 12 in field to I
remove acids, cadmium 1
carbonate (removes sulfide), 1
ascorbic acid (removes free 1
chlorine). |
Key: ICP * Inductively coupled plasma. 1
GFAA - Graphite furnace atomic absorption. 1
CVAA * Cold vapor atomic absorption. |
86
-------
presence of a large number of contaminants; this usually
results in lower levels of precision and accuracy than
those cited in the method.
4.2.3 Developing Alternatives When
Routine Methods are not
Available
If routine methods are not available to suit the parameters
of interest, it is often due to one or more of the following
factors:
• The detection limit of commonly available
instrumentation has been reached, and a lower
detection limit is required for the risk assessment,
• An unusual combination of chemicals are of
potential concern,
• The sample matrix is complex, and
• The chemicals of potential concern or other
analytical parameters are unique to a particular
site.
Consult an analytical chemist for specific guidance on
the potential limitations of alternative approaches. These
may include adaptation of a routine method or use of a
non-routine method. Be aware that certain conditions,
such as extremely low detection limits for some
chemicals, may be beyond the capability of current
analytical technology. Turnaround times and costs may
also be increased.
Adaptation of routine methods. Adapting routine
methods may be a solution when routine methods will
not provide the desired data even after compromises
have been made with respect to parameters such as
turnaround time and cost. Using toe completed Method
Selection Worksheet as the starting point, work closely
with an analytical chemist to formulate suitable
modifications to the routine method. Evaluate and
document any effects on data quality that will result
from the modifications.
Within the CLP, such analyses can be obtained by
special analytical requests. Before analysis of site
samples, it is advisable to confirm a laboratory's ability
to perform the adapted method with preliminary data.
Use of non-routine methods. Existing non-routine
methods that meet criteria can be used if a routine
method cannot be adapted to provide the necessary data.
Such analyses can be found in the research literature,
usually catalogued by analyte or instrument. On-line
computerized search services can be of considerable
help in identifying such methods. Work interactively
with an analytical chemist inreviewing selected methods.
Recognize that non-routine analyses require a greater
level of capability and experience from the analytical
laboratory, and that turnaround time can be longer
because the method may need alteration during analysis
if problems develop.
Development of new methods. Developing new
methods should be the option of last resort. The RPM,
risk assessor, and project chemist should consider
recommending the development of new methods only
for chemicals of substantial potential concern that cannot
currently be analyzed at appropriate limits of detection.
Although designing a method based on data available
for a given instrument and analytes may seem
straightforward, the process is time-consuming and
expensive. Unforeseen problems can often arise when
the method is implemented in the laboratory. Problems
can occur even when laboratory personnel have superior
training and experience. Consider the following points
when requesting the development of a new method:
• If possible, select a laboratory with a recognized
reputation for performance and flexibility in a
related area. Treat laboratory personnel as partners
in the development process. This is true whether
a commercial or a government laboratory is used.
• Identify sources for authentic standards of the
chemicals in question to support method
development Computerized databases such as
the EPA EMMI (see Exhibit 53) may be useful for
such a determination.
• Be aware that turnaround time for useable data
may be long (potentially several months) because
of the likelihood of trying different approaches
before discovering an acceptable procedure.
4.2.4 Selecting Analytical Labora-
tories
In selecting a laboratory to produce analytical data for
risk assessment purposes, identify and evaluate the
following laboratory qualifications:
• Possession of appropriate instrumentation and
trained personnel to perform the required analyses,
as defined in the analytical specifications,
• Experience in performing the same or similar
analyses,
• Performance evaluation results from formal
monitoring or accreditation programs,
• Adequate laboratory capacity to perform all
analyses in the desired timeframe.
87
-------
• Intra-laboratory QC review of all generated data,
independent of the data generators, and
• Adequate laboratory protocols for metbod
performance documentation and sample security.
For non-routine analyses, the laboratory should have
highly trained personnel and instrumentation not
dedicated to production work, especially if new methods
or untested modifications are requested.
Accreditation programs monitor the level of quality of
laboratory performance within the scope of their charters.
Many of these programs periodically provide
performance evaluation samples that the laboratories
must analyze within certain limits in order to maintain
their status. Prior to laboratory selection, request that
laboratories provide information about their performance
in accreditation programs. This information can be
used for evaluation of laboratory quality, in the case of
similar matrices and anal ytes. Laboratory adherence to
standards of performance such as the Good Laboratory
Practices Standards (Annual Book of ASTM Standards)
also provides a measure of laboratory quality.
4.2.5 Writing the Analysis Request
Include the following items in the analysis request:
• A clear, complete description of the sample
preparation, extraction, and analysis procedures
including detailed performance specifications. For
adaptation of routine methods, specify the routine
method and explicitly state alterations with
applicable references.
• Documented reporting requirements.
• Laboratory access to required authentic chemical
standards.
• A mechanism for the laboratory to obtain EPA
technical assistance in implementing method
modifications or performing non-routine methods.
If the analysis request is for a non-routine method,
reference the published material with a detailed
specification of procedures and requirements prepared
by the analytical chemist who has been working with
the RPM and risk assessor. The specification must
include the frequency, acceptance criteria, and corrective
action requirements for each of the following:
• Instrument standardization, including tuning and
initial and continuing calibration,
• QC check samples such as surrogate compound
and internal standard recoveries,
• Method blank performance (permissible level of
contamination),
• Spike sample recovery requirements,
• Duplicate analysis requirements, and
• Performance evaluation or QC sample results.
Allow time for the laboratory to review the analysis
request and question any pan of the description that
seems unclear or unworkable according to its experience
with the analytes or sample matrix. Preliminary data,
such as precision and accuracy data on a subset of the
analytes, can be requested to determine if the laboratory
can implement the proposed method. Should the criteria
not be met in the preliminary analyses, the analytical
chemist should advise the laboratory on additional
method modifications to produce die required data. In
some cases, even qualitative data can be used to note the
presence of chemicals of potential concern.
In all cases, require the laboratory performing the
analyses to contact the project chemist at the first sign
ofaproblemthatmayaffectdataquality. The RPM and
the site technical team can then judge the magnitude of
the problem and determine appropriate corrective action.
4.3 BALANCING ISSUES FOR
DECISION-MAKING
Resource issues. Resource limitations are a major
reason for sampling design modification. The number
of samples required to achieve desired performance
measures may exceed resource availability. Modifying
the sampling design and the efficiency of statistical
estimators can reduce sample size and costs, and improve
overall timeliness for the risk assessment. Analytical
methods such as field analyses may also reduce cost.
Systematic and geostatistical sampling designs can
often achieve the required performance measures with
fewer samples than classical random sampling (Gilbert
1987). Pilot sampling can be used to verify initial
assumptions of the SAP, increase knowledge of
contaminant distribution, and support S APmodifications
to reduce the number of samples. Explain resource
issues and record potential design modifications in
documentation developed during planning.
Completing a number of Sampling Design Selection
Worksheets (Exhibit 45) for different exposure areas,
88
-------
media, and sampling design altemati ves will enable the
RPM and risk assessor to compare and evaluate sampling
design options and consequences and select the
appropriate sampling design for each medium and
exposure pathway.
Computer programs are useful tools in developing and
evaluating sampling strategies, especially in trading off
costs against uncertainty, and identifying situations
when additional samples will not significantly affect the
useability of the data (i.e., the point of diminishing
returns). Each automated system has specific data
requirements and is based on specific site assumptions.
The major systems that support environmental sampling
decisions are listed, contacts for information given, and
brief descriptions provided in Exhibit 51.
Documenting design decisions. It is important to
document the primary issues considered in balancing
tradeoffs to accommodate resource concerns and their
impact on data useability. Several compromises among
options are discussed in this section. Features of
analytical options available for organic and inorganic
analytes are summarized in Exhibits 56 through 59.
Fully document all final sampling and analytical design
decisions, including the rationale for each decision.
During the course of the RI, continue to document
pertinent issues that arise and any plan modifications
which are implemented.
The goal of balancing issues in toe selection of analytical
methods is to obtain the best analytical performance
without sacrificing risk assessment requirements. The
selection of analytical methods often involves tradeoffs
among the required detection limit, number of analytes
involved, precision and accuracy, turnaround time, and
cost Some choices may conflict with others.
Cost should be considered only after the mostappropriate
methods have been determined. Methods requiring
specialized instrumentation, such as high resolution
mass spectrometry, will be more expensive. Methods
for use on matrices such as soil, can be more expensive
than similar methods for a simpler matrix such as water.
Less expensive methods often have higher detection
limits and less specific confirmation of identification.
However, the turnaround times are often quicker and a
larger number of samples can be analyzed. This often
significantly increases sampling precision and reduces
the probability of missing hot spots. Less expensive
methods are often chosen if the site has already been
characterized by broad spectrum analyses. In evaluating
routine methods, consider whether analysis of more
samples through use of less expensive methods can
provide a similar level of data quality to that achieved
through the use of more expensive methods on fewer
samples. By remaining aware of the effect of individual
issues on the data quality, the RPM can determine the
optimum choices.
<•• Field analysis can be used to decrease
cost and turnaround time, providing data
from a broad spectrum analysis are
available.
In addition to turnaround time for analysis, time must
also be scheduled for data review. This will not hinder
the availability of laboratory and field data for
preliminary use if a tiered data review sequence is
incorporated.
When using the tiered approach, consider the use of split
samples (i.e., sending sample splits for analysis by field
and fixed laboratories). Quantitative comparison can
then be made between the precision and accuracy of the
field analyses and those of the fixed laboratory.
Confirmation of identification by both field and fixed
laboratories also increases data confidence and
useability. It is recommended that field methods should
be used with at least a 10% rate of confirmation or
comparison by fixed laboratory analyses.
89
-------
EXHIBIT 56. COMPARISON OF ANALYTICAL OPTIONS
FOR ORGANIC ANALYTES IN WATER
Method
MOL
Quantitative
Confidence
Timeliness
Precision &
Accuracy
FIELD SCREEN/FIELD ANALYSIS (Assumes preparation step)
GC(PCB)
GO (Pesticides)
GC (VOA)
G C (Soil Gas)
GC (BNA)
PHOTO VAC
Detector
FIXED LABORATORY
CLP RAS
VOA
BNA
Pesticides
Dioxin
CLP LOW CONG
GC
VOA
BNA
1600 SERIES
GC
VOA
BNA
Dioxin
PCDDs, PCDFs
Key: V = Method strength
500 SERIES
GC
VOA
BNA
600 SERIES
GC
VOA
BNA
SW846
GC
VOA
BNA
21-002-OSS
90
-------
EXHIBIT 57. COMPARISON OF ANALYTICAL OPTIONS
FOR ORGANIC ANALYTES IN SOIL
Precision &
Accuracy
Quantitative
Confidence
Comparability
FIXED LABORATORY
CLP RAS
VOA
BNA
Pesticides
Dioxin (2,3,7,8 TCDD)
SW846
GC
VOA
BNA
1600 SERIES
GC
VOA
BNA
Oioxin
FIELD SCREEN
GC(PCB)
GC(Pesticides)
GC(VOA)
GC(Soil Gas)
GC(BNA)
PHOTO VAC
Detector
Key: V = Method strength
91
-------
EXHIBIT 58. COMPARISON OF ANALYTICAL OPTIONS
FOR INORGANIC ANALYTES IN WATER AND SOIL
Precision &
Accuracy1 Comparability 2
Quantitative
Confidence Timeliness
FIXED LABORATORY
CLPRAS
ICP
GFAA V
Flame AA
200 Series
GFAA
AA
Key: V = Method strength
CLP inorganic water assays are more accurate and precise than soil assays.
ICP and GFAA are comparable at medium to high ppb levels. For As, Pb, Se, Tl and Sb at less than
20 ppb, GFAA is the method of choice.
ICP-MS and ICP-Hydride methods are relatively new; therefore, precision, accuracy, and comparability
estimates based on large statistical sampling are not available.
92
21-002-06142
-------
EXHIBIT 59. COMPARISON OF ANALYTICAL OPTIONS* FOR
ORGANIC AND INORGANIC ANALYTES IN AIR
Method
MDL
Quantitative
Confidence
Timeliness
Precision &
Accuracy Comparability
FIXED LABORATORY
CLP VOA
Cannister
Tenax
CLP BNA
CLP Metals
2-5 ppb
2-30 ppb
(for most)
0.00001-
0.001 ug/m3
3-10ng/m3
V
V
V
V
Key: V = Method strength
The methods described are new Statements of Work.
93
21-OCO-OW-03
-------
Chapter 5
Assessment of Environmental Data for Useability in
Baseline Risk Assessments
This chapter provides guidance tor the assessment and
interpretation of environmental data for use in baseline
human health risk assessments. Ecological risk
assessments follow a similar logic but may differ in
some details of sampling and analytical methodologies
and minimum data requirements. The discussion of
data assessment is presented as six steps that define the
assessment process for each data useability criterion.
Exhibit 60 lists the six criteria in the order that a risk
assessor would evaluate them. It also gives references
to the sections in this chapter where they are further
discussed.
EXHIBIT 60. DATA USEABIUTY
ASSESSMENT OF CRITERIA
CRITERION I
Reports to Risk
Assessor
(5.1)
I
CRITERION II
Documentation
(5.2)
CRITERION III
Data Sources
(5.3)
I
CRITERION IV
Analytical Method and
Detection Limit
(5.4)
CRITERION V
Data Review
(5.5)
CRITERION VI
Data Quality
Indicators
(5.6)
The four basic decisions to be made from data collected
in the RI are:
• What contamination is present and at what levels ?
• Are site concentrations sufficiently different from
background?
• Are all exposure pathways and exposure areas
identified and examined?
• Are all exposure areas fully characterized?
The uncertainty associated with each data useability
criterion affects the level of confidence associated with
each of these decisions.
How to conduct the data assessment The risk assessor
or RPM examines the data, documentation, and reports
for each assessment criterion (I - VI) to determine if
performance is within the limits specified in the planning
objectives. The data assessment process for each
criterion should be conducted according to the step-by-
step procedures discussed in this chapter. Minimum
requirements are listed for each criterion. Potential
effects of not meeting the minimum requirements are
also discussed and corrective action options are
presented. Exhibit 61 summarizes the major impact on
assessment if the minimum requirements associated
with each data useability criterion have not been met.
Acronyms
CLP Contract Laboratory Program
CV coefficient of variation
CRDL contract required detection limit
CRQL contract required quantitation limit
DQO data quality objective
GC gas chromatography
ICP inductively coupled plasma
MDL method detection limit
MS mass spectrometry
QA quality assurance
QC quality control
RAGS Risk Assessment Guidance for Superf und
RI remedial investigation
RME reasonable maximum exposure
RPD relative percent difference
RPM remedial project manager
SAP sampling and analysis plan
SOP standard operating procedure
SQL sample quantitation limit
95
-------
EXHIBIT 61. MINIMUM REQUIREMENTS, IMPACT IF NOT MET, AND
CORRECTIVE ACTIONS FOR DATA USEABILITY CRITERIA
Data Usability
Criterion
5.1 Reports to Risk
Assessor
5.2 Documentation
5.3 Data Sources
5.4 Analytical
Method and
Detection Limit
5.5 Data Review
5.6 Data Quality
Indkalors
Minimum
Requirement
• Site description
• Sampling design with
sample locations
• Analytical method and
detection limit
• Results on par-sample basis.
qualified lor analytical
limitations
• Sample quantitation limits and
detection limits (or non-
detects
• Field conditions for media
and environment
• Preliminary reports
• Meteorological data
• Field reports
• Sample results related to
geographic location
(chain-of-cuslody records,
SOPs. field and analytical
records)
• Analytical data results for
one sample per medium
per exposure pathway
• Broad spectrum analysis for
one sample per medium
par exposure pathway
• Reid measurements data
for meda and environment
• Routine (federally
documented) methods used
lo analyze chemicals of
potential concern in critical
samples
• Defined level of data review
for an data
• Sampling variability
quantified for each analyle
• QC samples to identify and
quantify precision and
accuracy
• Sampling and
analytical precision and
accuracy quantified
Impact on Risk
Assessment If Criterion
Not Met
• Unable to perform
quantitative risk
assessment
• Unable to assess
exposure pathways
• Unable lo identify
appropriate
concentration for
exposure areas
• Potential for false
negatives or false
positives
• Increased variability in
exposure modeling
• Unqualified precision
and accuracy
• False negatives
• Potential for false
negatives or false
positives
• Increased variability and
bias due to analytical
process, calculation
errors or transcription
errors
• Unable to quantify
confidence levels tor
uncertainty
• Potential for false
negatives or false
positives
Corrective
Action
• Request missing
information
• Perform qualitative
risk assessment
• Request locations
identified
• Resampling
• Resampling or
reanalysis for
critical samples
• Reanalysis
• Resampling or
reanalysis for critical
samples
• Documented
statements of
limitation for non-
critical samples
• Perform data
review
• Resampling for
critical samples
• Perform qualitative
risk assessment
• Perform
quantitative
risk assessment
for non-critical
samples with
documented
discussion of •
potential imitations |
96
-------
The following activities should be performed for each
assessment criterion:
• Identify or determine performance objectives and
minimum data requirements.
Quantitative or qualitative performance objectives
should be specified in the sampling and analysis
plan for all components of the acquisition of
environmental data (as discussed in Chapter 4).
The first step in assessing each criterion is to
assemble these performance objectives and note
any changes. Performance objectives should also
be compared with the minimum acceptable
requirements for data useability presented in this
chapter. These minimum requirements can be
adopted as performance objectives if objectives
were not specified. For example, the requirement
that there must be a broad spectrum analysis for at
least one sample in each medium for each exposure
area would be a performance objective, if
performance were not specified during planning.
• Determine actual performance compared to
performance objectives.
The next step in the assessment of each criterion is
to examine results to determine the performance
that was achieved for each data useability criterion.
This performance should then be compared with
the objectives established during planning. Take
particular note of performance for samples or
analyses that are critical to the baseline risk
assessment All deviations from the objectives
should be noted. In those cases where performance
was better than that required in the objective, it
may be useful for assessment of future activities to
determine if this is due to unanticipated
characteristics of the site or to superior performance
in some stage of the data acquisition. Corrective
action is the next step where performance does not
meet performance objectives for data critical to
the risk assessment.
• Determine and execute any corrective action
required.
*• Focus corrective action on maximizing
the useability of data from critical samples.
Corrective action should be taken to improve data
useability when performance fails to meet objectives
for data critical to theriskassessmenL Corrective action
options are described in Exhibit 62. These options
require communication among the risk assessor, the
RPM, and the technical team. Sensitivity analysis may
be performed by the risk assessor to estimate the effects
of not meeting performance requirements given the
certainty of the risk assessment. Corrective actions may
improve data quality and reduce uncertainty, and may
eliminate the need to qualify or reject data.
EXHIBIT 62. CORRECTIVE
ACTION OPTIONS WHEN DATA
DO NOT MEET PERFORMANCE
OBJECTIVES
• Retrieve missing information.
• Resolve technical or procedural
problems by requesting additional
explanation or clarification from the
technical team.
• Request reanalysis of sample(s)
from extract.
• Request construction and
re-interpretation of analytical results
from the laboratory or the project
chemist.
• Request additional sample
collection and analysis for site or
background characterization.
• Model potential impact on risk
assessment uncertainty using
sensitivity analysis to determine
range of effect.
• Adjust or impute data based on
approved default options and
imputation routines.
• Qualify or reject data for use in risk
assessment.
21-002-062
Using a worksheet to organize the data assessment.
The level of certainty associated with the data component
of risk assessment depends on the amount of data that
meet performance objectives. The risk assessor
determines whether the data for each performance
measure are satisfactory (data accepted), questionable
(data qualified) or unsatisfactory (data rejected). The
worksheet provided in this chapter may be used as a
guide or organizational tool.
Use the Data Useability Worksheet, Exhibit 63, to
document data assessment decisions. Record the
decision as accepted, accepted with qualification, or
rejected for use in the risk assessment for each data
97
-------
EXHIBIT 63. DATA USEABILITY WORKSHEET
Data Useability Criterion
Decision
Comments
Reports to Risk Assessor
Documentation
A. Work Plan/SAP/QAPjP
B. SOPs
C. Field and
Analytical Records
III Data Sources
A. Analytical
B. Non-analytical
IV Analytical Methods
Decision: Accept. Qualified Accept, Reject
21-002-083
98
-------
EXHIBIT 63. DATA USEABILITY WORKSHEET
(Cont'd)
Data Useability Criterion
VI
Data Quality Indicators
A. Completeness
B. Comparability
C. Representativeness
D. Precision
E. Accuracy
Decision
Sampling
Analytical
Combined
Sampling
Analytical
Combined
Sampling
Analytical
Combined
Sampling
Analytical
Combined
Sampling
Analytical
Combined
Comments
•
Decision: Accept, Qualified Accept. Reject
21-002-063-01
useability criterion. Outline the justification for each
decision in the comments section.
The remainder of this chapter explains how to assess
data using the data useability criteria. Assessment of
Criterion I involves identifying the data and
documentation required for risk assessment (Section
5.1). Assessment of Criteria n through V examines
available data and results in terms of the assessment of
data useability criteria for documentation (Section 5.2),
data sources (Section 5.3), analytical method and
detection limit (Section 5.4), and data review (Section
5.5). Criterion VI includes the assessment of sampling
and analytical performance (Section 5.6) according to
five data quality indicators: completeness,
comparability, representativeness, precision, and
accuracy.
99
-------
5.1 ASSESSMENT OF CRITERION I:
REPORTS TO RISK ASSESSOR
Minimum Requirements
• Site description.
• Sampling design with sample locations,
related to site-specific data needs and data
quality objectives.
• Analytical method and detection limit.
• Results on per-sample basis qualified for
analytical limitations.
• Sample quantitation limits and detection
limits for non-detects.
• Field conditions for media and environment.
• Preliminary reports.
• Meteorological data.
* Held reports.
Data and documentation supplied to the risk assessor
must be evaluated for completeness and appropriateness,
and to determine if any changes were made to the work
plan or the sampling and analysis plan (SAP) during the
course of the work. The SAP discusses the sampling
and analytical design and contains the quality assurance
project plan and data quality objectives (DQOs), if they
have been developed. The risk assessor should receive
preliminary and final data reports, as described in the
following sections.
5.1.1 Preliminary Reports
«• Use preliminary data as a basis for
identifying sampling oranalysis deficiencies
and taking corrective action.
Preliminary analytical data reports allow tberisk assessor
to begin assessment as soon as the sampling and analysis
effort has begun. These initial reports have three
functions:
• The risk assessor can begin to characterize the
baseline risk assessment on the basis of actual
data. Chemicals of interest will be identified and
the variability in concentration can be estimated.
• Potential problems in sampling or analysis can be
identified and the need for corrective action can be
assessed. For example, additional samples may be
required, or the method may need to be modified
because of matrix interferences.
• RI schedules are more likely to be met if the risk
assessment process can begin before the final data
reports are produced.
The major advantage of preliminary review of data by
the risk assessor is the potential for feedback and
corrective action while the RI is still in process. This
can improve the quality of data for risk assessment.
5.1.2 Final Report
w Problems in data useability due to sam-
pling usually can affect all chemicals
involved in the risk assessment; problems
due to analysis may only affect specific
chemicals.
The minimum data reports and documentation needed
to prepare the risk assessment are:
• A description of the site, including a detailed map
showing the location of each .sample, surrounding
structures, terrain features, receptor populations,
indicationsof air and water flow, and a description
of the operative industrial process (if any),
• A description and rationale for the sampling design
and sampling procedures,
• A description of the analytical methods used,
• Resultsforeachanalyteandeachsample.quah'fied
for analytical limitations, and a full description of
all deviations from SOPs, SAPs, and QA plans,
• Sample quantitation limits (SQLs) and detection
limits for undetected analytes, with an explanation
of the detection limits reported and any
qualifications,
• A narrative explanation of the level of data review
usedand the resulting dataqualifiers. The narrative
should indicate the direction of bias, based on the
assessment of the results from QC samples (e.g.,
blanks and field and laboratory spikes), and
• A description of field conditions and physical
parameter data as appropriate for the media
involved in the exposure assessment.
It may not be possible to perform a quantitative baseline
risk assessment if any of these materials are not available
and cannot be obtained. The RPM or risk assessor
should attempt to retrieve missing deliverables from the
source.
Additional reports and data that are useful to the risk
assessor, such as data results on Contract Laboratory
Program (CLP) diskettes, are listed in Exhibit 19. Access
100
-------
to this information can improve the efficiency and
quality of the risk assessment. However, not having
access does not necessarily require the data to be qualified
or rejected. Minimum requirements for reports to the
risk assessor are listed in Exhibit 61.
5.2 ASSESSMENT OF CRITERION II:
DOCUMENTATION
Minimum Requirements
• Sample results related to geographic location
(chain-of-custody records, SOPs, field and
analytical records). __
Three types of documentation must be assessed: chain-
of-custody records, SOPs, and field and analytical
records. Chain-of-custody records for risk assessment
must document the sample locations and the date of
sampling so that sample results can be related to
geographic location and specific sample containers. If a
sample result cannot be related to a sampling date and
the point of sample collection, the results are unuseable
for quantitative risk assessment. Full scale chain-of-
custody procedures (from sample collection through
analysis) are required for enforcement or cost recovery.
SOPs describe and specify the procedures to be followed
during sampling and analysis. They are QA procedures
that increase the probability that a data collection design
will be properly implemented. SOPs also increase
consistency in performing tasks and, as a result,
determine the level of systematic error and reduce the
random error associated with sampling and analysis.
Knowledge mat SOPs were developed and followed
increases confidence that the quality of data can be
determined, and the level of certainty in risk assessment
can be established. The existence of SOPs for each
process or activity involved in data collection is not a
minimum requirement, but SOPs can be useful if data
problems occur, particularly in assessing the
comparability of data sets.
Field and analytical records document the procedures
followed and the conditions of the procedures. Field
and analytical records, such as field logs and raw
instrument output, may be useful to the risk assessor as
back-up documentation, but they are not minimum
requirements. QC data from blanks, spikes, duplicates,
replicates, and standards should also be accessible, in
either raw or summary formats, to support qualitative or
quantitative assessments of the analytical results. Like
SOPs, such records are critical to resolving problems in
interpretation, but they may not directly affect the level
of certainty of the risk assessment. Minimum
requirements for documentation are listed in Exhibit
61.
5.3 ASSESSMENT OF CRITERION III:
DATA SOURCES
Minimum Requirements
• Analytical sample data results for each
medium within an exposure area.
• Broad spectrum analysis for one sample per
medium per exposure area.
• Field measurements data for media and
environment.
Data source assessment involves the evaluation and use
of historical and current analytical data Historical
analytical data should be evaluated according to data
quality indicators and not source (e.g., analytical
protocols may have changed significantly over time).
The minimum analytical data requirement for risk
assessment is that results are produced for each medium
within an exposure area using a broad spectrum analytical
technique, such as GC-MS methods for organic analytes
or ICP for inorganic analytes. The useability of data
will almost always increase as more broad spectrum
analyses are performed for each exposure area. The
absence of a broad spectrum analysis from a fixed
laboratory results in an increased probability of false
negatives; all chemicals of potential concern at the site
may not be identified. In the absence of a broad
spectrum analysis, the best corrective action is to take
additional samples. If additional samples cannot be
obtained, the probability of false negatives and false
positives should be considered high, and the level of
certainty of the risk assessment is decreased.
The broad spectrum analysis, and any other analytical
data, are subject to the basic documentation and data
review requirements discussed in this chapter. The
location of the sample data point must be known, as well
as the method and SQL achieved for analytical results.
Guidance for the assessment of analytical data to
determine false positives and false negatives and the
precision and accuracy of concentration results is
provided in Section 5.6.1.
Field measurements of ph> jical characteristics of the
site, medium, or contamination source are a critical data
source, whose omission can significantly affect the
ability of the risk assessor to perform a quantitative
assessment Physical site information is also required to
perform exposure fate and transport modeling. Examples
101
-------
of such data are particle size, pH, clay content and
porosity of soils, wind direction and speed, topography,
and percent vegetation. RAGS, Part A, Exhibit 4-2,
"Examples of Modeling Parameters for Which
Information May Need to be Obtained During a Site
Sampling Investigation," (EPA 1989a) provides a list of
data elements according to medium modeling category.
These measurements must be collected during sampling.
The use of default options and routines to estimate
missing values allows the use of the model but increases
the uncertainty associated with the exposure assessments.
5.4 ASSESSMENT OF CRITERION IV:
ANALYTICAL METHOD AND
DETECTION LIMIT
Minimum Requirements
• Routine (federally documented) methods
used to analyze chemicals of potential
concern in critical samples.
The risk assessor compares SQLs or method detection
limits (MDLs) with analyte-specific results to determine
their consequence given the concentration of concern.
Assessment of preliminary data reports provides an
opportunity to review the detection limits early and
resolve any problems. When a chemical of potential
concern is reported as not detected, the result can only
be used with confidence if thequantitation limits reported
are lower than the corresponding concentration of
concern. The minimum recommended requirement is
that the MDL be no more than 20% of the concentration
of concern, so that the SQL will also be below the
concentration of concern. Chemicals identified above
this ratio of detection limit to concentration of concern
can be used with good confidence. For example, if the
concentration of concern for arsenic in groundwater is
70 ug/L for an average daily consumption of 2 L of
water by a 70 kg adult, the detection limit of a suitable
method for examination of groundwater samples from
such a site should be no greater than 14 ug/L. Minimum
requirements for analyticalmethods anddetection limits
are listed in Exhibit 61.
If the concentration of concern is less than or equal to the
detection limit, and the chemical of concern is not
detected, do not use zero in the calculation of the
concentration term. When the MDL reported for an
analyte is near to the concentration of concern, the
confidence in both identification and quantitation may
be low. This is illustrated in Exhibit 64. Information
concerning non-detects or detections at ornear detection
limits should be qualified according to the degree of
acceptable uncertainty, as described in Section 5.6.1.
The concentration of concern for ecological risk may be
different than the concentration of concern for human
health risk. In addition, aquatic life criteria should be
examined to determine if they are based on ecological
or human health risk.
5.5 ASSESSMENT OF CRITERION V:
DATA REVIEW
Minimum Requirements
• Defined level of data review for all data.
Data review assesses the quality of analytical results
and is performed by a professional with a knowledge of
the analytical procedures. The requirement for risk
assessment is that only data that have been reviewed
according to a specified level or plan will be used in the
quantitative risk assessment Any analytical errors, or
limitations in data that are identified by the review, must
be noted in the risk assessment if the data are used. An
explanation for qualifiers used must be included with
the review report.
All data should receive some level of review. The risk
assessor may receive data prior to the quantitative
baseline risk assessment that were not reviewed. Data
that have not been reviewed must be identified because
the lack of review increases the uncertainty for the risk
assessment These data may lead to false positive or
false negative assessments and quantitation errors.
Unreviewed data may also contain transcription errors
and calculation errors. Data may be used in the
preliminary assessment before review, but must be
reviewed at a predetermined level before use in the final
risk assessment
Depending upon data user requirements, the level and
depth of the data review are variable. The level and
depth of the data review may be determined during the
planning process and must include an examination of
laboratory and method performance for the samples and
analytes involved. This examination includes:
• Evaluation of data completeness,
• Verification of instrument calibration,
• Measurement of laboratory precision using
duplicates; measurement of laboratory accuracy
using spikes,
• Examination of blanks for contamination,
102
-------
EXHIBIT 64. RELATIVE IMPORTANCE OF DETECTION LIMIT
AND CONCENTRATION OF CONCERN: DATA ASSESSMENT
Relative Position of Method
Detection Limit (MDL) and
Concentration of Concern (COC)
Consequence
^^ Confidence
COC Limits
Non-Detects and
Detects Useabfe
Possibility of
False Positives and
False Negatives
Concentration
Non-Detects Not
UseaWe
Detects UseaWe
Possibility of False
Negatives
• Assessmentofadherencetomethodspecifications
and QC limits, and
• Evaluation of method performance in the sample
matrix.
Specific data review procedures are dependent upon the
method and data user requirements. Section 5.6.1
details procedures for evaluating QC samples for
laboratory and method performance. CLP data review
procedures are performed according to criteria outlined
in National Functional Guidelines for Organic Data
Review (EPA 1991e) and Laboratory Data Validation:
Functional Guidelines for Evaluating Inorganics
Analyses (EPA 1988e). Minimum requirements for
data review are listed in Exhibit 61.
5.6 ASSESSMENT OF CRITERION VI:
DATA QUALITY INDICATORS
Minimum Requirements
• Sampling variability quantitated for each
analyte.
• QC samples required to identify and
quantitate precision and accuracy.
• Sampling and analytical precision and
accuracy quantitated.
The assessment of data quality indicators presented in
this chapter is significant to determine data useability.
103
-------
EXHIBIT 65. CONSEQUENCES OF ALTERNATIVE SAMPLING
STRATEGIES ON TOTAL ERROR ESTIMATE
Group Data by
MedlunVStratum
+ ByAnalyte
/ Multipl«\
Judgmental
Modal
4
No
Y«
Non-Statiiitical
Traatmant
104
-------
•• Qualified data can usually be used for
quantitative risk assessments.
The assessment of data quality indicators for either
sampling or analysis involves the evaluation of five
indicators: completeness, comparability, represen-
tativeness, precision, and accuracy. Uncertainties in
completeness, comparability, and representativeness
increase the probability of false negatives and false
positives when the data are used to test particular
hypotheses as pan of the site evaluation. This increase
in uncertainty can affect the confidence of chemical
identification. Variation in completeness, comparability,
representativeness, precision, and accuracy affects the
uncertainty of estimates of average concentration and
reasonable maximum exposure (RME). Once the
indicator is examined or a numerical value is determined,
the results can be compared to the performance objectives
established during RI planning. This comparison
determines the useability of the data and any required
corrective actions.
A summary of the minimum requirements for data
quality indicators is presented in Exhibit 61, and the
evaluation process is illustrated in Exhibit 65. Specific
requirements for each indicator are presented in the
following sections.
5.6.1 Assessment of Sampling and
Analytical Data Quality
Indicators
The major activity in determining the useability of data
based on sampling is assessing the effectiveness of the
sampling operations performed. Samples provided for
analysis must answer the four basic decisions to be
made with RI data in risk assessment (cited at the
beginning of this chapter) that are translated into site-
specific objectives based on scoping and planning
decisions.
Independent data review evaluates laboratory results,
not sampling. Determining the useability of analytical
results begins with the review of QC samples and
qualifiers to assess analytical performance of the
laboratory and the method. It is more important to
evaluate the effect on the data than to determine the
source of the error. The data package is reviewed as a
whole for some criteria; data are reviewed at the sample
level for other criteria, such as holding time. Factors
affecting the accuracy of identification and the precision
and accuracy of quantitation of individual chemicals,
such as calibration and recoveries, must be examined
analyte-by-analyte. The qualifiers used in the review of
CLP data are presented and their effect on data quality
is discussed in this section. Exhibit 66 presents a
EXHIBIT 66. USE OF QUALITY CONTROL DATA FOR RISK ASSESSMENT
Quality Control Criterion
Spikes (High Recovery)
Spikes (Low Recovery)
Duplicates
Blanks
Calibration
Tune
Internal Standards _
(Reprodudbiity) 3
Internal Standards
(High Recovery)
Internal Standards
(Low Recovery)
Effect on MentH icatton When
Criterion to not Met
-
False Negative1
None, unless analyte found
in one duplicate and not the
other. Then either false
positive or false negative.
False Positive
-
False Negative
-
-
False Negative 1
QuantlUUve BlM
High
Low
Hl£s
High
High or
Low2
-
-
Low
High
Use
Use data as upper limit.
Use data as lower limit.
Use data as estimate-poor precision.
Set confidence level 5x blank.
Use data above confidence level.
as estimate.
Use data as estimate
unless problem is extreme.
Rojoct data or examine raw data and
use professional judgment.
Use data as estimate-poor precision.
Use data as lower limit.
Use data as upper limit.
1 False negative only likely if recovery is near zero.
~ Effect on bias determined by examination of data for each individual analyte.
3 Includes surrogates and system monitoring compounds.
105
-------
summary of the QC samples and the data use implications
of qualified data. Corrective action options are shown
in Exhibit 62.
Sample media can be more complex than expected in
environmental analysis. For example, sludge or oily
wastes may contain interfering chemicals whose
presence cannot be predicted in precision and accuracy
measurements. The risk assessor must examine the
reported precision [relative percent difference (RPD)]
and accuracy [percent recovery (%R)] data to determine
useability. Ranges used for rejection and qualification
of CLP data have been determined based on the analysis
of target compounds in environmental media These
ranges, documented in the Functional Guidelines (EPA
1991e, EPA 1988e) can be used in the absence of
specifications in the planning documents.
Completeness. Completeness for sampling is
calculated by the following formula:
Percent
Completeness
(Number of Acceptable Data Points') x 100
Total Number of Samples Collected
This measure of completeness is useful for data collection
and analysis management but misses the key risk
assessment issue, which is the total number of data
points available and acceptable for each chemical of
potential concern. Incompleteness should be assessed
to determine if an acceptable level of data useability can
still be obtained or whether the level of completeness
must be increased, either by further sampling or by other
corrective action. Any decrease in the number of
samples from that specified in the sampling design will
affect the final results. In this case, the option of
obtaining more samples should be reviewed.
Minimum Requirements
for Completeness
Impact When Minimum
Requirements Are Not Met
Corrective Action
Percentage of sample
completeness determined
during planning to meet
specified performance
measures.
100% of all data for analytes
in critical samples (at least
one sample per medium per
exposure area).
All data from critical samples
considered crucial.
Background samples and
broad spectrum analyses are
usually critical.
Higher probability of false
negatives.
Reduction in confidence
level and power.
A reduction in the number of
samples reduces site
coverage and may affect
representativeness. Data for
critical samples have
significantly more impact
than incomplete data for
non-critical samples.
Useability of data is
decreased for critical
samples.
Useability of data is
potentially decreased for
non-critical samples.
Reduced ability to
differentiate site levels from
background.
Impact of incompleteness
generally decreases as the
number of samples
increases.
Resampling or reanatysis to
fill data gaps.
Additional analysis of
samples already at
laboratory.
Determine whether the
missing data are crucial to
the risk assessment (i.e.,
data from critical samples).
21-002-081
106
-------
Typical causes for sample attrition include site conditions
preventing sample collection (e.g., a well runs dry),
sample breakage, and invalid or unuseable analytical
results. Incompleteness can increase the uncertainty
involved in risk assessments by reducing the available
number of samples on which identification and estimates
of concentration of chemicals at the site are based. The
reduction in the number of samples from the original
design further affects representativeness by reducing
site coverage and increases the variability in
concentration estimates. Only the collection of additional
samples will resolve the problem, unless the samples
involved were duplicates or splits. In this case, or if the
cause was laboratory performance, the extracts may be
considered for reanalysis.
Completeness for analytical data is calculated by the
following formula:
Percent (Mumher of Acceptable Samples^ « 100
Completeness Total Number of Samples Analyzed
The completeness for analytical data required for risk
assessment is defined as the number of chemical-specific
data results for an exposure area in an operable unit that
are determined acceptable after data review.
An analysis is considered complete if all data generated
are determined to be acceptable measurements as defined
in the SAP. Results for each analyte should be present
for each sample. In addition, data from QC samples
necessary to determine precision and accuracy should
be present. QC samples and the effects of problems
associated with these samples are discussed later in this
section.
Comparability. Comparability is not compromised
provided that the sampling design is unbiased, and the
sampling design or analytical methods have not changed
over time. If any of these factors change, the risk
assessor may experience difficulties in combining data
sets to estimate the RME. The determination of the
RME is based on the principal of estimating risk over
time for the exposure area. The ideal situation occurs
when samples can be added within the basic design,
decreasing the level of uncertainty.
•*• Anticipate the need to combine data from
different sampling events and/or different
analytical methods.
Comparability is a very important qualitative data
indicator for analytical assessment and is a critical
Minimum Requirements
for Comparability
Impact When Minimum
Requirements Are Not Met
Corrective Action
Unbiased sampling design or
documented reasons for
selecting another sampling
design.
The analytical methods used
must have common analytical
parameters.
Same units of measure used
in reporting.
Similar detection limits.
Equivalent sample
preparation techniques.
• Non-additrvrty of sample
results.
• Reduced confidence, power,
and ability to detect
differences, given the
number of samples
available.
• Increased overall error.
For Sampling:
Statistical analysis of effects
of bias.
For Analytical Data:
Preferentially use those data
that provide the most
definitive identification and
quantitation of the chemicals
of potential concern. For
organic chemical
identification, GC-MS data
are preferred over GC data
generated with other
detectors. For quantitation,
examine the precision and
accuracy data along with the
reported detection limits.
Reanalysis using comparable
methods.
21-002482
107
-------
parameter when considering the combination of data
sets from different analyses for the same chemicals of
potential concern. The assessment of data quality
indicators determines if analytical results being reported
are equivalent to data obtained from similar analyses.
Only comparable data sets can readily be combined for
the purpose of generating a single risk assessment
calculation.
The use of routine methods simplifies the determination
of comparability because all laboratories use the same
standardized procedures and reporting parameters. In
other cases, the risk assessor may have to consult with
an analytical chemist to evaluate whether different
methods are sufficiently comparable to combine data
sets. The RPM should request complete descriptions of
non-routine methods. A preliminary assessment can be
made by comparing the analytes, useful range, and
detection limit of the methods. If different units of
measure have been reported, all measurements must be
converted to a common set of units before comparison.
Representativeness. Representativeness of data is
critical to risk assessments. The results of the risk
assessment will be biased to the degree that the data do
not reflect the chemicals and concentrations present in
the exposure area or unit of interest. Non-representative
chemical identification may result in false negatives.
Non-representative estimates of concentration levels
may be higher or lower than the true concentration.
Non-representative sampling can usually only be
resolved by additional sampling, unless the potential
limitations of the risk assessment are acceptable.
It is important to determine whether any changes have
occurred in the actual sample collection that convert an
originally unbiased sampling plan into a biased sampling
episode. Bias in unbiased designs is difficult to assess
because no measure of the true value is known. Bias is
assumed in non-statistical designs.
Representativeness is primarily a planning concern.
The solution is in the design of a sampling plan that is
representative. Once the design is implemented, only
the sampling variability is evaluated during the
assessment process, unless contamination occurs in the
QC samples or blanks, or problems exist during sample
preparation that affect sample results. Incompleteness
of data potentially decreases representativeness and
increases the potential for false negatives and the bias in
estimations of concentration.
Representativeness is determined by examining the
sampling plan, as discussed in Section 3.2. In
determining the representativeness of the data, the
evaluator examines the degree to which the data meet
the performance standards of the method and to which
the analysis represents the sample submitted to the
laboratory. Analytical data quality affects
representativeness since data of low quality may be
rejected for use in risk assessments. Holding time,
sample preservation, extraction procedures, and results
Minimum Requirements
for Representativeness
Sample data representative
of exposure area and
operable units.
Documented sample
preparation procedures.
Filtering, compositing, and
sample preservation may
affect representativeness.
Documented analytical data
as specified in the SAP.
Impact When Minimum
Requirements Are Not Met
Bias high or low in estimate
of RME.
Increased likelihood of false
negatives.
Inaccurate identification or
estimate of concentration
that leads to inaccurate
calculation of risk.
Remaining data may no
longer sufficiently represent
the site if a large portion of
the data are rejected, or if all
data from analyses of
samples at a specific location
are rejected.
Corrective Action
Additional sampling.
Examination of effects of
sample preparation
procedures.
For critical samples,
roanalyses of samples or
resampling of the affected
site areas. For non-critical
samples, reanalyses or
resampling should be
decided by the RPM in
consultation with the
technical team.
If the resampling or
reanalyses cannot be
performed, document in the
site assessment report what
areas of the site are not
represented due to poor
quality of analytical data.
21-002-OU
108
-------
from analyses of blanks affect the representativeness of
analytical data (see Appendix V).
Precision. The two basic activities performed in the
assessment of precision are estimating sampling
variability from the observed spatial variation and
estimating the measurement error attributable to the
data collection process. Assumptions concerning the
sampling design and datadistributionsmust be examined
prior to interpreting the results. This examination will
provide the basis for selecting calculation formulas and
knowing when statistical consultation is required.
The type of sampling design selected is critical to the
estimation of sampling variability as discussed in
Sections 3.2 and 4.1. If the sampling design is
judgmental, the nature of the sampling error cannot be
determined and estimates of the average concentrations
of analytes may not be representative of the site.
<*• Determine the distribution of the data
before applying statistical measures.
The nature of the observed chemical data distribution
affects estimation procedures. The estimation of
variability and confidence intervals will become complex
if the distribution cannot be assumed normal or to
approximate normal when transformed to log normal.
Estimates of the 95% upper confidence limit of the
average concentration for the RME should be based on
an analysis of the frequency distribution of the data
whenever the database is sufficient to support such
analysis. Statistical tests may be used to compare the
distribution of the observed data with the normal or log
normal distribution (Gilbert 1987). Graphs of data
without statistical test results may also be acceptable for
some data sets. Statistical computer software can assist
in the analyses of data distribution.
Sampling variability. Exhibit 67 summarizes the
assessment procedures for the evaluation of variability
from different sampling procedures. The estimation of
confidence levels, power, and minimum detectable
relative differences requires assumptions about the
coefficients of variation from sampling variability for
Minimum Requirements
for Precision
Impact When Minimum
Requirements Are Not Met
Corrective Action
• Confidence level of 80% (or
as specified in DQOs).
• Power of 90% (or as specified
in DQOs).
• Minimum detectable relative
differences specified in SAP
and modified after analysis of
background samples if
necessary.
• One set of field duplicates or
more as specified in the SAP.
• Analytical duplicates and
splits as specified in the SAP.
• Measurement error specified.
Errors in decisions to act or
not act based on analytical
data.
Unacceptable level of
uncertainty.
Increased variability of
quantitative results.
False negatives for
measurements near the
detection limits.
For Sampling:
• Add samples based on
information from available
data that are known to be
representative.
• Adjust performance
objectives.
For Analysis:
• Analysis of new duplicate
samples.
• Review laboratory protocols
to ensure comparability.
• Use precision measure-
ments to determine
confidence limits for the
effects on the data.
• The risk assessor can use
the maximum sample results
to set in upper bound on the
uncertainty in the risk
assessment if there is too
much variability in the
analyses.
21-002-064
109
-------
EXHIBIT 67. STEPS TO ASSESS SAMPLING PERFORMANCE
1. Confirm statistical assumptions.
2. Summarize analyte detection data by strata: media within site or site subgroups
and strata within media.
3. Transform analyte concentration data so distribution is approximately normal.
4. Calculate the coefficient of variation for each analyte detected.
5. Using Exhibit 47 'Relationships Between Measures of Statistical Performance
and Number of Samples Required,' look up the range of power, confidence
level and minimal detectable relative differences for the calculated
coefficient of variation.
6. Compare the statistical performance measures required to those achievable
given the coefficient of variation and sample size.
7. If the performance objectives are achieved, go to Step 9.
If the required statistical performance levels are not met, then additional samples
must be taken or one or more of the performance parameters must be changed.
If samples are to be added, Exhibit 47 and the calculation formulas in Appendix
IV can be used to determine the number needed.
8. If the performance parameters are to be changed, the parameter to be changed
should be the one which will increase the probability of taking unnecessary
action as opposed to unnecessary risk.
9. Examine the results of the QC samples. Sample results must be considered to
be qualitative if no results are available for QC samples.
10. If the QC sample results indicate possible bias through contamination, take
appropriate corrective action.
each chemical of potential concern. Hie RPM or risk
assessor should discuss the implications of these
assumptions with a statistician to determine their
potential impacts on data useability.
•*• Determine the statistical measures of
performance most applicable to site
conditions before assessing data useability.
Once the statistical assumptions and observed analyte
variability are known, selected statistical performance
measures can be assessed to determine the data quality
achieved. Additional samples may be needed, or
modified DQOs required, as a result of evaluating
sampling variability. Three issues are involved in the
assessment of required statistical performance:
• Level of certainty or confidence,
• Power, and
• Minimum detectable relative difference.
The required level for each of these performance
measures should be included in the SAP as DQOs. The
user's data quality requirements defined by these
statistical measures determine the number of samples
that are taken during data collection. Recommended
minimum statistical performance parameters for
110
-------
discriminating contaminant concentrations from
background levels in risk assessment are provided in
Exhibit 68.
EXHIBIT 68. RECOMMENDED
MINIMUM STATISTICAL
PERFORMANCE PARAMETERS FOR
RISK ASSESSMENT
Null Hypothesis: On-sfte Contaminant
Concentrations are not Higher than
the Background
Confidence level:
80% minimum, reject null when true (take
unnecessary action).
2
Power
90% minimum, accept null when false (fail to
take action when action is required).
Minimum detectable relative difference:
10% - 20%, usually depends on concentration
ol concern.
1 (1 -false positive estimate) or (1 -a).
2 (1 -false negative estimate) or (1 -(3).
Source: EPA 19891.
21-OOZ-OM
First, summarize the sample results at the analyte level
by stratum and strata within media to determine whether
the performance objectives have been met. Sampling
error is not relevant if a particular combination of
stratum and analyte yields only a single data point In
that case, assessment proceeds to that of analytical error
for that stratum and analyte combination.
Hie distribution for stratum and analyte combinations
with multiple data points should usually be examined
for normality and transformed to log normal. The
coefficient of variation is calculated for each stratum
and analyte combination. If the distribution resulting
from the transformation is not normal, a new
distributional model will need to be identified and
validated in consultation with a statistician. Non-
parametric procedures which require no distributional
assumptions may also be used.
Conversely, the statistical performance achieved can be
determined, given the coefficient of variation. This
performance should be compared to the requirements
stated in planning. If the performance objectives are
achieved, the risk assessor can proceed to the assessment
of measurement error.
If the required .statistical performance objectives are not
met, additional samples must be taken, or one (or more)
of the performance parameters must be changed. If
samples are added, the tables and formulas provided in
Chapter 4 and Appendix IV can be used to calculate the
number of samples required. If a performance parameter
is changed, it should be the one that will increase the
probability of taking unnecessary action as opposed to
an increased probability of unnecessary risk. The
uncertainty level will then be reduced first, the minimum
detectable relative difference will be increased second,
and the level of power will be reduced last. Minimum
recommended levels for performance parameters in
risk assessment in the absence of site-specific DQOs are
80% confidence levels, 90% power, and 10-20%
minimum detectable relative differences (EPA 19890.
Exhibit 68 summarizes the recommended DQOs for
statistical performance parameters.
Measurement error. Measurement error is estimated
using the results of field duplicate samples. Field
duplicates determine total within-batch measurement
error, including analytical error if the samples are also
analyzed as laboratory duplicates. The estimate is of the
difference between analytical values reported for
duplicates. This type of variation has four basic sources:
sample collection procedures, sample handling and
storage procedures, analytical procedures, and data
processing procedures.
The formula for computing the relative percentdifference
between duplicates is:
V *100
where R, and R2 are the results from the first and second
duplicate samples, respectively. Precision is a measure
of the repeatability of a single measurement and is
evaluated from the results of duplicate samples and
splits.
Low precision can be caused by poor instrument
performance, inconsistent application of method
protocols, or by a difficult, heterogeneous sample matrix.
The last effect can be distinguished from the others by
evaluation of laboratory QC data.
If splitsamples have been analyzed by different methods
or different laboratories, then data users have a measure
of the quality of individual techniques. Splits are
particularly effective when one laboratory is a reference
laboratory. Ifbotbsetsofdataexhibitthesame problems,
then laboratory performance can usually be ruled out as
a source of error. Splits are also useful when using non-
routine methods or comparing results from different
analytical methods.
Ill
-------
Accuracy. Accuracy is a measure of overestimation or
underestimation of reported concentrations and is
evaluated from the results of spiked samples. The
procedure for determining accuracy will vary according
to differences in the number of measurements and the
precision of the estimates. Data that are not reported
with confidence limits cannot be assigned weights
based on precision and should not be combined for use
(Taylor 1987).
Spiked samples are particularly useful in the analysis of
complex sample types because they help the reviewer
determine the extent of bias on the sample measurement.
A set of standards at known concentrations is mixed into
a portion of the sample or into distilled water prior to
sample preparation and analysis. The analytical results
are compared to the amount spiked to determine the
level of recovery. It is important to note that unless
every sample is spiked, spike recoveries indicate only a
trend rather than a specific quantitative measure.
Accuracy is controlled primarily by the analytical process
and is reported as bias. The absolute bias of a sampling
design cannot be determined unambiguously because
the true value of the chemicals of concern in the exposure
area can never be known. However, statistically based
sampling designs described in Chapter 4 are structured
to produce unbiased results.
Bias can be estimated using field spikes on field
evaluation or audit samples to assess the accuracy and
comparability of results. These estimates will reflect
the effects of sample collection, handling, holding time,
and the analytical process on the result for the sample
collected.
Bias is estimated for the measurement process by
computing the percent recovery (%R) for the spiked or
reference compound as follows:
%R =
(Measured Amount - Amount in Unspiked Sample) X 100
Amount Spiked
Because of the inherent problems associated with the
spiking procedure and the interpretation of recovery,
spikes are considered minimum requirements only if
specified in the SAP. Held matrix spikes are currently
not recommended for use in soils (EPA 19890.
Held blanks are evaluated to estimate the potential bias
caused by contamination from sample collection,
preparation, shipping and/or storage. Results for the
analysis of field blanks indicate whether contamination
resulted in bias, but they are not estimates of accuracy.
Bias pertaining to analytical recoveries is computed as
follows:
Percent
Bin
Amount*Amount in Unsnikcd Samole) x 100
Amount Spiked
Minimum Requirements
for Accuracy
Impact When Minimum
Requirements Are Not Met
Corrective Action
Field spikes to assess
accuracy of non-detects and
positive sample results if
specified in the SAP.
Analytical spikes as
specified in the SAP.
Use analytical methods
(routine methods whenever
possible) that specify
expected or required
recovery ranges using
spikes or other QC
measures.
No chemicals of potential
concern detected in the
blanks.
Increased potential for false
negatives. If spike recovery
is low, it is probable that the
method or analysis is biased
low for that analyte and
values of all related samples
may underestimate the
actual concentration.
Increased potential for false
positives. If spike recovery
exceeds 100%, interferences
may be present, and it is
probable that the method or
analysis is biased high.
Analytical results
overestimate the true
concentration of the spiked
analyte.
Consider resampling at
affected locations.
No correction factor is
applied to CLP data on the
basis of the percent recovery
in calculating the analyte
concentration.
If recoveries are extremely
low or extremely high, the
risk assessor should consult
with an analytical chemist to
identify a more appropriate
method for reanalysis of the
samples.
21-002-085
112
-------
Blanks are of primary concern for the analysis of bias
involved in sampling because of the difficulty in
performing field spikes and the availability of appropriate
reference standards and matrix for evaluation samples.
Results from blanks can be used to estimate the extent
of high bias in the event of contamination. The following
procedures should be implemented to prevent the
assignment of false positive values due to blank
contamination:
• If the field blanks are contaminated and the
laboratory blanks are not, the RPM or risk assessor
can conclude that contamination occurred prior to
receipt of the samples by the laboratory. If the
contamination is significant (i.e., it will interfere
with the determination of risk),consider resampling
at affected locations.
• If it is not possible to resample, the RPM or risk
assessormust assess tbeeffectof the contamination
on the potential for false positives. Often, this
determination can be made by examining data
from samples located nearby. If all samples and
blanks show the same level of a particular chemical,
the presence of the chemical in the samples is most
likely due to contamination.
• If the laboratory blanks are contaminated, the
laboratory should be required to rerun the
associated analyses. This is especially important
in the case of critical analytes or samples. Before
reanalyses, the laboratory must demonstrate
freedom from contamination by providing results
of a clean laboratory blank. Note: If laboratory
blanks are contaminated, field blanks will generally
also be contaminated.
• If reanalysis is not possible, then the sample data
must be qualified. The Functional Guidelines
provide examples of blank qualification.
Chemicals detected in the associated samples
below the action level defined in the Functional
Guidelines are considered undetected.
Data qualifiers. All data generated by the routine
analytical services of the CLP are reviewed and qualified
by Regional representatives according to the guidelines
found in the Functional Guidelines as modified to fit the
requirements of the individual Regions.
<*• Use data qualified as U or J for risk
assessment purposes.
Analytes qualified with a U are considered "not
detected." If precision and accuracy are acceptable (as
determined by the QC samples), data are entered in the
data summary tables in the data validation report as the
SQL or corrected quantitation limit (MDL corrected for
dilution and percent moisture), and qualified with a U.
Note that the same chemical can be reported undetected
in a series of samples at different concentrations because
of sample differences.
Data qualified with an R are rejected because
performance req uiremen ts in the sample or in associated
QC analyses were not met. For example, if a mass
spectrometer "tune" is not within specifications, neither
the identification nor quantitation of chemicals can be
accepted with confidence. Extremely low recoveries of
a chemical in a spiked sample might also warrant an R
designation for that chemical in associated samples
because of the risk of false negatives (see Appendix VI).
Data qualified with a J present a more complex issue. J-
qualified data are considered "estimated" because
quantitation in the sample or in associated QC samples
did not meet specifications. The justification for
qualifying the data should be explained in the validation
report Draft revisions of the Functional Guidelines
propose that the justification be included on a qualifier
summary table submitted with the validation report.
Data can be biased high or low when qualified as
estimated. The bias can often be determined by
examining the results of the QC samples. For example,
if interfering levels of aluminum are found in inorganic
analysis of the interference check sample, the sample
results are probably biased high because the signal
overlap is added to the signal being reported. When
volatile organic compounds are qualified J for holding
time violations, the results are usually biased low because
some of the volatile compounds may nave volatilized
during storage.
Data associated with contaminated blanks are not
consideredestimatedandarenotflaggedJ. Thepresence
of the blank contaminant chemical in the analytical
samples is questionable at levels up to 5 to 10 times
those found in the blank, depending on the nature of the
analyte. An action levelisdetenninedforeach chemical
based on the quantity found in the blank. Data above the
action level are accepted without qualification and data
between the contract required quantitation limit (CRQL)
and the action level are qualified U (undetected).
Estimated organics and inorganics data that are below
the CRQL or contract required detection limit (CRDL)
are qualified as UJ. This qualifier signifies that the
quantitation limit is estimated because the QC results
did not meet criteria specified in the SAP.
Other qualifiers may be added to the analytical data by
the laboratory. A set of qualifiers (or flags) has been
defined by the CLP for use by the laboratories to denote
113
-------
problems with the analytical data. These qualifiers and
their potential use in risk assessment are discussed in
RAGS (EPA 1989a).
5.6.2 Combining the Assessment of
Sampling and Analysis
Once the quality of the sampling and analysis effort has
been assessed using the five data quality indicators,
combine the results to determine the overall assessment
of a particular indicator across sampling and analysis.
Combining the assessment for completeness,
comparability, and representativeness is discussed in
thissectionasaqualitativeprocedure. Statisticalmodels
are available for combining data sets with different
variability and bias. The risk assessor should consult a
chemist or statistician if the magnitude of the sampling
and analysis effort warrants the use of a formal statistical
treatment of comparability.
The basic model for estimating total variability across
sampling and analysis components is presented in Exhibit
69. An example of a non-statistical approach to
combining the assessment results is given in Exhibit 70.
Using this approach, each data quality indicator is
assessed to determine whether a problem exists in either
sampling or analysis. This assessment leads to different
combinations of problem determination. For example,
completeness may have been a problem in sampling
[YES] but not a problem in analysis [NO]; the
combination is [YES/NO].
Basic guidance is given on the combinations of sampling
and analysis once assessment patterns based on the
detenninationofaproblemhavebeenestablished. This
guidance is qualitative in nature and is presented to
assist in organizing the data assessment problem for the
application of professional judgment If the assessment
pattern is [NO/NO], the issue of combining results is not
a problem. Conversely, if the pattern is [YES/YES], the
issue of combining results is an issue of the effects of the
combined magnitudes. Instances of combined sampling
and analysis problems for a single indicator will have
significant effects on the risk assessment uncertainty.
The most complicated assessment pattern to interpret is
encountered when a problem occurs in one area but not
in another (e.g., in sampling but not in analysis). This
situation is briefly discussed for each indicator in the
following sections.
EXHIBIT 69. BASIC MODEL FOR ESTIMATING
TOTAL VARIABILITY ACROSS SAMPLING AND
ANALYSIS COMPONENTS
where
ot - total variability
a_ = measurement variability
a = population variability
= ov
where
2222
+ °h +
-------
EXHIBIT 70. COMBINING DATA QUALITY INDICATORS FROM
SAMPLING AND ANALYSIS INTO A SINGLE ASSESSMENT
OF UNCERTAINTY
Data Quality
Indicators
Completeness
Comparability
Representativeness
Precision
Accuracy
Assessment of Problems
Sampling Analytical
- __
YES
NO
...._ —
YES
NO
- _
YES
NO
_
YES
NO
MV^^H -I^^^M
YES
NO
_
YES
NO
___ ___
YES
NO
-t^"~*
YES
NO
— W __M—
YES
NO
•»W^» «W_BH
YES
NO
Combined Sampling 1
and Analytical . 1
Determination 1
YES/YES 1
YES/NO 1
NO/YES 1
YES/YES 1
1
YES/NO I
NO/YES 1
1
YES/YES I
YES/NO 1
I
NO/YES •
1
YES/YES 1
YES/NO 1
I
NOA'ES
YESA'ES
YES/NO
NOA'ES
*
The combination [NO/NO] indicates that the data quality indicator will not affect the
level of uncertainty in data useabilfty.
21-002-070
115
-------
Completeness. A sample is considered incomplete for
all analytes. Analytical incompleteness is usually related
to particular analytes. In this instance [YES/YES], the
effect on the risk assessment will vary according to
chemical. For some chemicals, the data points will be
lost in both sampling and analysis.
The effects of a loss in the number of sample points for
a particular chemical can be substantial. For example,
if collection of 10 samples was planned and one sample
could not be collected because of site access problems,
one was broken in transport, and the laboratory
experienced analysis problems with three samples for
the chemical of potential concern causing the data to be
rejected, then only five data points remain.
If the assessment pattern is [YES/NO], the effects are
distributed across all chemicals involved in the risk
assessment If the pattern is [NO/YES], the effects are
localized to the particular chemical affected.
Comparability. Comparability problems in sampling
are primarily due to different sampling designs and time
periods. Seasonal variations are treated like spatial
variations because the risk assessment is calculated as
risk over time. Data can be averaged and considered as
a single data set For analytical data, comparability
problems are related primarily to the use of different
methods and laboratories. A pattern of [YES/YES] will
indicate that the risk assessor will have considerable
difficulty in combining the various data sets into a
single assessment of risk. In situations of [YES/NO] or
[NO/YES], the problem of sampling comparability is
more difficult to resolve. Models exist for determining
comparability between methods and integrating results
across laboratories. These models involve the general
statistical approach to confirming data sets with different
but known variability and bias (Taylor 1987).
Representativeness. Representativeness in sampling
is critical to the risk assessment. Non-representativeness
affects both false negatives (chemicals not identified)
and estimates of concentration and, therefore, affects
estimates of RME. Analytical representativeness
involves the question of whether the analytical results
represent the sample collected. For example, holding
times and sample preservation can cause the analytical
results not to be representative of the sample collected.
These questions should be treated separately in the
discussion of effects.
Precision. The contribution to imprecision from
sampling variability often exceeds that from analytical
variability in the measurement process. If precision is
a problem in both sampling and analysis, the risk
assessor should focus on the impact of sampling
variability on theestimateof RME. Analytical variability
will be minimal in comparison to the effects of sampling
variability unless the sampling variability is untypically
low and the analytical variability is untypically high.
Accuracy. The assessment of accuracy in sampling is
focused primarily on recoveries from spiked or
performance evaluation samples. Analytical
performance and potential blank contamination are
reflected in analytical spike recoveries. If the pattern is
[YES/YES] for accuracy, this may require assessment
of calibration, or of potential blank contaminants, and
integration of their possible effects by comparison of
results from laboratory and field QC samples.
If the accuracy pattern is [NO/YES], then the issue is
analytical performance. Low variability in sampling as
measured by low coefficients of variation for chemicals
of potential concern should increase the risk assessor's
concern over an analytical accuracy problem.
High sampling variability (CV>25%) will greatly reduce
the effects of analytical bias on the level of certainty of
the risk assessment.
116
-------
Chapter 6
Application of Data to Risk Assessments
This chapter provides guidance for integrating the
assessment of data useability to determine the overall
level of uncertainty of risk assessment. This guidance
builds on each of the previous chapters.
• Chapter 2 explained the risk assessment process
and the roles and responsibilities of key
participants. Exhibit 5 defined a continuum of
level of certainty in the baseline risk assessment
result based on the ability of the risk assessor to
quantitate or qualify the level of uncertainty
associated with the analytical data.
• Chapter 3 defined six data useability criteria and
examined preliminary issues that must be
considered while planning sampling and analysis
activities to increase the certainty of the analytical
data collected for the risk assessment.
• Chapter 4 presented strategies for planning
sampling and analysis activities based on the six
data useability criteria.
• ChapterSdesaibedhowtouseeachdatauseabUity
criterion to determine the effect of sampling and
analysis issues on data quality and on the useability
of data in baseline risk assessment.
The Data Useability Worksheet (Exhibit 63) assists the
risk assessor in summarizing data quality across the
various assessment phases. This worksheet is the basis
for this chapter's discussion of the impact of analytical
data quality on the level of certainty of the risk
assessment
6.1 ASSESSMENT OF THE LEVEL OF
CERTAINTY ASSOCIATED WITH
THE ANALYTICAL DATA
This section explains how to assess the level of
confidence in sampling and analytical procedures in the
context of the four major decisions to be made by the
riskassessorwith environmental analytical data. Exhibits
in this section apply the data useability criteria, defined
in Chapter 3 and appearing on the Data Useability
Worksheet, to these four decisions. Data useability
criteria affect the level of confidence involved in each
decision. The level of certainty in the data collection
and evaluation component of risk assessment affects the
overall certainty of the risk estimate.
6.1.1 What Contamination is Present
and at What Levels?
The risk assessor's first task is to use analytical data to
determine what contamination is present at the site and
at what levels (i.e., what potential exists for increased
risk from the contamination). Exhibit 71 lists the
criteria from the Data Useability Worksheet that affect
this decision. The most critical analytical data question
to be answered before calculating the risk is the
probability of false negatives or false positives. False
negatives are of greater concern in risk assessment than
false positives, since false negatives may result in a
decision that would not be protective of human health.
False positives cause the calculated risk to be biased
high, and are of concern because taking unnecessary
action at a site is costly.
*• The major concern with false negatives
is that the decision based on the risk
assessment may not be protective of human
health.
Probability of false negatives. False negatives occur
when chemicals of potential concern are present but are
not detected by the sampling design or the analytical
method. The probability of false negatives can be
determined by using the following parameters from the
Data Useability Worksheet- analytical methods, data
review, sampling completeness, sampling
representativeness, analytical completeness, analytical
precision and accuracy, and combined error.
«• False negatives can occur if sampling is
not representative, if detection limits are
above concentrations of concern, or if spike
recoveries are very low.
Sampling strategies can increase the probability of false
negatives if too few samples were taken or if sections of
the site were not sampled. The probability of false
negatives increases if sampling of any exposure pathway
was not representative.
Knowledge of analyte-specific detection limits is critical
to determining the probability of false negatives.
Recovery values from spikes, internal standards,
Acronyms
RAGS Risk Assessment Guidance for Superfund
SAP sampling and analysis plan
SOP standard operating procedure
117
-------
EXHIBIT 71. DATA USEABILITY CRITERIA AFFECTING
CONTAMINATION PRESENCE
Worksheet
Reference
Data Useability
Criterion
Data Collection and
Evaluation Decision
1
2B
2C
3A
4
5
6A
6C
6D
6E
Reports to risk assessor
Documentation (SOPs)
Documentation (analytical records)
Data sources (analytical)
Analytical methods
Data review
Completeness (analytical)
Representativeness (sampling)
Precision (analytical)
Accuracy (sampling and analytical)
What contamination is
present and at what
levels?
surrogates, and system monitoring compounds are used
to assess the level of accuracy and precision in laboratory
data and determine whether the detection limits stated in
the analytical methods have been met
• The probability of false negatives for an analyte is
high if the concentration of concern is at or below
the detection limit. This probability should have
been documented during planning if no analytical
methods were found with detection limits below
the concentration of concern. If the concentration
of concern is very near the detection limit, a false
negative can occur because of "drift" in instrument
response. This behavior may not be reflected in
data from spike recoveries or blanks.
• The probability of false negatives is low if spike
recoveries are acceptable, or biased high as
documented during data review, and the detection
limits are below the concentration of concern for
each analyte.
• Theprobabilityoffalsenegativesisdirectlyrelated
to the amount of bias if spike recoveries are biased
low and detection limits are below the concentration
of concern for each analyte. The effect is more
pronounced the closer the concentration of concern
is to the detection limits.
• The possibility of false negatives should be
carefully evaluated whenever sample extracts have
been highly diluted (i.e., diluted beyond normal
method specifications).
Probability of false positives. False positives occur
when a chemical of concern is detected by an analytical
21-002471
method but is truly not present at the site. Assessment
of the following parameters from the Data Useability
Worksheet can be used to determine the probability of
false positives: analytical methods, datareview, sampling
accuracy, analytical completeness, analytical precision
and accuracy, and combined error.
*• Fa/sa positives can occur when blanks
are contaminated or spike recoveries are
very high.
Sampling and analysis uncertainties connected with
false positives can be assessed by examining the results
of quality control samples. Blank contamination is the
mostimportantindicatorof probability of false positives,
particularly when accompanied by high spike recoveries.
As described in Chapter 5, samples can be contaminated
during sampling, storage, or analysis. Field and
laboratory blanks identify this problem by determining
the level and point of contamination. Sample matrix
interferences can also cause false positives. High spike
recoveries indicate that matrix interference has occurred.
• The probability of false positives is high if the
chemical of potential concern has been detected in
any blanks. False positives should be suspected
for any sample value less than 5 times the blank
concentration (10 times for common laboratory
contaminants). High spike recoveries combined
with blank contamination increase the likelihood
of false positives.
• The probability of a false positive for an analyte is
directly related to the amount of bias if chemicals
of potential concern are detected in blanks and
spike recoveries for the analyte are biased high.
118
-------
• The probability of false positives is highest when
the reported concentration is near the detection
limit for an analyte.
• The probability of false positives is low if chemicals
of potential concern have not been detected in any
blanks and spike recoveries are not biased high.
6.1.2 Are Site Concentrations
Sufficiently Different from
Background?
Background samples provide baseline measurements to
determine the degree of contamination. Background
samples are collected and analyzed for each medium of
concern in the same manner as other site samples. They
require the same degree of quality control and data
review. Background samples differ from other samples
in that the sampling points, as defined in the sampling
and analysis plan (SAP), are intended to be in an area
that has not been exposed to the source of contamination.
Historical data, when available, are particularly useful
in selecting sampling and analysis techniques used to
determine therepresentadveconcentradons of chemicals
of potential concern in background samples. Historical
data can help to delineate physical areas that are
background and provide a basis for temporal trends in
the concentration of chemicals of potential concern.
Exhibit 72 lists the criteria from the Data Useability
Worksheet that affect this decision.
As pan of the risk assessment process, the risk assessor
must determine if background samples are
uncontaminated. The entire data collection process will
be simplified if chemicals of potential concern are not
found in background samples. If chemicals of potential
concern are found in the background samples, the risk
assessor must determine whether they are at naturally
occurring levels, of anthropogenic origin, due to
contamination during the sampling process, or are site
contaminants.
Both naturally occurring chemicals and anthropogenic
chemicals have significance for risk assessment.
Naturally occurring chemicals are those expected at a
site in the absence of human influence. Metals are
naturally occurring chemicals that are often included in
risk analysis; they are often present in environmental
media in varying concentrations. For example, soils of
high organic content, such as humus, would have a low
concentration of metals by weight, while soils with a
high clay content would contain higher metal levels.
Anthropogenic chemicals are defined in RAGS (EPA
1989a) as chemicals that are present in the environment
due to man-made, non-site sources (e.g., industry,
automobiles). Chemicals of anthropogenic origin may
include organic compounds such as phthalates
(plasticizers), DDT, or polycyclic aromatic hydrocarbons
and inorganic chemicals such as lead (from automobile
exhaust). Guidance highlights for background
concentration issues for risk assessment are:
• Organic chemicals of potential concern found in
background samples should not be considered
naturally occurring. They may be present because
they are either site contaminants or are of
anthropogenic origin. They also could be a result
of contamination during sampling.
• The risk assessor may eliminate chemicals from
risk assessment calculations if their concentrations
fall within naturally ocurring levels and are below
the concentration of concern.
• Contamination ofbackground samples is indicated
if chemical concentrations are higher than naturally
occurring levels. Such contamination may come
EXHIBIT 72. DATA USEABILITY CRITERIA AFFECTING
BACKGROUND LEVEL COMPARISON
Worksheet
Reference
1
2A
3A
6A
6B
60
6E
Data Useability
Criterion
Data Collection and
Evaluation Decision
Reports to risk assessor
Documentation (SAP) and historical data
Data sources (analytical)
Completeness (sampling)
Comparability (analytical)
Precision (analytical)
Accuracy (sampling and analytical)
Are site concentrations
sufficiently different from
background?
21-002-072
119
-------
from anthropogenic sources or from problems in
sampling or analysis activities. The risk assessor
may include analytical data with other site data or
perform a separate risk assessment based on best
professional judgment.
• Anthropogenic chemicals should not be eliminated
from the risk assessment.
• Statistical analysis may be necessary to determine
if site levels are distinctly different from those
found in background samples when background
results approach site concentration levels.
• Statistical analysis may be necessary where
chemicals of potential concern are detected in site
samples at very low concentrations. It is difficult
to distinguish a difference between background
and site sample concentrations at levels close to
the detection limit.
f Statistical analysis may determine if site
concentrations are significantly above
background concentrations when the
differences are not obvious.
6.1.3 Are All Exposure Pathways and
Areas Identified and Examined?
The identification and examination of exposure path ways
is discussed in detail in RAGS. Exhibit 73 summarizes
the criteria that the riskassessor must assess to determine
the probable level of certainty that all exposure pathways
and areas have been identified and examined.
The nature of the exposure pathways and areas to be
examined is critical to the selection of a sampling design
and analytical methods. If the pathways and areas are
not identified properly, the resulting characterization
may be inappropriate. Theriskassessorshoulddetennine
which pathways and areas are not adequately assessed
and determine the effect on the risk assessment if they
are excluded from study. Guidance highlights for
exposure pathway identification for risk assessment
are:
• Recommend acquisition of additional samples
from the inadequately represented exposure
pathway or area if feasible. (Sampling
considerations presented in Chapter 3 should be
re-examined).
• Investigatewnethercomputersimulationmodeling
is feasible if additional samples cannot be collected
from an inadequately represented pathway or area.
For example, air flow models could be used to
estimate transport of volatile contaminants if the
contamination of soil and water at a site is fully
characterized but no air samples were obtained.
• Note in the report that the risk could not be
determined for a pathway or area, or use simple
chemical/physical relationships to estimate
exposure if additional samples cannot be collected
from an inadequately represented pathway and no
simulation models are appropriate. For example,
equilibrium partition coefficients can be used to
estimate movement in the vadose zone of soil if
insufficient data exist to calibrate a groundwater
transport model.
6.1.4 Are All Exposure Areas Fully
Characterized?
Assessing how well exposure areas have been
characterized involves evaluation of completeness,
comparability, and representativeness across analytical
and sampling data quality indicators. Exhibit 74 lists
the criteria from the worksheet that affect this decision.
To be fully characterized, the exposure area must have
EXHIBIT 73. DATA USEABILITY CRITERIA AFFECTING EXPOSURE
PATHWAY AND EXPOSURE AREA EXAMINATION
Worksheet
Reference
1
2A
3B
6A
6B
Data Useability
Criterion
Reports to risk assessor
Documentation (SAP)
Data sources (non-analytical)
Completeness (sampling)
Comparability (sampling)
120
Data Collection and
Evaluation Decision
Are all exposure
pathways and areas
identified and
examined?
21-002X173
-------
been appropriately sampled. Broad spectrum analyses
must also have been conducted for the mediaof concern
and analyte-specific methods used where appropriate.
The uncertainty in data collection and analysis depends
on the evaluation of completeness, comparability and
representativeness as discussed in Section 5.6. Based
on these indicators, the risk assessor should determine
the magnitude of the effect of data confidence on the
riskassessment. Guidancehighlights for characterization
of exposure areas for risk assessment are:
• Use the data but note the level of confidence
associated with assessment of the affected exposure
area if it is not significant.
• Statistical interpretation procedures (e.g.,
sensitivity analysis) may be used if the confidence
level associated with data for an exposure area is
significant but does not warrant resampling and
reanalysis.
• If the uncertainty associated with the data is high,
the risk assessor may determine that an exposure
pathway or area is not fully characterized.
6.2 ASSESSMENT OF UNCERTAINTY
ASSOCIATED WITH THE BASE-
LINE RISK ASSESSMENT FOR
HUMAN HEALTH
The level of certainty in making each of the four
decisions discussed in Section 6.1 contributes to the
overall uncertainty in data collection and analysis
components of risk assessment. The critical factor in
assessing die effect of uncertainty on the en vironmental
analytical data component of risk assessment is not that
uncertainty exists, but rather that the risk assessor is able
to qualify and/or quantitate the uncertainty so that the
decision-maker can make informed decisions. The
certainty levels for risk assessment, represented in
Exhibit 75, are based on the ability to quantitate the
uncertainty in analytical data collection and evaluation.
However, data collection and evaluation is only one
source of uncertainty in the risk assessment. Other
components of the risk assessment process, such as
toxicity of chemicals and exposure assumptions,
influence the four decisions to be made and contribute
significantly to the uncertainty of the baseline risk
assessment.
The most quantitative level of risk assessment occurs
when the uncertainty in data can be determined
quantitatively. The next level occurs when the
uncertainty can be determined qualitatively, or the
impact of the uncertainty is assessed using sensitivity
analysis. The least desirable situation occurs when the
uncertainty in data is unknown. This situation can occur
if the minimum requirements given in Chapter 5 for the
data useability criteria have not been achieved.
*• The primary planning objective is that
uncertainty levels are acceptable, known
and quantitatable, not that uncertainty be
eliminated.
EXHIBIT 74. DATA USEABILITY CRITERIA AFFECTING
EXPOSURE AREA CHARACTERIZATION
Worksheet
Reference
Data Useability
Criterion
Data Collection and
Evaluation Decision
1
2A
2B
2C
3A
3B
6A
6B
6C
6D
Reports to risk assessor
Documentation (SAP)
Documentation (SOPs)
Documentation (field records)
Data sources (analytical)
Data sources (non-analytical)
Completeness (sampling and analytical)
Comparability (sampling and analytical)
Representativeness (sampling and analytical)
Precision (sampling)
Are all exposure areas
fully characterized?
21-002474
121
-------
EXHIBIT 75. UNCERTAINTY IN DATA COLLECTION AND EVALUATION
DECISIONS AFFECTS THE CERTAINTY
OF THE RISK ASSESSMENT
Decisions To
80 Made
What
contamination is
present and at
what levels?
Risk Assessment
Process
Are site
concentrations
sufficiently
different from
background?
Are all exposure
pathways and
areas identified
and examined?
Are all
exposure
areas fully
characterized?
—
••
MM
m^tm
r~
Data Collection
and Evaluation
Exposure
Assessment
Toxicity
Assessment
Risk
Characterization
Nature of Risk
Assessment
Quantitative
(uncertainty
explicitly stated)
Quantitative
(uncertainty not
known)
Qualitative (no
uncertainty
estimate)
21-OOZ-075
122
-------
Appendices
I. DESCRIPTION OF ORGANICS AND INORGANICS DATA REVIEW PACKAGES
II. LISTING OF COMMON POLLUTANTS GENERATED BY SEVEN INDUSTRIES 153
III. LISTING OF ANALYTES, METHODS, AND DETECTION OR QUANTITATION LIMITS FOR
POLLUTANTS OF CONCERN TO RISK ASSESSMENT 167
IV. CALCULATION FORMULAS FOR STATISTICAL EVALUATION 235
V. "J" DATA QUALIFIER SOURCE AND MEANING 239
VI. "R" DATA QUALIFIER SOURCE AND MEANING 245
VII. SUMMARY OF COMMON LABORATORY CONTAMINANTS, CONCENTRATION REQUIRE-
MENTS, AND RISK ASSESSMENT IMPLICATIONS 249
VIII. CLP ANALYTICAL METHODS SHORT SHEETS AND TCL COMPOUNDS 253
IX. EXAMPLE DIAGRAM FOR A CONCEPTUAL MODEL FOR RISK ASSESSMENT 263
123
-------
APPENDIX I
DESCRIPTION OF ORGANICS AND INORGANICS DATA REVIEW PACKAGES
The purpose of Appendix I is to familiarize the reader with a model for data review
deliverables. This appendix consists of the following items:
o A description of the data reporting format,
o An example of a data review summary, and
o Example data review forms.
Please note that the example forms are designed for the validation of Contract
Laboratory Program (CLP) data packages. An example form is included for each analytical
fraction (volatiles, semivolatiles, pesticide/Aroclors and metals) and for samples from
soil/sediment and aqueous matrices. These forms nevertheless include the necessary
information for the review of most types of data (analytical results, sample
quantitation/detection limits, data qualifiers, etc.) not associated with the CLP.
1. DATA REPORTING FORMATS
Whenever an analytical laboratory is requested to analyze field samples for a specific
site, the RPM (in consultation with the technical project team) must ensure that the laboratory
will provide adequate documentation to support all current and future uses of the data.
Potential uses of the data can include data validation, monitoring, modeling, risk assessment,
site characterization, Record of Decision defense, enforcement, and litigation.
Data packages produced by analytical laboratories should contain all the documents that
were produced or used by the laboratory for that particular analysis. The required documents
should include a narrative (detailing the exact method performed, deviations from the method,
problems encountered, and problem resolution), chain-of-custody records, laboratory logbook
pages, and raw data and tabulated summary forms for all standards, quality control and field
samples.
The documents should be organized in a logical manner and the entire data package
should be paginated. Generally, the laboratory should be required to produce a data package
with documents ordered in the following manner:
1) Narrative
2) Tabulated summary forms for laboratory standards and quality control samples
(in chronological order by type of quality control sample/standard by date of
analysis by instrument)
3) Tabulated summary forms for field sample results (in increasing RAS, SAS, or
project sample number order)
4) Raw data for field samples (in increasing RAS, SAS, or project sample number
order)
5) Raw data for laboratory standards and quality control samples (in chronological
order by type of quality control sample/standard by date of analysis by
instrument)
6) Laboratory logbook pages
7) Chain-of-custody records
125
-------
APPENDIX I (continued)
It is often convenient to require that the laboratory data package resemble as closely as
possible the data packages required by the current CLP RAS SOWs for organics and
inorganics, that the tabulated summary forms provided in those SOWs be utilized and modified
appropriately, and that the data qualifiers in those SOWs be applied to the data as appropriate.
The following sections describe specific requirements for the content of each document
contained in the laboratory data package.
NARRATIVE:
A narrative must be provided describing the analytical methods and exact procedures
performed by the laboratory, as well as any deviations from the method. Problems
encountered during analysis, problem resolution and any factors which may affect the validity
of the data must be addressed. The narrative must include the laboratory name and RAS,
SAS, or project sample numbers cross-referenced to the laboratory sample identification
numbers, and must be signed and dated by the laboratory manager.
Any telephone communications between the laboratory and sampling personnel (or other
parties outside of the laboratory) to resolve sampling discrepancies or analytical problems must
be documented in detail on telephone communication logs. Those telephone logs must
explicitly detail the problems requiring resolution, the agreed to resolution, and the names and
affiliations of the communicating parties. All telephone logs must be appended to the
narrative.
An example calculation of a positive hit and a detection/quantitation limit for each type
of sample analysis must be provided. All equations, dilution factors and information required
to reproduce the laboratory results must be provided.
TABULATED SUMMARY FORMS:
Laboratory Standards and Quality Control Samples
Tabulated summary forms must be provided for all laboratory standards, tunes, blanks,
duplicates, spikes, and any other types of laboratory quality control samples/standards. The
tabulated summary forms must contain information pertinent to the type of laboratory quality
control sample/standard which was analyzed. Typical entries include: concentrations spiked,
concentrations detected, spike compound names, results of statistical calculations (%R, %D,
RPD, RSD, CV, RRF, SD, etc.), sample identification numbers, dates/times of analysis,
instrument IDs, lab file IDs, and QC limits.
The exact format of each tabulated summary form will depend on the particular analysis
method requested and the quality control procedures specified in that method. However,
comprehensive tabulated summary forms must be prepared for all quality control
samples/standards analyzed by the laboratory. For example, typical tabulated summary forms
for volatile organics analyses include but are not limited to:
Surrogate results: Tabulate the sample identification numbers, surrogate compounds added,
concentration added, percent recoveries, and QC limits for all standards, blanks, quality
control samples and field samples. Flag outliers.
Matrix spike and matrix spike duplicate results: Tabulate the matrix spike compounds added,
concentration added, percent recoveries and relative percent differences for the spiked
compounds, and QC limits. Flag outliers. List the sample identification numbers. Results for
126
-------
APPENDIX I (continued)
all non-spike compounds must be tabulated on the form used to summarize field sample
results.
Method/laboratory blanks: Tabulate the sample identification numbers, lab file IDs, and time
analyzed for field samples and matrix spike samples which pertain to each blank on a separate
form. The form must also contain the GC column, instrument ID, laboratory sample
identification number, lab file ID, and date/time of analysis for the blank itself. Results for
each blank must also be tabulated on the form used to summarize field sample results.
Tuning results: Tabulate the m/e, ion abundance criteria, and percent relative abundances and
list the tune compound name, instrument ID, lab file ID, and date/time of injection which
pertain to each tune analysis on a separate form. The form must also contain tabulated sample
identification numbers, lab file IDs, and date/time of analysis for all field samples, matrix
spike samples, blanks, and standards which pertain to that tune. Flag outliers.
Initial calibration results: Tabulate the target compound names, relative response factors for
each target and surrogate compound at each standard concentration, mean relative response
factors and percent relative standard deviations for all target and surrogate compounds, and
QC limits for each initial calibration on a separate form. The form must also contain the
concentration of the calibration standards, instrument ID, lab file IDs, and dates/times of
standard analyses for that initial calibration. Flag outliers.
Continuing calibration results: Tabulate the target compound names, mean relative response
factors from initial calibration, relative response factors from continuing calibration, percent
differences, and QC limits for all target and surrogate compounds for each continuing
calibration on a separate form. The form must also contain the concentration of the
continuing calibration standard, instrument ID, lab file ID, and dates/times of initial and
continuing calibration standard analyses which pertain to that continuing calibration. Flag
outliers.
Internal standard results: Tabulate the sample identification numbers, internal standard
compound names, QC limits, retention times and area counts of the quantitation ion for each
internal standard compound in the continuing calibration standard and all field samples,
matrix spike samples, and blanks which pertain to that continuing calibration on a separate
form. The form must also contain the instrument ID, lab file ID, and date/time of continuing
calibration standard analysis. Flag outliers.
MDL study results: Tabulate the target compound names, concentrations spiked and detected
for each MDL spike analysis, and the standard deviation and calculated MDL for each target
compound. (Note: The narrative must explain the MDL procedure utilized to generate the
values. The formula and associated constant values utilized in the calculation of the MDL for
each analyte must be provided. The column, instrument ID, trap composition, and operating
conditions must be clearly displayed on the raw data.)
Field Samples
The exact format of the tabulated summary form for each field sample will depend on the
particular analysis method requested. However, comprehensive tabulated summary forms must
be prepared for each field sample analyzed by the laboratory. At a minimum, the target
compound names, concentration units, positive hits and numerical detection/quantitation limits
and any laboratory qualifier flags for each target compound must be tabulated on a separate
form. Definitions must be provided for all qualifier flags used by the laboratory. For each
127
-------
APPENDIX I (continued)
sample, the tabulated form must also contain the RAS, SAS, or project sample identification
number, laboratory name, laboratory sample ID, lab file ID, sample matrix type, and level of
analysis (low, medium, high). The percent moisture/solids, weights and volumes of sample
prepared/purged/extracted/digested/analyzed, initial and final extract/digest and extract
clean-up volumes, injection volume, clean-ups performed, dilution factor, measured pH, and
dates that sample was received/extracted/digested/analyzed should be included as appropriate
to the analysis method.
RAW DATA:
Raw data must be provided by the laboratory for all laboratory quality control samples,
blanks, spikes, duplicates, standards, and field samples. The exact format and content of the
raw data will depend on the particular analysis method requested. However, any and all
instrument printouts, strip chart recordings, chromatograms, quantitation reports, mass spectra
and other types of raw data generated by the laboratory for a particular project must be
provided in the data package. Typical raw data for organic GC/MS analyses includes but is
not limited to:
o Reconstructed total ion chromatograms,
o Instrument quantitation reports containing the following information:
laboratory sample identification number, RAS, SAS or project sample number,
date and time of analysis, RT and/or scan number of quantitation ion with
measured area, analyte concentration, copy of area table from data system,
GC/MS instrument ID, lab file ID, column, trap composition, and operating
conditions,
o Raw and enhanced mass spectra for all positive field sample results and daily
continuing calibration standard reference spectra for all positive field sample
results,
o Mass spectra and three library searched best-match mass spectra for all
tentatively identified compounds reported, and
o Instrument normalized mass listing and the mass spectrum for each tune.
Typical raw data for inorganic analyses includes but is not limited to:
o Instrument printouts and strip chart recordings containing the following
information: laboratory sample identification number, RAS, SAS or project
sample number, date and time of analysis, absorbance/emissions values, analyte
concentration, instrument ID, lab file ID, and operating conditions, and
o Standard curve raw data, plotted standard curves, linear regression equations,
and correlation coefficients.
LABORATORY LOGBOOK PAGES:
Copies of standards preparation logs, sample preparation/extraction/digestion logs,
sample analysis run logs, personal logs, and any hand written project-specific notes must be
included. The initial and final volumes of sample prepared/purged/extracted/digested, initial
128
-------
APPENDIX I (continued)
and final extract/digest and extract clean-up volumes, injection volumes, and dilution factors
must be clearly labelled.
CHAIN-OF-CUSTODY RECORDS:
All chain-of-custody records provided to the laboratory during sample shipment or
generated by the laboratory during sample receipt, storage, preparation, and analysis must be
included. Chain-of-custody records include but are not limited to: signed and dated field
chain-of-custody forms, signed and dated shipping airbills, sample tags, SAS packing lists,
RAS Traffic Reports, internal laboratory receiving records, and internal laboratory
sample/extract/digest transfer records.
129
-------
APPENDIX I (Continued)
2. DATA REVIEW SUMMARY
ORGANIC DATA SUMMARY FORMS UTILIZED
BY REGION III IN THE CLP
DATE:
SUBJECT:
FROM:
TO:
THRU:
OVERVIEW
Case consisted of four (4) low level water and two (2) low
level soil samples, submitted for -full organic analyses.. Included
in this data set was one (1) equipment blank and one (1} trip
blank. The trip blank was analyzed for volatiles only. The
samples were analyzed as a Contract Laboratory Program (CLP)
Routine Analytical Service (RAS).
SUMMARY
All samples were successfully analyzed for all target compounds
with the exception of 2-Butanone and 2-Hexanone in the volatile
fraction. All remaining instrument and method sensitivities were
according to the Contract Laboratory Program (CLP) Routine
Analytical Service (RAS) protocol.
MAJOR PROBLEM
The response factors (RF) for 2-Butanone and 2-Hexanone were less
than O.O5 in one of the continuing volatile calibration. The
quantitation limits for this compound in the affected samples
were qualified unreliable, "R". (See Table I in Appendix F for
the affected samples.)
MINOR PROBLEMS
Several compounds failed precision criteria for initial and/or
continuing,.calibrations. Quantitation limits and the reported
results for these compounds may be biased and, therefore, have
been qualified estimated, "UJ" and " J", respectively. (See Table
I in Appendix F for the affected samples).
130
-------
APPENDIX I (Continued)
2. DATA REVIEW SUMMARY
Page 2 of 3
The soil semivolatile MS/USD analyses were originally
extracted within the technical and contractual holding
times. Re-extractions were required because of surrogate
recoveries, and these re-extractions were performed outside
of holding times. Surrogate recoveries were again outside
of the QC limits, therefore, original sample results are
being reported.
The maximum concentration of compounds found in the trip
blanks, field blanks, or method blanks are listed below.
All samples with concentrations of common laboratory
contaminants less than ten times (<10X) the blank
concentration, and uncommon laboratory contaminants less
than five times (
-------
APPENDIX I (Continued)
2. DATA REVIEW SUMMARY
Page 3 of 3
o The pesticide/PCB analyses of all soil samples and associated
QC samples had surrogate recoveries in excess of the QC limit.
Since no positive results were reported for any pesticide or
. PCB compounds for any of the samples in this case no data was
affected. (See Appendix F) .
o The reported Tentatively Identified Compcunds (TIC's) in
Appendix D have been reviewed and accepted or corrected.
o All data for Case were reviewed in accordance with the
Functional Guidelines for Evaluating Organic Analyses with
modifications for use within Region III. The text of this
report addresses only those problems affecting usability.
ATTACHMENTS
APPENDIX A - Glossary of Data Qualifiers
APPENDIX B - Data Summary. These 'include:
(a) All positive results for target compounds with
qualifier codes where applicable.
(b) All unusable detection limits (qualified "R").
APPENDIX C - Results as Reported by the Laboratory for All
Target Compounds
APPENDIX D - Reviewed and Corrected Tentatively Identified
Compounds
APPENDIX E - Organic Regional Data Assessment Sumnary
APPENDIX F - Support Documentation
132
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
3
i
1
§
c
*^
•
1
g-
e
i/t
1
I
4_
Of
U
*
—
M
I
t
41
0
O
c
I
s»
V
0
a
—
I
<
_J
O»
•^
O
«*
41
^
^
<
U
C
u
>
3gO— -ggj) — t/ >> — 5 _5j o
fc H _ E "^ e**ii «*^^ ^ ** "? t TI
S*^ V *^ 2K P U O *0 A C ^ COO ^ i 4* > -«~ — <0 u C
*^ ** *• L ^) _^ 4_ .O S O <0 Ol C ^ O ^ — ' ^ T3 * C C **
0 •
3 '"D
— > O O
J <- X
41 O C
•° t §<
£8oS
t u > z
<
UJ
^
£ < 0
41 < Q »-
i < o a —
E
— «• « > E
<• u — —
- C 0
_ U. —
•< «*. a. u u
e —
O tit
*- a M
J= CX
o «i a — -
0 — vt < C
*- _
1 i - 2 c
y; a. u «Q
Q. i_ «-»-«-
E o o o o
«rt
133
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
w»
i
?
I
s
o
z
o
M ^
«J 5
c
o
I
1
1
1
VI
1
1
€
•
.AC
1
D
O»
C
1
«>
•
o
M
C
<
•S
O
C
u
I.
41
0.
^
^
*^
o
II)
o
1
u
c
a
>
§£« J66§ Is. IS S§ .
CO— -£— 35—- ** C. «§Ct-— ••— t-P — — T)
liilfjfliillllliliisilil!
X
0
C
2
1
*^
o
o
1
1 .
>c •*
O "
a
™" IT'S
~> a> t_ ro
01- >•
O a» — •
•S £ o5
13 i: S*
0 C — »*
j; o "o o
•- a > x
*
u
.C < 0
.5s^2"i
*> u a. o — • •
^ u. a. 2 o
M *
§ I.
£ C >-
a • » —
0— • « C
3 «• >-< •
If l||
• a. «- •
- 010 3 >.
o a: — *. o
Q. <->-<->-
134
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
f
s
te
_i
3
O
I
u
i
§
u
I
••
S
p
3
z
ktf
L
u
».
*•
M
M
i
«#
.*
a
z
1
a
ae
u
o
c
c
o
S
rM ni o «- o o o o
m m in tn
>
S E • C
eQ~g-33— •-«. 1 ? X_"S J u e J J! *
3*'«ik^v^^4iLo0aiCko**«<<^*n^c ^£
-*C^L
or
V
3 °>
«l 0
— t.
U "o
2 S
w o
§ X
u "
X —
1 3-
1 5
**
£ C
5 I
«* "D
— M
£ §
O "-
"O ** "
•1 •— «>
c ** a
0 — ^2
Z e **
o — .
0 J |
i si :
_J — O
— a i.
— a • o
x «- a.
o — » u
&c
• a ** o
-•S51 -
— u o — * «i
§— » Q. C *D —
«* E O 4» 3
*- vt ** >. O -
** •«— "O ** a ~ £
3 n u D o a
= Z
< u
111-
• 0 — > —
1 2^5.
— i. a. — • — •
« 3 0 O O
C >
< x. a. u «J
135
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
u
2
5 =
a
o
IB
5
g
J
3
_j
Cf
^
9
W>
j_
at
4f
1
4A
|
j
U
a
«»
qf
u
•
»_
M
Jtf
|
«
«
-o
o
C
1
VI
_*
O
oc
u
«n
«
>
4B
C
«B
(j
1
O*
|
oooointnooomomoinoJOOinoooooo
o ^> •- o o v- m *M o o •- * •* o •— o •- in oj •-
r\j rsi o ^ o O o O
m «n in in
>
SE «• E
:, co - - -§ 3j5s §§
co— e— 33— *•«- HCU— • «• — <-E — — "°
^ E C 9 — • .^ g ^^ «. 0034tMCw3 — • T3 '^
S— «— x_£oo0Q.CT>Cc»u.M0«i> — -><«oc
•* «• i. t. -B — i.3tLO»a»ci.uf— • — -a « c c -
<<irt»->fMCi
2
2 I
•~ c.
>
« —
u o
u
o c
<- 0
«* u
S X
u *•»
X —
** «
r S-
1 2
**
2 c
«« •—
1 1
u •*-
' i
I 2
^ C
g £
•Q ** •
** •»
S « S
£ 1 •"
*« •!" *rf
a . — .e
- o J
J - S
0 1 t
w . TJ
w «l
O •*• ** 0
1 5| g
£ gS "S
E ** CL «-
i « « e
X ** O.
O -- « W
& .iz i;
a T) 3 — i-
V O E w
M ** —
._ u o — o>
§-^ O. C TJ —
•1 6 O «» 3
— l_ A •— M M
u .«<
a «• « — • i.
— ~ T3— C «>
c D — c * a
a — > f
3 • » D o o
-> « 3 -» •<
3 z
«c u
-5 * c.T
O VQ»
^: 6 Q. 41
•• « a "0 e
« o — > —
1 S^-o fe
— ' i- a. — —
« 3 O O O
U u. — O o
^
< u. a. o o
136
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
s
e
c
<
« -^
I- O
u
^ c
t- o
I!
«
f C
** •—
Ir
««
• o
*• •*
.— §
Isu
111
T3 a — T3
a> E o
~ C C 3
C « C —
O 3 3 O
u o a >
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
Local ion
I
z
•i
Report Number
u
c
c_
0
V
Volatil
2222-2 2—2 22
C" O «*
« C C
«w o o a
O C Q. ,c
•— O 0 O 0 *-
-^ COOG.OC«-0 41
0 T3 COO C*3 Q. C O
O *V «* «* O JC— « C «- • .C O O u.
•0 CCC C *-» «- jC « Q. jC ~ *- C €»O
— o «» a o a o O <- Q. O *- o o a C —
c. "o jc .c JE .c o— o o *- o o o — • *- «» jc
o o — •-*-•** *- L.JI E «. o C E L. jc c .cu ^^
o -o — *-*«««* o o o » o a— aoo u o •"« o —
c o — o ^: — o o o o -»o**co£jc«-— — a «*i- c« *
f«O« «lOOOeO «*«**•*— O— O — O • IM«ft.«l M€* O
".3 £ — £ o — — — — « — -c — £*.e — a o jc — PIE*CO*- c« »^
5S°o— cc — — — o— «• c *D— •— E« c i •— x: « o - c o 2 c *
»_ S — »_>.OOOOO*-O^-— O— §0 — JCOrw4»tnQ«-K«*M»U« -WC
o£>-<'f*'A* • • O • 13 -.U >- E • • O «_ *MCE4f«l. - D O >•»-«*
jC t_ — £ O O « * » - JC *• » • — t. .^-c.— *«(«_«_• i «| - O £**«-•>*
3
• o
> t-
41
c. —
o
O •"
**
ir
-£
•3 —
•2S
if Identlfll
tatlont Idi
5 J
41
*• O
-J«
.J •-»
gll
— X
w 41 O
0 O «
41 O O 4*
ae — — «-
*• C C 3
C « « — •
0330
—I
ae -^
cj -» 3 oe
138
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
3
5
u
§
3
c «
c a
a .e
w &-•_!. t. ** fc. .C W U. .C t' U
— «ciati a «o ~ O. o •- t> O
•- "O.c^:.c ^ o —• uoc.«t«o -«
• o ^«»*»«* «« ^^ E*-OCei- ^
•o — «-««« « ou«io a— «• o o u
S ••-•.£ ~"OOO O *'««'(_OfJ£l • >^
i.cu 3 «- t_ t- >. £i.aoi-u<->oj: o
o o
c —
«l £
I
O "
<- tl
o +-
«>-—«> c
3 O X <- «
139
-------
APPENDIX 1 (Continued)
3. DATA REVIEW FORM
3
3
O
3C
(9
§
0.
^
u
8.
c
a
o>
t_
o
3.
I =
~ I-S S.I
••« .£— «
CCC C ««c. .c.
... • .0 »
«-
O
—
jC
go W-Q.O" 5. o a c —
— «OU.«O — w tuc
. •a — »-«•«« . ou.o Q.— .So "5 . '5 So . —
co— «.e —ooo o — « w >. o j= £ i—• — a ut. c. - c u 3>.Lkb fkaokuwoj: o • o ~ .c •-
5 £ — £ • — — — —«-— c— «••£ — ooS— m e *^ c o H- c~ *-
!|f K8"lllsli?sf|Ii-IiX5I*tSlrf2l5Sz
-
1 -- f.UM >..f - . . •
« c
II
X.
k, **
go
- ei
coo
s-1-
O « M
fc* ••-
iis
140
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
p
fl !
a 2
? e
I i
«M
i
1
O O O O O O OOOOOOOOOOOOOOOOOOOOOOOOOOO
5 S
« -
*» E
^ ?S S 2 §•
.*' «•«•-' Q.JC a.ot.M
II
— OOCC
OO « —
_ .
2 *» — o o >-o •;«'»• «• o
c jr*iMt« v*«0
« MUKI^ C«M I MZ3E X
JC •— i . . • . . — i , «l
o • u i- a i. — i. >._ i- t- i. o x^i c
*^>f\J>^ij;OfO£.Ci « OUJ= Q.'^
o ^ o ~r « — o — to-ou^— «-i no
N • «£ .c a f ti a • «.e — « c t
C «>T — >-«i«i i •) . .1 i~-0 -
L. O
t_
—rf «^
Sg
s°
O X
u «^
-I
!J
TJ U
.22
«^
« c
£.2
2S
4^
II
J **
I
" «l O
U ** ** •—
o 3 3
141
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
ii
B
i
g
s
I
f_
5
V)
L.
|
z
8
V
%
0
•
t
VI
oooooooooooooooooooooooooooooooo
c- — 0
•I O u —
£ C 4> <=
«* 0 .C _ 41
•>£«•< e a c
C >•£>*• 4f *«^ f**CC >«4>C
41 W £«C •* 4D»-4f Q.A4I4I Q.U4I
S4> « a. o x.c co — • aco —10^4- -o <_ >.
3 *- • 6 C o. o c a .cuu X£CC - cm £ — c « ~ <0>- 4<4;».t.CK>c8.
-4I- 4« Q. 41 -C «• (L>-f £Q.OO4I -^"C X — . — « — >. *£ •
^ -^ ** rC 3 ** 5 O^ ^*-o^r*-oLcV4z •*<:• ^V'^**-^"^ «c
0£COOC^l>COCOO^£*-*u3C 4l**-*'<-*4O*'OU4iau « UC*'1 O«I<\J» OOOCCO
, u • . — -— . — . . • , 4I4I.CC — — X.3 -4>£~~4l4»4lC— 4»
X
41
•i ~>
4* 41
<- O
c-
"o c
c- O
«w u
c
0 >.
U "
~~n
- 8-
^_c
dent If led dur
tlont (dentil
| j
a o
;!!
EOT)
— *• 41
sli
— X
" 4) O
° ** O.
u_ a
o 6 f
— ^- -^ o
oe v — c.
w c a »
£^^o
- C C 3
o
« ^ •! «
142
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
i
&
$
I
ssssssss
OOOOooo o
a
•o
C
O
OL
«•—
— 3 o a «
— • l-C^'O O«- rtuiu-*ui<«zuj a a»
E o-p
— »• £
—j **
§11
7:« o
u ** *.» —
u o o >
143
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
i
1
1
?
e
u
dl
i
o
tm
**
§
u
X
I
• c
i*
SI
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
1
I
y
o
g
3
ta
g
Bi
I
c
•8
amp
j.
o
j
a
3
3
-:i
o
°4N
r-SJ
5— i-
- 3 **
O « »
-i. U
I»H
— • cc
-« e
— £
*•—
CL>»
oa.
•- o
a. i.
-
— x— c
o 0041
£ £il
« .
u— owcconc
S-g.5 ---•ii'g.i oSgCio-o^ccot-ieouo •-.£>.;:
oo2jc«2 — u—-»>ocoO.t!™o2T— o o o — o ~ II <5 o — '£.t^
• j. o u «x. >.o— « i- ?6 oT-g t « £ — 2 x-i t £ o x£ c
--OJO--— •^w*ffM^:k£AOk<'-^AiM*i£OfO££t • oi..ca.<—
9^^—*OO XO*^^^*<*« U O.C*«O O%^O^M-^ O-^«-u^lA—•*•** CO
U W-^ CfMZ MXZ X*' OZ^ C M^fMO-U X«JZ X^Y^rCJZ £ «I*O
J=— • --«-•— !•«>—«•• . « _ . .«,«!..«.. , , _0-
o.j3«>j----j ryio>^-z%rzfM<\jo
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
3
q
C
y
i
s
e
i
I
•I
I
I
£
•o
i
lyySMl Sic
.11
C TJ TI
o
— D
•- «t
C C —«^6boii
«^« o *^ •—
c'o'SiS.c'So'ap
§
U
I
I
SS
£J
T> —
82
8.S
§5 .
a35
5J1
c o «»
' o :
-» i
3
146
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
s
I
I
<
II
e
c
I
a
a
>.
•
V V
«, «;
C C
o o
o u
g£
II
6 O O
Cl •
T) a
o
o o «r
o a o >
fOKiMr*irofi»ai
tooooooooooooooo
IIOin-OK>«rOIO»»»MMf»lKlM'OKlV>MMM
Q. >•
f^ O €1 • O CX
— C C C I- O
x o «< • a «-
j: MM M o Q.O
w c C C •« • C
>.M
— X — . C
0 OO
CQ.CCV
tt t— i 47 ,C 41 C
^j — > — . o.^. a— *-» a>
4t *^ 0 O «iT9f«OOO(0 <0 3
-s •'•s*- c«*2— ""2 o— 41 —
—"• OG.O — wv— ooo^c«co
*-COO£OCi-C— **4» O — ' ^ t- O — ' — -DE.C>*— — -t — ,CO*«-*
0«C-t.Ol.4i04f-D«*C C X---- OU£ *. — i .*S.O^:]cQ.— w— 0
— JI 0 0 0 0 X - ^.04104.^0—OuC COrtOOOU m — £ X U
^ a.— — — — o..ea.o «- M c ^ «^ « ^ -< •«- « « u • c *_ — — c c a.^ ^
oo^^^jr-o— wocoTLg S^u— ooo- o<- i. o « — "5 .-
• ^UU OX* XO~*4t«_OEo> 0«-«»_— l_X- >•-»-(. OXjCC
— «MO — — — -— ji r«j .c u^jQ o »- -^ — rvj — . £ o^ o ^^ « T o t- £ a.--
— -
147
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
t
d
te Location
>.
«
I
D
t.
\
te
u
<•»
w
«t
M
C.
i
OK
«
O
£
_••
1
VI
3
§
LJ
<0
k
UJ
o
M
X
c
3
oc
u
•JJ
I
Cl
5
I
3OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
1 1 fc 2
5 "S. c Z c .2 £ «
« s |lftf_ ^ =!« lilll III
§3 ** • EC OL V C « fflto X£CC of*»
S«'««^ WI^*MV f *«^v x*^9a4ii*'V
4'3c2"~'>" ~ "J1 i^ C 'S. ** i01" •'^'-•-CtrtCQ.
c^+'j£S»'je%. *c»8fbOkC^f M^M f'xCC Q.--"'*1-^:"
'9o*'Coje-'L.wo*'Ou.> «- c — ? ^ S <>j • <><><>cc^>
zo^z-o^^o^at-o^accDxc**— • SI-«-MC<-«»' c c c TJ i c
• 2J
— ' "3
o <-
cS
o u
ft
41 -C
*«
If led during
identified Ir
II
•> «
S3
12
= 3
. o-i
— «>
E o —
j-i—
X
c « o
tia
8x"
** 2 —
•g-ii
*- •» — w
§•§§•?
ae — -^ t-
*• a a w
a — —
t. w ** O
*- C C D
O 3 D a
0 "* 3 "
ac
u
148
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
I
>
4
U
_J
0-
5
VI
u
3
1
t_
u
u
««>«.««.02^2;o«o:oggoggooooo
a •— «
•D x — o v
co — c c
— a. — 3 one
^«- i-C C C — «> o o o cj JO ri ri e(j rS
CDXCOtt^ £"<«-O -^O-^O>- OUO<-l>t-u4-t.l.
^r «**£*- «-*-o— - «-O-o-fc.^c4uuuuuuu
a.<->— •eo.T3Q.TJ4i^»-D-o-*'n~r— -na.Exooooooo
X
a> X
O «
t. &.
§0
0 <-
5-1
^
?l
«_
«-* **
C C
II
a a
600
w o «
a ** *•
"el
c Ul
000
"g ™ * tj
(- «B M w
S-c c-2.
o o o o
oc — — u
*-* a « «i
u ** ** —
a — —
*- C c 3
C o c —
U O O >
-j -» ^ ec
O 3
LJ
149
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
o
•a
§
3
i
I
I 5 g
1
I =
•§„
IJ
•g
II
^•••
— C C C
f •
-
a.
~ I
— >< — c
— 4»
I i?o
c "S.C c w we
«
c
s
— xjoooo^j — ^.
f B-— — — — • TL^ a.
U O JC f<£ — U—
— .c >•
Q. j: *^
— ~- rv—
q.^«M —«-«^-cx^34r zzz—rutvi'oalB «\T-^
.
. c i- -- - — c c Q. j: *^
0— OI-1.0
C
•o
gl .
HI
Wl O >
150
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
^
o
UJ
*1
<
•
I
«
1
B
§
«
1
2
*
i
<
&
<
c
tr
f
u
S
1
u
_J
u
(\J
•o
l\l
a.
i
t—
i
I
'^
i
i
M
<£
^
£
0
<>
g
Hi
u
c
•
8
«l
1
A
(Sample Number
u
j)
2
(Traffic Repor
(Remarks
«
«
o
o
c
1
A
<_
*r*
U
«
U.
g
M
3
5
•o
^S
«A
£
u
k
«l
a.
"g
0
j
«
•*
•
"o
>
1
•A
.
i- — «
•1 O «. w
f C • «
*• • JC _ «l
— O.C« Cf tl«
c >gx « «- T> «wcc x«c
O t> « Q. « XwC CO — « C »> — > • f u T3I-X
C 3fi ECQ.OC « ^X" XJCCC U£U
c f — o — • x CN«->Cf *> £Oi- vfc-ccmco.
— •U.QCwaC — • O. X « 1L 0 £«) &l-f £a.OO«< — o— ci o c o. c —ow — . 5 3 <_ <\j ^~-
— •t-wi.i.wf — k-t>«ot.«ia-r«i >-— c x— — ~ x -.e -
CfW£3V>£@. CVO£I-OC.C X£ N f a f XH- H. a.<- -£.
Ofcooc— i-cocoo — ^: ^ o 3 c v— «Cu/uxij<«) -i a>
<- «i— "- " — x o • «. — I.IJBU c « i « «i2o<^tiT o t. C. C- o M ^
•'•OWCOJI— l-*.o».Bo««t-i I. C — • OMCKf OOOCCO
— Ci-jjjli.^g— •~-€.««CjCCO«lX- MXvCMMM««M
»g«»iae£>»«>«J3aE'OaeBKC«>^t 3 i. » n c L. » . ccc-ojic
io
21
• ^
i.
§
o
x
*«
1
«
**
1
3
S
6
T3 —
§0
~
II
as
J> 9
' 0
:&
*• c
1 41
_J •
C •»
^±•0'
-1^
O — *^
r-si
*- C "-*
cot*
3 *^
o « «
4^ — •
•z«
S3
3 «
IA O >
3"
151
-------
APPENDIX I (Continued)
3. DATA REVIEW FORM
|
I
g
§
=
i I
a
x
§
u
£
I
8.2
21
«lc
<- o
:&
~ c
c o
**
II
-------
APPENDIX II
LISTING OF COMMON POLLUTANTS GENERATED BY SEVEN INDUSTRIES
Appendix II identifies seven industries that generate waste which contains pollutants that
are known to pose human and environmental hazards. This appendix is intended to aid the
reader in three ways:
o To assist in the identification of target compounds and potential exposure pathways.
o To predict associated contaminants that potentially yield interferences.
o To assist in early identification of sites that contain high levels of compounds that
may not be included as target analytes for routinely available methods.
The data for these tables were obtained by searching the USEPA Toxic Release Inventory
System using the Standard Industrial Classification (SIC) codes listed below:
Industry SIC Code
1 Battery Recycling 3691, 3692
2 Munitions/Explosives 2892
3 Pesticides Manufacturing 2842, 2879
4 Electroplating 3471
5 Wood Preservatives 2491
6 Leather Tanning 3111
7 Petroleum Refining 2911
The appendix consists of seven tables and depicts the pollutants associated with each of
the seven industries, the CAS number of each pollutant, and the matrices where each pollutant
has been found. The list is not inclusive of ail pollutants or industrial sources. The seven
industries were selected based on the recommendation of the Risk Assessment Subgroup of the
Data Useability Workgroup because of the frequency of occurrence of the pollutants produced
by those industries in Superfund sites.
153
-------
jo!
I
•g
I
I
I*
I* 5
II
o o
» I
154
-------
>.»-
2r*cs«r~*
-------
I
I
III
01:
j"
juuzl
SBg*
^"•J^sSSS
3H*IiS*SS'
2iS25=
caa:aull]ZQr-fr--JuiiJUzlu:3-
QVOzqHZOgzgSS^SQzua^
B^i^ZSjBOOSQSPSaQsSx • 5-8§QH^Qfl§
5^hu"-'Q-J5-o:S2o2-j9l3x-.a>.c9.jC£:.32
oot;S-?-:i:>-aEH"O'<=riww-.HJ«.s
-------
.s
sssss-s—sRsa
SSglligSSSSRpS
r»-»Aor>r>o«~«5
-------
I
s
I
3
I
I
>• >•>•>•
> 5-
si 1
aigg
in
u>
5 o "•*§ w£2 p 2
SU 3 g§§ ||g g j|gz
3§3eiii9i §li
MdZ
^^•>•r!
s!S
° O
158
-------
B
1
B:
liii iisls silii|||3i|iiiljj8gi
oi2|o2|g:g3Q3{nig^^ii^S|>|l3g25|^
ll|l 5
ti§* "
159
-------
I
Compound
1
at.
II
It
II
£ 3
•5 2
& 8
1
oc
160
-------
Pr->f-
o —
fgsSSilBs^lgSuSjlSgS
5sfe&ai2s2-:pS5iv?v
161
-------
I
>•>•»• s-
§o\r*r»f
ss;:
Dio55H«25
-------
I
w
I-
1
at.
Q3?5j J:Jz[3ou{3Kfcorio*fcffl5ixSu3aci£G3<
g^ssg^slggltsi&^iissl^siig^gsSIsgS
163
-------
I
.s
I
I
1
et
8
;ui
e-z
is;
Ul
'tii
164
-------
i i
(11
in
?H £9
2S oxgj
165
-------
APPENDIX III
LISTING OF ANALVTES, METHODS, AND DETECTION OR QUANTITATION LIMITS
FOR POLLUTANTS OF CONCERN TO RISK ASSESSMENT
The purpose of this appendix is to familiarize the reader with the variety of EPA
methods that are available for analysis of pollutants of concern in risk assessment. The
appendix facilitates appropriate method selection for pollutants in the matrix of interest.
Appendix III consists first of a summary of definitions of commonly used detection
limits and quantitation limits. Tables I, II, and III depict detection limit estimates achievable
for 33 organic and inorganic pollutants of potential concern to risk assessment in air, soil, and
water matrices respectively. The detection limits listed herein are provided for guidance and
may not always be achievable. Specific quantitation limits are highly matrix-dependent.
Table IV provides a summary of each method of analysis for these pollutants. The 33
pollutants listed were chosen because they are highly toxic and/or have reported cancer risks,
and occur at a frequency of greater than 2% in 141 National Priorities List (NPL) sites.*
Tables V-A and V-B provide an additional comparison of analytical methodologies for
selected organic compound classes and inorganic analytes including method detection ranges
and the applicable analytical system and preparation procedures.
*Source: CLP Statistical Database (STAT).
167
-------
APPENDIX III
GLOSSARY
Instrumentation
CVAA =
ECD =
ELCD =
FID =
FLAME =
Fluor =
FPD =
GC«
GC-MS =
GFAA =
HPLC-
HYDAA =
ICP =
LC =
MS =
NPD«
PIDs
UV =
Cold Vapor Atomic Absorption
Electron Capture Detector
Electrolytic Conductivity Detector
Flame lonization Detector
Flame Atomic Absorption
Fluorescence
Flame Photometric Detector
Gas Chromatography
Gas Chrornatography-Mass Speetrometry
Graphite Furnace Atomic Absorption
High Pressure Liquid Chromatography
Hydride Atomic Absorption
Inductively Coupled Plasma
Liquid Chromatography
Mass Speetrometry
Nitrogen/Phosphorus Detector
Photoionization Detector
Ultraviolet
QuantHation/Detection Limits
CRDL s Contract Required Detection Limit
CRQL = Contract Required Ouantrtation Limit
EDL * Estimated Detection Limit
MDL = Method Detection Limit
NA« Not Available
POL * Practical Quantrtation Limit
Methode/Sample Preparation
CLP SOW
Dl
EPA
EPA AIR
EPADW
EP Extracts
MCAWW
QTM
SDDC
SMEWW
SW846
TO
XTN
3510
3540
3550
5030
Contract Laboratory Program Statement of Work
Direct injection of liquid samples; solid samples mixed, then injected
Guidelines Establishing Test Procedures for the Analysis of Pollutants
under the dean Water Act
Compendium of Methods for the Determination of Toxic Organic
Compounds in Ambient Air
Methods for the Determination of Organic Compounds in Drinking Water
Extraction procedure toxicity test extracts
Methods for Chemical Analysis of Water and Wastes
Quick Turnaround Method
Silver diethyldithiocarbamate
Standard Methods for the Examination of Water and Wastewater
Test Methods for Evaluating Solid Waste
Toxic organic
Extraction methods that could be used include 3510,3520, 3540 and 3550
Separatory Funnel Extraction of Liquid Samples
Soxhlet Extraction of Solid Samples
Sonication Extraction of Solid Samples
Purge and Trap
168
-------
Z S
2 J
H z
=> tu
O-Q
H
UJ
— U
g
u
S
3
o
H
Z O <
< U U
C
O
Q
a
g.
Q
U
o
eo
c
A
II
a
Ul
Q
o
<
z
z
<
z
CO
U
o
CO
U
O
CO
6
o
u
O
o
v
i
•a
UJ
a
o a
"%S
cLr-
S
Q m
0. Wl
u
o
S
u
I
.!
_ m
— p-
169
-------
H
z 2
2 3
P
cxo
tu
2
s <
— o
o
o
X
Hi
U
Ul
o
I
01
_
< 2
Z O
< U
<
U
<
z
E
I
U)
<
Z
CO
O
O
o
O
a8
U. (I]
uu
o o
CO
U
O
CO
I
CO
li
U U
O O
CO
CO
6
o
o e -•
fils
u
•o
|3
5 5-
170
-------
z S
2 3
II
z m
< H
D U]
CXQ
z
U)
«g
H O
E
t
o
II
^
a
1°
Q
U]
CO
U
O
uu
o o
oo
s
u
O
28
u. w
uu
o o
CO
0
o
CO
2
6
<
z
a8
SS
o o
CO
U
o
171
-------
z 5
2 J
j- z
H 2
£u
^* UJ
3£
OQ
U]
i
ES
TR
138
g g CO
U < 06 SJ
O. H O
2s S es
^ Cfl <
5
LU
Z *
-- A O3
fjM «^- ^
^°§
j 2j z
< 2 eo
Z O <
< u o
172
-------
id
i
g
in
a
o
x
H
(-
z i
QUANTITATIO
DETECTION LI
Z
INSTRUME
ATION
METHOD
u.
O
U]
u
H
P
UJ
0
lETHOD REFERE
^
U)
< i*^
2 UJ
~~ 03
m Z 5
t O "^
j S z
< S t/j
z o <
< u u
oo
q
ts
II
Q
a:
U
0,
y
u.
O
1
M
"S3
•5"
^
w
1
f Work for Inori
o
•g
3j
ts
vi g
sl
(W METHOD INO
ledia, Multi-Concei
O 2
CO >
dl
co
(J
Z
^
o w S
ON C ^^'
7 *•• ^
^* l^
00
"So
6
II
_J
a
1
0
.1
^
t-
s
3
2
<
*
T3
JS
^3
js
$
1
ra
W METHOD 206.:
Technique)"
MCAW
Furnace
oo
oo
II
Q
Ul
s
§
'8
1
.2
1
i
•8
cL
§
U
1
1
METHOD 6010 "1
icopy"
5? 2
li.
CO CO
00
00
E
o
II
Q
HYDAA
*
^^
•a
i
X
i
D
1
§°
««
o
METHOD 7061 ",
SW846
oo
oo
E
q
II
Q
06
U
U)
<
^
IB
i
00
*V9
"2
^
1
9
J
1
"o
•g
B
O
CO g
gi
>W METHOD INO
ledia, Multi-Concei
y
E .^
s ""
•5, o
u 3
03 t—
OO
o
II
a
Ul
.2
i
1
g
f
•a
•S
4)
2
$
1?
W METHOD 210.
ion, Direct Aspirati
0 J
2 <
oo
E
S
0
ll
a
S
at,
O
o
§
|
^
t
i
*g
•5
4)
JS
1.
If
st
W METHOD 210.:
ion, Furnace Techi
^1
S <
OC
o
o
II
a
a.
y
J
g
i
<
i
1
•8
"S.
U
1
i
METHOD 6010 "]
$
00
%
00
I
>T)
o
II
Q
00
q
o
II
O
a.
O
>>
a.
I
fas
II
173
-------
H
z S
2 3
S?z
,< 0
QUANTI1
DETECTI
H
Z
[U
S
INSTRU1
ATION
METHOD
0
S
H
UJ
u
1
J
I
**•
o
h
si
il
if
S .«j
0 S
3
i2
Chromium,
7440473
00
00
q
II
a
5
w
tt.
o
|
¥
•g
*
g
|
oo
If
MCAWW METHOD 218.
Absorption, Direct Aspirati
00
~oo
o
II
a
u.
o
o
1
E
1
U
1
4>
5?
oo *_^
1}
MCAWW METHOD 21 8.:
Absorption, Furnace Techn
1
r—
0
II
a
UI
a,
U
§
1
o
2
I
X
1
i
SW846 METHOD 6010 "]
Spectroscopy"
oo
1
8
II
Q
^
9l±*
U]
u,
eu
u
|
1
;§.
{
i
X
i
SW846 METHOD 7195 "(
Extracts"
1
*
•
X
o
Chromium,
7440473
oo
1
0
II
O
fu
O
3
i
£
*$
Q.
I
•f
g
•§
s
1
s
1
J
SW846 METHOD 7 196 "(
I
II
|
u]
B.
a.
v.
<£
J
|
uj
§
J
1
g
i
SW846 METHOD 7197 '(
Extracts*
I1
8
o
^ 00 00
s> «s «s
5 60 00
o E S
- 99
II ^ *
u
in
(X
(I)
U J
Q Q
u u
•8*?
1-
5,(>L
u U
174
-------
H
0 3
HI
H £j
^ f~*
•••) U]
CXQ
1
g
U)
Z
INSTRU
ATION
Q
0
w
2
Q
3
H
P
0)
U
i
U]
u.
Ul
06
Q
O
H
ia
2
Is
ANALYTE/
COMMON t
CAS NUMB
BO
1
O
II
u
o
cd
U
t.
•g
s
•§
s
^^
ci"
•n
o
•o
u
^r
1
u"
4)
i
i
tj
1
1
0*
I
i
u
Z
3
oo
^
CO
•s*
is
^ o
U H
Amenable to
Chlorination
eo
Q
O
II
BO
O
U4
Q
06
U
cf
(S
o
Q
I
O
II
»
BO
m
H
u
tt.
O
u
o
BO
BO
S
fn
m
H
s
u
a
175
-------
cxo
INSTRUME
ATION
HOD
tt.
O
s
uu
U
z
UJ
a.
a
u.
ui
ct
o
O
ME
C/5
Q
O
UJ
2g
1 =
3. Z
z o <
< O U
3°
1
Q
U
U
U
0
oo
3
§
Q
-
U
O
oo
oo ^
j* 00
I =
3 f^
O r *fl
5 g
I?
00
m
II
J
U
o
176
-------
§1
S2
H
U)
uu
— 00
S z 5
Sll
< 2 c/j
Z O <
< U U
oo
SP
o
0£
O
60
60
05
U
60
a
oo
60
tn
60
BO
3
Of
u
Ui
6
o
a
u
6
o
8
u
o
o
w
6
o
3
q
oo
2
Q
-
U
O
s
»0
U
BO
I
O
CO
6
o
CO
3
o
CO
O
uS
Q
Q m
.' o*
a. °
O
u
CO
3°
S
C>1
u
177
-------
CXQ
u
I*
^2g
m
< S CO
Z O <
< U U
5°
eo
I
ofi
u
3
a.
CO
U
O
CO
is
ffl
-------
Cfl
H
Z S
2 3
^i
1"
D U]
ao
i
1
s
*^ ^v
H O
15
Q
o
i
0
U
H
pi
^
U)
METHOD REFERENC
LU
s
•7 tU
.— , Z OQ
E§§
< S to
2 O <
< U U
oo
*
II
J
S
O
Q
Ou
6
o
II
•s 1
•3 2
II
if
si
•§ §>
JS 0
•^ o
CLP SOW METHOD QTM
Multi-Concentration Samples (
Chromatography Techniques"
U
£
o
j=
o
75
^" Irt
••«' p-
00
00
3
O
II
U
CO
6
o
1
1
.—
'3
X
i
(O
1
w
I
*o
i
4J
2
CO
CLP SOW METHOD ORG "
Media, Multi-Concentration"
U
g
JS
*«
g
J5
7 5>
^" Sri
•-" r-
00
3°
II
J
§
O
£
6
o
•S 3
U o
•- §
3 2
o •=
f 5
If
— .°
||
U *•
« ^>
CLP SOW METHOD QTM
Multi-Concentration Samples 1
Chromatography Techniques"
oo
1
q
II
S
CO
S
U
3
1
J5
f
1
I
CO
«
3
I
a
S
Q
^^
SW846 METHOD 8240 "Gas
Organics"
00
op
O
II
J
1
CO
2
U
O
1
1
1
•i
.>>
1
8
i
6
1
•a
1
(*••
^
I
o
a
CO
CLP SOW METHOD ORG "
Media, Multi-Concentration"
i
•5
Jj
o
|
N "*
^^ S
— c^
00
00
3
O
II
s
Q
y
w
U
o
>
M
U
i
ff
jj
>
.5
g
00
o
SW846 METHOD 8010 "Hal
00
3°
q
II
t
CO
i
o
1
-1
f
2
1
1
f
B
8
Q
^^
SW846 METHOD 8240 "Gas
Organics"
oo
3
O
II
00
3
00
OO
3
00
3
2 S?
CO
U
O
U
O
CO
U
O
179
-------
li
Hi
< H
- 53
u
2
£•* O
IS
Si
u
ta
06
a
u,
a
on
Q
O
X
•^ MM
^- flQ
S|i
J S Z
< 2 oo
z o <
< o u
BO
3
O
II
a
o
sc
a
a;
u
§?
BO
3
»^
d -•
II II
S 21
BO
BO
3
O
II
a
q
in
U
O
o
o
CO
U
a
CO
6
o
u
o
BO
3
o
II
sr
a
u
o
BO
3
3°
q
tX
BO
BO
9
BO
3°
o
u a!
o
u
o
CO
2
6
o
CO
i P 9
I § §
u
u S
Q
O
ss
as r-
o
180
-------
»^l
S
i
.
3
s
CA
CO
U!
CA
QG
g
i
u
1
bb
O
ll
•^ u
w ^
H H
= i ^
•j Si z
TAB
ON LIMITS FOR {
SOIL/SEDIME]
£
£
H
Z
DS AND DETECTION/QUA
O
X
H
iC
H
z2
23
H 2
P ^*
Z uJ
Is
^y Ci
I
i
Ul
2 z
2 5
E OF METHOD
H
0)
U
METHOD REFEREN
u
-« Q>
III
< 2 co
Z O <
< U U
BO
BO
g
O
u
1
s
SW846 METHOD 8020 "A
>
5 «
II
Cx
CL
CO
2
u
O
Volatile
k*
t
S
I
CO
I
•f
1
§
a
SW846 METHOD 8240 "G
Organics"
00
BO
3
o
II
CX
(A
o
CO
2
6
o
i
1
i
to
"3
^
1
.a
1
•a
1
Vb.
O
I
J
CO
*
CLP SOW METHOD ORG
Media, Multi-Concentration'
^^
Chloroethene
(Vinyl Chlorid
75014
BO
BO
3
*
II
l
2
Q
6
-• a
"32
S"O
q
3 I
2 |
^ M
1 Analytical Service:
lie Analysis by Quic
j&
|5
Mt •
CLP SOW METHOD QTM
Multi-Concentration Sample;
Chromatography Techniques
BO
^J
oo
g
00
^*
II
£
ft.
8
W
^
Volatile Organics"
SW846 METHOD 8010 "H
oo
04
3
O
»-«
it
ex
a.
CO
s
Volatile
Ui
•S,
f
8
CO
1
i
00
2
|
S
e
a
SW846 METHOD 8240 "G
Organics"
BO
BO
3
o
II
U
u
CO
2
6
o
i
1
2
i
en
'S
i
s
A
•B
1
5
1
CO
•
CLP SOW METHOD ORG
Media, Multi-Concentration'
f
B ®
Dichloromethai
(Methylene Ch!
75092
BO
^*
BO
3
O
\ft
II
g
a.
CO
2
g
Volatile
Q
f
e
CO
1
i
00
I
3
SW846 METHOD 8240 "Gi
Organics"
BO
ei
O
II
j
%
U
CO
2
u
O
1
"a
i
«
QO
_>»
|
.8
1
•a
1
<*«
o
§
1
a
CO
•
CLP SOW METHOD ORG
Media, Multi-Concentration"
22
Ethenyl Benzer
(Styrene)
100425
oo
^<
BO
3
O
ft
II
£
O.
CO
2
g
o
I
O
i
|
CO
i
I
•g.
1
u
2
SW846 METHOD 8240 "Gi
Organics"
BO
BO
3
o
II
J
ex
(A
O
CO
2
u
o
13
2
GO
j>S
|
CO
o
t of Work for Organ
§
8
CO
B
CLP SOW METHOD ORG
Media, Multi-Concentration"
U1
. I
s >*
Tetrachloroethe
(Tetrachloroeth
127184
181
-------
z
5 W
CXQ
S<
Ul
2
< 5 10
z o <
< u u
U
eo
3
cn
O
00
3
3 «
S 2
u u
o o
METHOD 801
SW
\0 .3
*f
co O
eo
eo
3
O
CO
O
O
eo
3°
S°
II
-J
00
o
Q
eu
o
CO
6
o
CO
oo
•a
eo
3
as
o
§
6
o
00
3
o
II
BO
3
§
«?
u
o
CO
I
182
-------
at
i
i
Z 2
0 j
Ii
PD
^ UJ
13 UI
OQ
H
Z
UJ
Ii
is
a
o
g
u
2
u.
O
u
£
£
UJ
0
z
UJ
05
uu
u.
UJ
ei
a
0
e
U)
2
UJ
>«-
ANALYTE/
COMMON NA\
CAS NUMBER
INORGANICS
eo
3
00
3
i
U
a.
y
•u.
O
oo
O
(N
oo »J oo sd
s "So = 2°
03 o 3
— o
-------
H
z 2
o 3
H
UJ
lz
H O
METHOD REFERENCE
UJ
5-
zg
< 2^
z o <
< u u
oo
3
q
vi
" "
0 0
i
.j eo
^b =
Q
U4
q
wi
Q
U
q
wi
II
u
g
3
O
-J
Q
BO
3
I'd
— 00
0 ~
I
3
8
u
2
3°
**
O
at
U
a
Ul
uj
U.
O
cS S3 5
O
a.
to
S
A
fl
184
-------
U
5
CA
O
o
X
f-
z i
O j
H °
i§
< H
3 U)
CXQ
1
H
Z
w
. §
ll
METHOD
i
_j
H
U]
U
Of
Ul
METHOD REF1
Ul
2 ^
ZLLJ
- 00
Ell
— i 2 z
< 5 CO
z o <
< U U
"So
q
_
**
IS
» 7
Is
u.
0
BO
-«
1
£
U
od 7191/SM
Technique)"
s i
So ***
II
(N £•
« 8
MCAWW METHOD
"Chromium (Atomic .
a
5
g
3 p-
iS
~ ^
u ?
_J
"So
q
II
j
Q
Ul
u.
i
§
1
|
o
|
§
P
s
m
od
(M
MCAWW METHOD
Extraction)"
"Sb
3
q
—
d
3
O
II
J
Q
U]
U.
1
*•!
•t
iromium, Hexi
S
B
r-
o\
| J
« S
"* K
11
W3 O
$ 5
oo "5
S M
t MCAWW METHOD
(Atomic Absorption, '
e
"a
a
S
o:
E"
^3
g
.C
U
'si
q
II
U
Q
O
.H
B
exavalent (Ato
Dissolved H
S
1
2 ,
in -g
oo' -C
S S
MCAWW METHOD
Absorption, Furnace '
Z
HI
tt.
.
TJ
1
i
1
§
8
^j
M
U
o
"•
jr
SMEWW METHOD
_J
"So
g
q
II
U
Q
^
^
u.
O
w
tt.
necipitation)"
avalent (Coj
i
E*
.3
I
u
B
oJ
SW846 METHOD 71
Z
.
"So
3
8
u
u
a
S
§
1
I
5
S
X
"Chromium,
1 3500CR D
f
2
i
2
CO
1
SW846 METHOD 71
(Colorimetric)"
"So
3
O
II
••4
§
•§,
M
S§
S
1
1
Q
^.X
1
1
§"
§
u
B
oo
SW846 METHOD 71
Polarography)"
ao
3
II
e
O Q
U U
3°
II
I
•S-?
87
18S
-------
X =
O
H
z S
2 3
< z
|i
ID u]
CXQ
i
m
••B
INSTRUJ
ATION
O
O
1
Ii.
o
£
fe-
ll)
U
z
Ul
oi
UJ
u.
U]
oe
a
0
ac
&
2
U]
z
-< K
•_ yj
Nl
jSz
< S co
Z O <
< U U
_J
"So
g
S
ii
3
l.o
II
f
8"
5
u
¥
1
i
E
<
1
•3
^
i
^
^*
0
o
o\
O
C
^
£
0
^
CO
•o
i
3
0 O
IM **
« •<
« :E
•a « .
c S
5 g •
U < C
3°
S
II
g
_J _J
OO 90
3 3
O O
ts m
II II
U U
a a
Ui W
oo
3
8
a
U]
a o
u ia
eo
3
I
BO
3
3°
lio
eo
3
s
3°J
|1
"* o
II •*?
00
3
ea _J
^ "M
o 3
— o
111
o
a
o
u
u
1
I I
•s!
1
I S
•s
S
§
X
I
CO
(N
g
PI
186
-------
= _= 5
X 5 U H
15* I
t? J &o *
s
-------
3
i
H
I
I
U
Q
O
H
zS
O j
t? z
< o
il
3 U)
o* o
1
r-
z
INSTRUM
ATION
O
O
IS
U)
2
u.
O
Ul
j~
fc;
U]
i REFERENC
Q
0
UJ
2?
Ul
s
< «!
z "1
~- CO
w z «5
£§D
J2Z
< S CO
z o <
< u u
z
co
6
o
u
1
S
1
CI9
|
.§
^
M
U)
."2
.er
i
cr
J
as -a
!
Is
11
!«•
> °°
U] %
2 w
co 2
S -
-1
o T
•« O5
£ u
< s^
Z
CO
6
o
u
I
s
o
.1
u
i
ttl
"O
g"
§•
j
•
a
o
i
H
>
^
U)
5
"
s
GO
1
O
o
1
f
o
^
I
•a
o<
|
I
^"8
W £
2^
co 2
eo
3
q
o
II
I
U
9
o
II
O1
ofi
u
_J
9
•*
mn
q
d
II
O
u
o
U
o
a
8
o
GO
6
o
O
U
*
u
o
(O
1
.
U
_J
3
q
6
CO
6
o
188
-------
c/j
83
O
I
U
Cs.
O
o
• 3 U]
CXQ
U
Q
1
25 """ cd
u £
j M
CO ~
*S
§
J
1
H
fc
ETECTION/QUA
<
C/3
O
o
H
OUS MATR
AQ
METHOD REFERENC
uu
ffl
J 2 Z
< 2 CO
z o <
< u u
SMEWW METHOD 6630C "Li
Method II"
c
^ t^
o§
9
II
e
U
ia
-OR
r Sa
pectr
hniqu
OD L
n Wate
Mass S
Tec
CLP SOW METH
Low Concentration
Chromatography-
Capture (GC-ECD)
3°
o>
06
U
8
Ul
u
o
eo
3
u
Q
U
w
6
o
o
_ d §P
9
d
II
Q
ia
8
6
o
BO
3
Q
CO
U
O
-J
3
0
CO
0
o
3
8
d
a
Q
U
S
O
on Gas Chro
Ext
Liqui
METHOD 6630B "Liqu
SMEW
Method
189
-------
f-
i
S Lti
CXQ
U)
2
CA
U
*~%
p«E
5? -J co
CO
S ^
£ §
c« 5
b S
s <
OF METHOD
METHOD REFERENCE/T
u
1*
51!
z o <
< u u
-------
I
I
o
u
i
u
b.
O
g
z
Q
28
i
u
§
I
O
U
U
f-
z i
2 3
J? 2
Ii
IE
ao
ia
^ ^
H* O
15
2!
1
Q
O
I
S
u
a:
§
I
O
tu
2
CO
_
< S
Z
<
CO
O <
U U
u
to
oo
ID
en
3
O
II
g
U
O
BO
3
O
II
ii
98
Z d
§
r 8
f^i O
u ii
Q
2
O
u
o
CO
6
o
a
u
Oi
oo
3
U-l
q
d
II
O
a
§
o
II
Q
CO
6
o
•g o
S -S
I
II
O
_j
ao
d
O
o
o
9
Q
j
U3
Q
O
191
-------
UJ
2
at
i
l
I
.
o
S
s
I
u.
s
Q
O
|
U4
z W
Z CO
Z, O <
< U O
-------
H
Of
£
I
U
Ct.
o
z
Q
I
1
S
Q
I
H
H
*"^, *™—
z 2
2 3
H Z
^2
P
7 U]
OQ
t
Z
U)
2
INSTRU
ATION
Q
0
1
2
S
I
OD REFERENCI
X
&
2
LU
"*T
<"
-i LU
^^ •"••
-. *• CO
w Z 5
>il
_1 2t Z
< 2 co
Z O <
< U U
_J
g1
r-
Tf
II
0
2
CO
U
0
•
•3
'I
1
t»
1
z
*
S
Q
O
1
2
O.
W
Q
O_
^
.' <»
B-S
O.S
_J
&io
3
S
d
II
Q
U
Q
O
Ul
u
o
*
1
^
.s
1
.a
8
a.
o «>
S?
UM S
O •§_
5 a
• I
•§ s
PS
«*
s *
§1
1!
s i
£ 2
°0
^3
U O
~J
S1
r~
• CO
II
co 2
j
~5o
3
2
o
d
II
Q
2
CO
1
f
1
i
a
o
1
1
!
•0
1
ETHOD6630B '
2
^^
^ '
2 '.
CO .
b
3
q
o
II
a
U
Ul
6
o
ETHOD66
H)
co
•s
3
q
CO
O
00
3
OO
3
II
U
Q
3
f.
II
U
a
3°
OO
3
3°
CO
U
O
u
o
U
o
co co
* 5
U U
o o
CO
0
o
•o
fM
VO
§
a.
U]
193
* I
•
» >> T
m o —
-------
UJ
_)
<
Z
<
H
z 5
23
< o
M h"
Z
CXQ
I
01
§
B! z
15
E OF METHOD
u
METHOD REFERENC
a«
"7 [l]
* CO
O
si
S ^
o <
u u
]
"oil
p
q
u
a
g
CO
U
O
"8 c
M 2
'55 TJ
JW
'
* *
i
o
II
a
Q
a.
i
g
•
C
8
1
•s.
o
'5
8100 " Poly nuclear Aromi
•s.
EPA METHOD 610/SW846
j
BO
™
C^
II
Q
CO
*g
1
U
O
*
-3
'1
1
CO
i
EPA METHOD 625 "Basett
j
i
0
II
CO
2
1
U
O
eo i>
P
If
•S g
« §
ition of Organic Compouni
I Capillary Column Gas Ch
11
EPA DW METHOD 525 "D
Water by Liquid-Solid Extrac
Mass Spectrometry"
_J
•0
fsj
II
_J
a
CO
2
I
U
a
i
O
1
i
-Liquid Extraction Gas Chi
I
SMEWW METHOD 6410B
Mass Spectrometric Method"
j
8
II
1
U
g
U)
U
*!
31
3 i
s §
cat Analytical Services for
anic Analysis by Quick Tu
?!
CLP SOW METHOD QTM
Multi-Concentration Samples
Chromatography Techniques"
j
>
§
o
II
Q
CO
JS
i
U
o
t
O
o
s
oo
3
-Liquid Extraction Chroma
J
•
SMEWW METHOD 6440B
j
3"
o
""
II
O*
SO
Jg
i
U
J
£•>
§
latography-Mass Spectrom
imn Technique"
is
?^
SW846 METHOD 8270 "Ga
Semivolatile Organics: Capill
_j
3°
§
0*
II
1
CJ
X
•
(A
1
o
1
>>
SW846 METHOD 8310 "Po
194
-------
H
Z S
2 3
M
OQ
u
Z
u
B
ETHOD
OF
METHOD REFERENCE/TI
w
z
SSI
21s
< O U
eo
3
Q
Q
00
O
Q
U
S
O
CO CO
3 3
o o
-------
(U
8
b
u
u
z
a
on
w
u.
Q
o
u
--. *• OQ
£ z I
Sil
< 2 co
Z O <
< U U
SP
"1
II
Q
S
*
o »
— —
II II II
U U U
Q Q Q
S S 2
CO
U
O
co
U
o
co
to
co
(I)
CO
^ 3
CO
6
O
§
U
ss
u
8
CO
• u
V t5
CQ O> *"*
i U
^S S3
w >•» ^P
s § *
z-S.58
196
J
o
CO
O
CO
Q
Z.
O
a.
C
_J
o
CO
U
O
-------
==§ g
u
J c«
g
Z 2
O j
li
•— i_
;=> u
CXQ
I
15
Q
O
I
S
U)
FERENCE
METHOD
u
-------
1
o
2ft
eo
3
ao
-------
CXQ
I
1
I
OF METHOD
METHOD REFERENCE
U)
i*
z u
•* ffl
< 2 CO
z o <
< u o
00
3
S
O
•2.
BO
3
O>
a.
BO
3
3
06
O
_)
"So
£
OL.
U
60
3
9
o
II
Q
=
o
-------
o
t-
H
z 2
2 3
< z
c F
t"j ^J
^* ("*
CXQ
INSTRUMENT-
ATION
/TITLE OF METHOD
METHOD REFERENCE
2 ^
111
z o <
< u u
z
o
3
U]
6
o
sS
4) S
00 **
s, 1
£»•«
v- 9
!l
Volatile Organic Compounds i
iromatography with Photoioni
ss"
DW METHOD 502.2 "\
Capillary Column Gas Ch
uctivity Detectors in Sent
III
1
^
o
_- 8
-- r~
J
eo
a
—
d
II
_J
Q
2
CO
§
O
.S
«]
|
g* .
o £*
u fi
1 1
Measurement of Purgeable Or|
is Chromatography-Mass Spe
DW METHOD 524.2 "N
r by Capillary Column Ga
II
d
00
3
(S
8
c!
II
1
Q
tu
CO
U
O
a
i
u
1
w
eo
2
0
ff
|
CO
i
cj
WW METHOD 6040B "(
rometric Analysis"
UJ u
ll
_]
I^H
3
0
in
II
_J
Q
2
CO
6
o
a
o
g
Purge and Trap Packed-Colur
•metric Method I"
s g
II
2S
§4
1|
»l
II
U O
14
j
"So
a
d
II
U
a
2
CO
2
u
o
1
a
o
g
Purge and Trap Packed-Colui
WW METHOD 6230C "
matographic Method II"
UJ g
Z
Q
a
O
§
Purge and Trap Capillary-Co!
WW METHOD 6230D "
matographic Method"
II
j
^^
oc
^
ir!
II
O.
CO
I
J2
*^J
JS
>
g
**•
I
Chromatography-Mass Spect
16 METHOD 8240 "Gas
00
C/J
s
3
O
u
_J
O
CO
CO
O
a
- *
3"
O
o
200
-------
i
z S
2 3
< o
^~* u\
^ &••
3 W
CXQ
1
U)
2
INSTRU
ATION
Q
O
S
Ul
ft.
0
U
g
U)
u
z
U)
as
u
u.
tu
DC
Q
O
X
H
UJ
2
Ul
2
< a:
,^ Z «
Uu *^^ v*
si. O ^
j 2z
< 2 CO
Z 0 <
< O U
"Bo
3
O
6
II
j
a
o
S
o
}
*
NO
1
13
2
i^
U)
CO
o
i
"S
2
?
9O
i
/I
i
a
o .
£ e
2 -i
2i
< j
a, •
u X
u
1
8
t O
**! n
"^ is
—^ «J
~ BO
00 3
O\ ^D
^D ^3
u ii
u u
CO
o
3°
00
q
o"
II
O v<3
II II
Q a
BO BO
3 3
O —
II II
a a
3°
o
HI
BO
3
m
q
o
II
Q
2
BO
3
O
3
U]
u
O
CO
z
U
O
CO
2
o
a
CO
CO
U
O
3°
a:
u
CO
201
-------
ao
H
INSTRU
ATION
O M
w u
SaC H
£ "! u •<
ess? s
ft GQ M> M
S
-------
g
I
H
Z 2
2 3
Si
fc U
Z £JQ
^ fjq
• 5
.s|
3|
0 *
It
^1
« §
•3 2 •»
EPA DW METHOD 502.2 "V
Trap Capillary Column Gas Chi
Conductivity Detectors in Serie.1
"So
3
q
-
"5)~M
3 3
3
•i
if
3 J3
2u
• s
Q 0
WW Method 6210
Capillary Column
Ui X
EPA DW METHOD 524.2/SM
Organic Compounds in Water b
Spectrometry"
203
-------
Cfl
1/3
g
z
O
3
CXQ
o
tu
- z
< 5 to
Z O <
< U U
eo
3
s
O
II
O
o
3
o
CO
U
O
oo
3
BO
o
eb
3
s
eo
3
**!
O
II
Q
5
BO
3
Q
eo
II
Q
5
o
CO
?
u
o
O
O
O
tu
U
O
Q
O
I
o.
I
*»
eo
3
q
o
II
Q
Q
HC
6
o
S H 0
204
-------
*
C/3
d S U H
CA
c
t-
z 2
2 3
^*i ^"
2 o
p
OQ
Z
w
2
D
N
co c
5^c
ITLE OF METHOD
ERENCE/T1
METHOD REF
w
2
— ^ CD
rjj rv vi
> 2 z
< 2 CO
z o <
< u u
_J
"ob
a
o
o
II
u
Q
2
a
^
M
U
O
•s-s
a >>
8,2
<2|
1 §
si
latile Organic Compounds
matography with Photoion
n
** a
go
° s
DW METHOD
Capillary Colun
luctivity
:tors in Series"
<; a. £ 5
t» 2 o 4^
W H O Q
«
i
1
o
u
ll-
"7 §
•» s
* ^^
d
00
3
O
II
U
Q
O
Cu
0
°
o
1
2.
2 •*"•
latile Aromatic and Unsatu
id Trap Gas Chromatograp]
> s
?&
^?.
o ?^
W1 -O
DW METHOD
pounds in Water
< 6
ft. O
uu U
J
BO
3
0
ts
II
u
a
2
CO
2
6
.5
i 1
a |f
|3V
111
iVW Method 6210B (Meth
isurement of Purgeable Or;
romatography-Mass Specti
a 3 g
I^i
Si!
til
Pi
•* «s eu
II*
cu ^ •,
s-£
3 S
"1
s?
*» .B
DW METHOD
nic Compounds :
^ u
cu 2»
to o
eo
3
Q
3°
Q
J
o
t
S1
a
a:
u
CO
U
a8
E HI
06
o o
U
O
CO
§
O
CO
S2
00 r-
205
-------
zS
O j
H z
gl
il
3 m
oo
H O
ii
1
uu
S^,
§!l
< 2 to
z o <
< o u
Q
o
3
q
o
Q
00
Q
.J
"oo
II
— *
o •»'
II II
BO
3
3
d
3°
II
V)
U
O
O
V
U
o
CO
y
o
CO
I
co
o
U
O
S
§
Ul
2
Ou
206
-------
I/)
ca
US
o
o
1
u
&j
i i
w u
U £••
u «c
W5
^ y
EC *~3
S o
i
H
H
CX
CXQ
E OF METHOD
w
u
06
U
u_
Q
O
1
2
O
o
H
to
j 2 z
< 2 CO
Z O <
< u u
q
ui
II
^
BO
3
a
a!
u
3
O
cx
ai
O
CO
s
o
CO
2
u
O
BO
3
0>
06
U
O
3
u
o
pect
^\
I
f
1
S
6
3
O
a
o
x
2?
CO O
_j
"So
oo
d
II
|
u
o
u
JB
£
8
2
•g
*
0/SME
<
Z
o
o
II
I
CO
U
o
o
5
oo
1
i
8.
ac g
Bl
28
a. *5
s? -
vo
Q
O
U]
u
^a
I o
|6
BO
3
9
d
II
a
2
0
O
BO
3
S
O
a
2
I
6
o
oo
3
_J
a
2
CO
U
O
-2-So
207
-------
CO
O
a
o
H
z i
23
.
Sf
i f
i 2
2 *
• *
EPA DW METHOD 524.2/SM
Organic Compounds in Water b
Spectrometry"
^
u
•o
8 '§
u —
JS -C
** CJ
1^'
s>s?
^ >^r^ r**
_J
"So
g
oo
0
II
_J
a
Z
CO
U
O
3
O
|
3
U
1
1
1
y
P
SMEWW METHOD 6230C "P
Chromatographic Method II"
Z
a8
O.W
U U
0 0
3
O
3
3
£>
!
I
i
£
P
SMEWW METHOD 6230D "I
Chromatographic Method"
_j
"So
3
OO
O
II
J
Q
Z
Q
3
•?
u
o
•
10
U
mated Volatile Organ
&
SW846 METHOD 8010 "Halo
j
1°
o
II
g
CO
U
0
J8
1
o
>
£
£
I
CO
1
>»
I
J
n
SW846 METHOD 8240 "Cast
Organics"
BO
3
O
co
•8
g •§
JS
g js
° £
•55
u
I
V)
U
O
. °°.
d d
II II
J J
Q a
35 2
U
o
CO
=5
o
o
Q
U
1
S
Q
O
a.
ui
so
3
s
o
II
a
JS 2
208
-------
C/9
£
c«
(/>
tt
co
o
CO
f-
z i
O j
HO
pg
*7 ^
iS
2,S
OQ
1
H
z
U)
2
3 _
£o
fe C
15
METHOD REFERENCE/TITLE OF METHOD
UJ
i*
-. z S
& z 1
>i3
J S Z
< S to
z o <
< U U
— I — J
"So "So
3 3
O oo
— r-i
II II
J U
Q Q
2 2
CO
35
u
o
e
1 0/SMEWW
nic Compounds i
netry"
^ a 5
EPA DW METHOD 524.1/SMEWW Method 6210B (Meth
Method 6210C (Method II) "Measurement of Purgeable Or]
Water by Packed Column Gas Chromatography-Mass Specti
IT
3
« 0
3 ^
•£ U
= v
5 J
5 ^fN
•— JM iTN
•S|S
™" ^* lx^
o *•£ IL?
j •«»^' p*»
_j
"So
3
S
o
II
U
o
z
CO
S
CJ
O
u
•I «,
§>i
If
It
EPA DW METHOD 524.2 /SMEWW Method 6210D "Mes
Organic Compounds in Water by Capillary Column Gas Chi
Spectrometry"
_j
"So
3
S
e>
II
_)
Q
S
CO
35
6
o
s
o
§
SMEWW METHOD 6230C "Purge and Trap Packed-Colui
Chromatographic Method II"
<
Q
U
U)
u
O
S
O
J3
SMEWW METHOD 6230D "Purge and Trap Capillary-Co
Chromatographic Method"
_j
"So
3
wi
II
i
a
E
CO
35
U
o
JD
>
£
S
SW846 METHOD 8240 "Gas Chromatography-Mass Spect
Organics"
_j
"So
3
q
II
cc
U
CO
35
U
o
<*.
o
.52
£
1
*J
§;L
._ *
CLP SOW METHOD LC-ORG "Chemical Analytical Servi
Low Concentration Water Samples for Organic Compounds
o
3
5
03 ^
4)
^ g
e S*
.s £
u5 52,
g
I
i
I
a
Chromatography-Mass Spectrometry (GC-MS) and Gas Chr
Capture (GC-ECD) Techniques'
w>
^1
_
J
1°
O
II
O*
«
U
CO
35
u
o
i
'S
1
i
-22
"3
j>»
09
O
CLP SOW METHOD ORG "Statement of Work for Organ]
Media, Multi-Concentration"
_j
"So
3
8
O
II
-J
a
35
O
•?
U
o
EPA METHOD 602 "Purgeable Aromatics*
j
"So
3
O
o
II
J
O
2
g
EC
U
O
•a-s
a i>
8.3
U
I1
M
.8.3
EPA DW METHOD 502.2 "Volatile Organic Compounds i
Trap Capillary Column Gas Chromatography with Photoion
Conductivity Detectors in Series"
i_]
etf>
a
00
s
O
II
J
Q
2
Q
eu
U
O
1
3 v
15
EPA DW METHOD 503. 1 "Volatile Aromatic and Unsatui
Compounds in Water by Purge and Trap Gas Chromatograp]
209
-------
U
5
z i
0 j
H 1
H ^J
aS
t
s
g z
f^ O
ss
Q
O
1
NCE/TITLE OF ME
S
w
METHOD REF
la
< *
ZUJ
-» fM
H§|
Z O <
< U U
•J
00
3
O
0
CO
2
6
o
—
"§
5 s
S
*S >s
a*
I 2
C BO
£ S
?|
y a
2 o
S §
1/SMEWW Method
ater by Packed Colui
s?
<" .S
EPA DW METHOD
Organic Compounds
Spectrometry"
0
I
I ^
fll
£ 5.2
"5b
3
S
o
o
CO
1
U
O
*
M
W ••
£?
<*> >>
i*
1 F
3 §
Su
* a
8«
i I
2 /SMEWW Method
ater by Capillary Col
«r *
Ks
EPA DW METHOD
Organic Compounds
Spectrometry"
_j
~--
o
II
S
CO
1
g
=
^B
5
>
£•
^J
M
I
CO
1
>>
"Gas Chromatograph
§
SW846 METHOD 82
Organics"
a
o
U
CO
§
O
a
CO
U
O
t
s
6
II
- 8
Q
o
II
Q
U
o
U
o
CO
6
o
U
O
S
6
II
a
U
o
-------
CA
I
*
en
g
u
to,
O
«<
Q
CA
22
§
o>
o
£
Q
C/3
Q
O
8
u
CO
8
0»
f-
z 2
2 J
U— — .
22
p£
;— U
5 w
< H
3 U
CXQ
1
Z
HI
3 2
r™* ^3
CO *""'
SS
METHOD REFERENCE/TITLE OF METHOD
Ul
2 .
< ai
Ztu
— CQ
W 2 «fi
>°i
j2z
< 2 CO
Z o <
< u o
_J
"So
3
3
6
II
j
Q
a
3
u
1
u
o
•S-3
3 £>
-x Q
tf A uw METHOD 502.2 "Volatile Organic Compounds in Water by Purge
Trap Capillary Column Gas Chromatography with Photoionization and Electr
Conductivity Detectors in Series"
"JT
u §
c ?,
£ •£
Jj
•5 "3 5
2 9 OO
•K «S f-
H b2
_J
00
3
O
d
II
_I
Q
2
Q
ey
U
0
EPA DW METHOD 503. 1 "Volatile Aromatic and Unsaturated Organic
Compounds in Water by Purge and Trap Gas Chromatography'
-j
"So
3
—
Tf
d* d
00 BO
3 3
f) — »
O •*
II II
U J
Q Q
22
CO
1
O
O
.5
EPA DW METHOD 524.1/SMEWW Method 6210B (Method D/SMEWW
Method 62 IOC (Method II) "Measurement of Purgeable Organic Compounds
Water by Packed Column Gas Chromatography-Mass Spectrometry"
j
00
3
•*
O
II
_J
0
2
CO
1
U
O
u
is
EPA DW METHOD 524.2/SMEWW Method 6210D "Measurement of Purg
Organic Compounds in Water by Capillary Column Gas Chromatography-Ma
Spectrometry"
a*
3
o
d
II
J
0
(i)
U
u
o
1
2
SMEWW METHOD 6040B "Closed-Loop Stripping, Gas-Chromatographic-
Spectrometric Analysis"
j
"So
3
s
O
II
_J
Q
CO
2
1
u
O
SMEWW METHOD 6230C Purge and Trap Packed-Column Gas
Chromatographic Method II"
<
Z
s8
0, U
1 1
u u
O O
SMEWW METHOD 6230D "Purge and Trap Capillary-Column Gas
Chromatographic Method"
J
oo
3
o
ni
II
j
2
CO
2
1
u
0
V
1
SW846 METHOD 8240G "Gas Chromatography-Mass Spectrometry for Vol
Organics"
_j
00
3
O
II
]
5,
SM'
e
CO
2
u
o
o c
•1 I
CLP SOW METHOD LC-ORG "Chemical Analytical Services for the Analy
Low Concentration Water Samples for Organic Compounds by Gas
Chromatography-Mass Spectrometry (GC-MS) and Gas Chromatography-Elec
Capture (GC-ECD) Techniques'
?
rs
f-g
o!
§£
2 §
8 •§ S
^ « (N
fi y.s
-J
oo
3
o •
II
II
/™v
s^
u
CO
^
u
o
•2
CLP SOW METHOD ORG "Statement of Work for Organics Analysis - Mul
Media, Multi-Concentration"
211
-------
Z 2
0 j
I
u:
i
o
u
§
u
I
o
II
J5 c
*%
U
5
Cfl
§
a
2
Z
<2 co
-c, vj
825
3
00
II
o
2
3
i
d
II
O
00
3
q
d
II
00
oo
00 00
3 3
o
-------
I
H
fc
Z S
O j
HI
?"*
Z G
3fe
CXQ
1
H
Z
S
2
Ofi §
Is
ITLE OF METHOD
METHOD REFERENCE/T
u
*s
< oc
M UU
H § |
< 2 CO
Z O <
< u u
J
"So
q
^™
II
a
oi
u
CO
s
U
O
1 1
"Chemical Analytical Services for the Analysi
s for Organic Compounds by Gas
Iry (GC-MS) and Gas Chromatography-Elect
CLP SOW METHOD LC-ORG
Low Concentration Water Sample
Chromatography-Mass Spectrome
Capture (GC-ECD) Techniques"
u
g
•£ — •
8 >2
o ^2 ^2
•c u J2
H BIS
00
3
0
II
.^
PC
u
CO
2
U
O
1
1
1
Cfl
£?
O
£
1
<*•
0
S
CLP SOW METHOD ORG "Sta
Media, Multi-Concentration"
"M
3
°
II
o>
u
a
y
U
O
emical Analytical Services for Multi-Media,
Organic Analysis by Quick Turnaround Gas
CLP SOW METHOD QTM "Ch
Multi-Concentration Samples for
Chromatography Techniques"
.j
3
8
0
II
Q
S
O
3
u
u
o
«
hod 8010/SMEWW Method 6230B ' Purgeal
EPA METHOD 601/SW846 Metl
Halocarbons"
"eb
3
vO
^*
II
O
CO
IS
6
o
EPA METHOD 624 "Purgeables
Z
Q
Ul
g
1
atile Halogenated Organic Compounds in Wi
igraphy"
EPA DW METHOD 502.1 "Vol
by Purge and Trap Gas Chromato
d
3°
S
O
II
a
2
Q
^
1
U
O
It
atile Organic Compounds in Water by Purge
matography with Photoionization and Electro
EPA DW METHOD 502.2 "Vol
Trap Capillary Column Gas Chro
Conductivity Detectors in Series"
3
!5
-
"So "So
3 3
eN «
o —
II II
u -J
Q Q
2 S
CO
s
U
O
.S
WW Method 6210B (Method I)/SMEWW
isurement of Purgeable Organic Compounds
romatography-Mass Spectrometry"
EPA DW METHOD 524. 1/SME
Method 6210C (Method II) "Mei
Water by Packed Column Gas Ch
J
3
en
O
d
II
a
CO
1
U
O
—
9 ca
WW Method 6210D "Measurement of Purge
Capillary Column Gas Chromatography-Mas
EPA DW METHOD 524.2/SME'
Organic Compounds in Water by
Spectrometry"
_J
3
8
O
II
Q
CO
1
0
0
rge and Trap Packed-Column Gas
SMEWW METHOD 6230C "Pu
Chromatographic Method II"
213
-------
11
D
CXQ
u
D
CA
U
U
2
.j
1
H
S
•<
0
O
DETE
CA
§
O
S
ETHOD REFERENCE
tu
S
2
.
oo
o
iri
II
Q
U
U
L)
6
'rap Caplary-Co!
i .a
|l
> F
a s
IS
g E
B ^
S|
o 8
2 o S
.a S S
^BS
214
-------
« «r 8
CO
1
CO
s
J
s
a
i
u
tu
U)
u.
O
u
y
E
u.
O
o
a:
u
u
UJ
oi
Q
o
u
I
2
O
Z
Q
O
1
o
oi
9
3
Q
o
X
u
•8
•5 5
2 £
i
-------
1
«
•
3
CO
u
a.
1
u
I
eo
1.
i
o
•g
•3
'
1
z
U4
U
U
a£
LU
u.
u
oc
O
O
UI
Q
O
I
§
I
§
-O
Q
O
I
2
s
VO
Q
O
I
8
a
o
I
S
VO
a
o
I
IS
o
p
II
2
216
-------
00
oo
uu
u
u
Ul
u.
u
I
I
o
Q
u
fl
H
a
o
I
tN
O
Q
O
O
Q
O
I
217
7
e
a
o
U)
-------
O
3
g
1
at
oe
oo
O
i
i
a
Ul
u
UU
at
uu
u.
uu
at
Q
O
8
Q
O
(-
Q
O
s
I/I
a
o
I
a
o
i
§
U)
(S
Q
O
X
U]
i
D
2
U
< »
Si
I
< (XI
w u
il
Ui f-
X <
H OS
^S
3
CO
218
-------
Q
O
U)
s
I
u
si
o
H
U
E
X -
•<
Cfl
U
H
§
Q
O
I
u.
O
uu
H
H
U
U]
tu
LU
Di
Q
O
I
UJ
Sis £
* 3 «.
Q
O
I
U
Q
O
I
CO
2
CO
O
8
Q
O
Z
m
Q
O
I
8
-------
8
u
3
"
I
3
u
z
Ui
at
ui
u.
Ui
06
Q
O
X
uS
-------
Q
O
U
H
U
a
o
u.
O
f-
H
Ui
U
Z
u
(2
Ui
U.
U]
OS
Q
O
X
U
1
u
3
8 «
s i
ll
I
i
•a
3
o
§
b
1
u
•a
"3
U
"8
V)
e
I
.5 8
II It II
U
.•8
I
u
•o
X
U
Q
O
U
Q
O
I
IT)
r*>
m
Q
O
1
§
I
S
a
o
I
»
X
W
U
a
o
X
Q
O
s
O
5
ui
X
I
a
z
OH
Ui
DU
O
o
I
X
UJ
Ul
2
CO
I
U]
a
z
Q
O
I
CO
\o
221
-------
Q
O
uu
tu
O
U
I
|
•s
Q
I
U]
u
u
Oi
U
u.
m
oc
Q
O
I
2
O
O
I
O
Z
03
O
Q
O
O
m
Q
O
Q
Q
O
I
a
a
o
*
Q
8
£
222
-------
Q
O
u.
O
O
u
a.
5!
u
ia
5
s
u
u
u
z
EFERE
OC
Q
O
ac
&
2
m
IS
Q
x
«
u
0
i
if*
m
^1 °
^1 3«
tu H
5 "
sol 2
09
*
Q
00
a.
fl
a
o
I
:S
X
-------
§
U)
2
1
y
Cu
a.
Ul
in
1
CO
U)
at
U
«
I
i
y
E
S
I
1
o
Q
O
I
&
U]
s
I
s
O
1
1
1-
2
S
s
O
I
I
I!
6O
O
CO
O
I
I
otf
o
u.
co
Q
O
X
IX
o
Ul
Ul
u.
UJ
Q
O
I
tu
CO
m
o
(M
Q
O
Q
O
1
2
a
o
O
O
I
pa
o
vO
O
O
I
a
o
I
Q
O
tu
O
o
I
to
00
a
o
I
U]
oo
224
-------
u
tu
a:
u
u.
U]
OS
a
o
I
3
o
8"
1
o
•S
o
a
§
8
Q
O
I
8
r-
Q
O
W
S
i I
Q
O
to
O
O
I
IU
Q
O
Q
O
X
u
r~-
a
o
a
o
I
2
a
o
1
2
a
o
tu
a
o
225
-------
U
Q
O
2
S
u
s
a,
= E
•• -^ tf
*z
! u
[=
t-
§
X
Q
O
b
uu
u
Ul
Q£
UJ
u.
cu
a:
Q
O
I
•aS
Ji
I*
-I
J *
>1
li
I
=1 s
•s£
e» .2
'(0 M
£2
isr
Js
* V
e
5
S
"3
x
u
I
I
u
£ o >>
H 2 u
U
4>
1
•o
o
o
§
5
§
"t
I
o
O
<
S
.o-
I
I
I
I
I
S
1
JB
S
"o
o
<
-
-
Q
O
uu
I
r-
§
Q
O
I
Q
O
UJ
p
I
226
-------
1
i2
3
"3
UJ
c
o
I
O
O
UJ
o
U)
Q£
UU
U.
U]
as
O
O
I
c/l
oo
Q
O
X
UJ
i
Q
O
Q
O
ui
oo
Q
O
o
««
-------
APPENDIX III
Table V- A
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
ORGANIC COMPOUNDS
Drinking Water (USEPA, Offica of Water)
EPA
Compound Class Matted No.
Acnotoin and Acrytonttrila
Baaa/Nautmls, Acids and
Pasted**
Banzidina*
Carbamat** and Uraa
DA^MM^A*
~VMUUM
Chlorinated Acids
^ftiLnjinsvtstrl llurknrArlirtfia
\simonnmS9Q riyuiuGMiwwi»v
CNorinated Paafcidas
1 2-Dfcremoalhana and
1 2-Dibromo-3-Ch)orepfODana
DHhiocareafltate Pasfcidas
Extractabla Oiganica
HaJoathara
Nttroaremalic* and laophorena
Nrtfogan and Phosphorous
NUiosanwias
N-Maftyfcaifaamates and
Nll^lmfc •iliMiirmtuhiiai
OrganohaldB Paateidas and
PCB*
f^iit^tM'tfii^M taMlhlif^ PaVftiM^dMB
wiyWIOpllvafnMUV rvJltvMMI*
Pwctilonntttion Screening of
PCB*
P«*tada and PCBs
Pastkadas and PCBs
Oiganochlorina
Phanote
Phttialat* Estora
PurgaaWa Aromattcs
603
625*
605
632
515.1
612
SOB
504
630
525*
611
609
507
607
531.1
617
614
622
506A
505*
60S*
604
606
602*
Analytical System
GC-FID
GC-MS
HPLC/Etoctrecham
HPLOUV
ECO
CapMaiy Column
GC-ECD
ECO
Capilaiy Column
GC-ECD
Cokximatric
GC-MS
Capilaiy Column
QC-ELCO
GC-FID + ECO
NPD
CapiUary Column
GC-NPD
HPLC
Fhioraaoanoa Datector
GC-ECD
QC-FPDorNPD
GC-FPD
ECD/ELCDPackador
Capillauy Column
GC-ECD
Capilaiy Column
GC-ECD
GC-FID
GC-ECD
GC-PtD
Sampla
Intioduetion/
Pnaoamtion
P&T
XTN
XTN
XTN
XTN
XTN
XTN
XTN
CSjUbamlion
XTN
XTN
XTN
XTN
XTN
Dl
XTN
XTN
XTN
XTN
XTN
XTN
XTN
XTN
P&T
Dataction Limit/
Ranoa foobt
0.5-0.6
0.09-44.0
0.08-0.13
0.003-11.1
EDL, 0.1-1,0
0.03-1.34
EDL 0.01-0.5 (most
<0.1)
0.01
1.9-15.3
0.1-1.0
0.3-3.9
0.01-15.7
EDL(EslinwtedD.L)
0.1 -5.0 (moat <1.0)
0.154.81
0.5-4.0
0.002-0.176
0.012-0.015
0.1-5.0
0.1-0.3
Vahabia
Pasticida 0.005-1.0
Haibtckla 02-7.0
PCB* 0.1-0.5
0.002-024
0.14-16.0
029-3.0
02-0.4
* Fraquanfly nsquastod method.
228
-------
APPENDIX III
Table V-A
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
ORGANIC COMPOUNDS (continued)
Industrial and Municipal Wast* Water (USEPA, Office of Research and Development)
Compound Class
Purgeabte Hatocarbons
Purgeabte Organics
Purgeable Organics
Purgeables
Volatile Aromatics and
Unsaturated Compounds
Volatile Halocarbons
Volatile Habcarbons
2,3,7,8-Tetrachtorodibenzo-p
dioxin
Triazine Pesticides
Sample
EPA Introduction/
Method No. Analytical System Preparation
601* GC-ELCD P&T
524.1 GC-MS P&T
Capillary Column
524.2* GC-MS P&T
Capillary Column
624* GC-MS P&T
503.1 GC-PID P&T
502.1 GC-ECD P&T
Packed Column
502.2* GC-ELCD/PID P&T
Capillary Column
613 GC-MS XTN
619 GC-NPD XTN
Detection Limit/
Range (ppb)
0.02-1.81
0.1-1.0
0.02-0.2
1.6-75
0.002-0.03
0.001-0.01
0.01-0.10
0.002
0.03-0.07
Aqueous and Solid Matrices (USEPA, Office of Water)
EPA
Compound Class MetfK)
Semivolatile Organics 1625
Tetra- through ccta- 1 61 3
chlorinated dioxins
andfurans
Volatile Organics 1624
Sample
Introduction/
dNo. Analytical System Preparation
Isotope Dilution by XTN
GC-MS (Capillary
Column)
Isotope Dilution by XTN
high resolution
GC-high resolution MS
Isotope Dilution by P&T
GC-MS (Capillary
Column)
Detection
Range (ppb)
most 20-1 00 ppb
(dependent on
% solids)
10-1 00 parts per
quadrillion in water
1-10 parts per trillion
in soil
5-1 00 ppb
(dependent
on % solids)
* Frequently requested method.
229
-------
APPENDIX III
Table V- A
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
ORGANIC COMPOUNDS (continued)
Solid Matrices (USEPA, Test Methods for Evaluating Solid
Wast*, SW846, November, 1986.)
EPA
Compound Class Method No. Analytical Svstem
Acrolein, Acrylonitrite,
Acetonitrite
Aromatic Volatile Organics
Chlorinated Herhiddes
wf Huiii munj nviMwHiKpo
Chlorinated Hydrocaibons
Nftroaromatics and Cydic
Ketones
Organophosphorus Pesticides
Organochlorine Pesticides and
PCBs
Phenols
Phthalate Esters
Polynudear Aromatic
Hydrocaibons
Polynudear Aromatic
Hydrocarbons
Purgeabte Hatogenated Volatile
Organics
Purgeable Non-Halogenated
Volatile Organics
Semivolatile Organics
Volatile Organics
8030
8020*
8150
8120
8090
8140
8080*
8040
8080
8100
8310
8010
8015
8270*
8240*
GC-FID
GC-FID
GC-ECDorELCD
GC-ECD
GC-FID or ECD
GC-FPDorNPD
GC-ECD
GC-FID
GC-ECD
GC-FID
HPLC/UV and Fluor
GC-ELCD
GC-FID
GC-MS
Capillary Column
GC-MS
Sample
Introduction/
Preparation
5030
5030
3550
3550
3550
3550
3550
3550
3550
3550
3550
5030
5030
3550
5030
Detection Limit/
Range (POD)
0.5-0.6
0.2-0.4
0.1-200
0.03-1.3
0.06-5.0
0.1-5.0
70-1000
0.14-16
0.29-31
Not Reported
0.013-2.3
0.03-0.52
rtot Reported
Not Reported
11.6-7.2
• Frequently requested method.
230
-------
APPENDIX III
TABLE V-B
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
INORGANIC ANALYTES
Analyte
Total/Dissolved Metals
Total/Dissolved Metals
Total/Dissolved Metals
Aluminum
Antimony
Antimony
Antimony
Barium
Barium
Beryllium
Beryllium
Boron
Calcium
Calcium
Cobalt
Cobalt
Copper
Copper
Cyanide
Cyanide
Cyanide
Cyanide,
Amenable to
Chlori nation,
without
distillation
Cyanide
Gold
Gold
Iron
Iron
EPA
Method No.
1620
6010
7000
7020
204.2 CLP
7040
7041
7080
7081
7090
7091
212.3
215.2
7140
7200
7201
7210
7211
335.2
335.2
355.1
4500-CN-H
Standard Method
for the Examin-
ation of Water
and Wastewater
1989
335.3
231.1
231.2
7380
7381
Analytical System
ICP
ICP
AA
AA
GFAA
AA
GFAA
AA
GFAA
AA
GFAA
Spectrophotometric
Titrimetric
AA
AA
GFAA
AA
GFAA
Total, (Titrimetric,
Spectrophotometric)
Midi (Distillation,
Total, Cotorimetric,
Automated UV)
Amenable to
Chtorinatton
(Titrimetric,
Spectrophotometric)
Spectrophotometric
Total, Spec-
trophoto-
metric
AA
GFAA
AA
GFAA
Sample
Preparation
3005,3010
3005.3010
3005.3010
3005.3010
*
3005,3010
3005.3010.3020
3005,3010
Nitric acid, reflux
3005,3010
3020
Hydrochloric acid
*
3005,3010
3005-3010
3020
3005,3010
Nitric acid, reflux
-*
**+
***•
pH>12
#•#
Nitric acid. Aqua
Regia
Nitric acid, Aqua
Regia
3005,3010
Nitric acid, reflux
Detection Limit
Range (pptrt
1.000
4300-5700
70
20
30
2.0
50-200
1.0-30
200
100,000
4800-5200
3400-4600
50
3700-4300
1.0
10
5.0
10
20
10
100
1.0
4400-5600
1.0
731
-------
APPENDIX III
TABLE V-B
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
INORGANIC ANALYTES
(continued)
Analvte
"*•"•****
Indium
Indium
Magnesium
Manganese
Manganese
Molybdenum
Molybdenum
Molybdenum
Molybdenum
Nickel
Osmium
Osmium
Osmium
Palladium
Palladium
Platinum
Platinum
Potassium
Rhenium
Rhenium
Rhodium
Rhodium
Ruthenium
Ruthenium
Selenium
Selenium
Selenium
Silver
Silver
Sodium
Thallium
Thallium
Tin
Tin
Titanium
Titanium
Vanadium
Vanadium
Zinc
Zinc
EPA
Method No.
235.1
235.2
7450
7460
7461
246.1
246.2
7480
7481
7520
252.1
252.2
7550
253.1
253.2
255.1
255.2
7610
264.1
264.2
265.1
265.2
267.1
267.2
270.3
7740
7741
7760
7761
7770
7840
7841
282.1
282.2
283.1
283.2
7910
7911
7950
7951
Analytical System
AA
GFAA
AA
AA
GFAA
AA
GFAA
AA
GFAA
AA
AA
GFAA
AA
AA
GFAA
AA
GFAA
AA
AA
GFAA
AA
GFAA
AA
GFAA
AA-Hydride
GFAA
AA Hydride
AA
GFAA
AA
AA
GFAA
AA
GFAA
AA
GFAA
AA
GFAA
AA
GFAA
Sample
Preparation
Nitric acid. Aqua
Regia
Nitric acid, Aqua
Regia
3005.3010
3005,3010
Nitric acid, reflux
*
*
3005.3010
3020
3005.3010
Nftrfc,sulfuric
acids
Nitric acid
3005,3010
Nitric acid
Nitric acid
**
•*
3005.3010
Nitric acid
Nitric acid
Nitric acid
Regia
Nitric acid
Hydrochloric acid
Hydrochloric acid
**
3020
3005,3010
3005,3010
Nitric acid, reflux
3005,3010
3005,3010
3020
**
**
**
**
3005.3010
3020
3005,3010
Nitric acid, reflux
Detection Limit/
Range (ppb)
3000
30
970-1030
10
0.2
100
1.0
10.000
-
4900-5100
300
20
.
100
5.0
1000
20
1000-2200
5000
200
50
5.0
200
20
30-50
W»*^W«W
5.0
1200-2800
0.2
4800-5200
1.0-10
800
5.0
400
•TWW
10
49400-50600
50
wW
5.0
o!os
232
-------
TABLE V-B
SUMMARY OF ROUTINE METHODS BY PROGRAM AND COMPOUND CLASS
INORGANIC ANALYTES
(continued)
Sample Preparation Methods
3005 Acid Digestion of Waters for Total Recoverable Dissolved Metals for Analysis by Flame Atomic Absorption
Spectroscopy or Inductively Coupled Plasma Spectroscopy.
3010 Acid Digestion of Aqueous Samples and Extracts for Total Metals for Analysis by Flame Atomic Absorption
Spectroscopy or Inductively Coupled Plasma Spectroscopy.
3020 Acid Digestion of Aqueous Samples and Extracts for Total Metals for Analysis by Furnace Atomic
Absorption Spectroscopy.
CLP preparation methods are categorized by water/soil, ICP, AA, and GFAA instrumentation.
• CLP methods are based on the 200 series Methods for Chemical Analysis of Water and Wastes. U.S.
Environmental Monitoring Systems Laboratory. Cincinnati, Ohio. March, 1983.
• Water sample preparation for GFAA uses nitric acid, hydrogen peroxide and mild heat. SOW 788, D-5.
• Water sample preparation for ICP and AA uses nitric acid, hydrochloric acid and mild heat. SOW 788, D-5.
• Soil sample preparation for ICP, AA, GFAA uses nitric acid, hydrogen peroxide and mild heat.
• Hydrochloric acid is used as the final reflux acid for several analytes. SOW 788, D-5,6.
** Nitric and hydrochloric acids are used for digestion.
*** Total cyanide is determined by a reflux-distillation procedure using a sodium hydroxide scrubber.
**** Cyanide amenable to chtorination is chlorinated at pH greater than 11.
233
-------
-------
APPENDIX IV
CALCULATION FORMULAS FOR STATISTICAL EVALUATION
Appendix IV provides calculation formulas to enable responsible risk assessment personnel to determine the
minimum number of samples necessary to meet statistical performance objectives. This appendix also provides
statistical guidelines on the probability that a given sampling plan will identify a hot spot, and the probability that no
hot spot exists given none was found after sampling.
Calculation Formulas to Determine the Number of
Samples Required Given Coefficient of Variation and
Statistical Performance Objectives
The minimum number of samples, n, required to achieve a specified precision and confidence level at a
defined minimum detectable relative difference may be estimated by the following equation:
For one-sided, one-sample t-test
For one-sided, two-sample t-test
2^Df+ 0.5Z».
where: Z is a percentite of the standard normal distribution such that P(Z i Z^) - a, Z. is simaarty defined,
and D - MDRD/CV, where MORD is the minimum detectable relative difference and CV is the coefficient of
variation. NOTE: Data must be transformed (ZJ, for example:
Confidence Level
1-0 a Z
0.80
0.85
0.90
0.95
0.99
0.20
0.15
0.10
0.05
0.01
0.842
1.039
1.282
1.645
2.326
0.80
0.85
0.90
0.95
0.99
Power
P
2.00
0.15
0.10
0.05
0.01
0.842
1.039
1.282
1.645
2.326
As an example of applying the equation above, assume CV « 30%, Confidence Level - 80%. Power - 95%.
and Minimum Detectable Relative Difference - 20%. For infinite degrees of freedom (t distribution becomes a
normal one), Z.= 0.842 and Z, -1.645. From the data assumed. D - 20% /30%. Therefore.
n 2 [(0.842 + 1.645)/(20/30)J* + 0.5 (0.842)*
n 2 13.917 + 0.354 - 14.269
n 215 samples required (round up)
Source: Adapted from EPA 1989C.
235
-------
APPENDIX IV
(continued)
Calculation Formulas For The
Statistical Evaluation Of The
Detection Of Hot Spots
Hot Spot Will Be Identified: Example * 1
These formulas are useful in evaluating the probability that a particular sampling plan will identify a hot spot.
Let R represent the radius of a hot spot and 0 be the distance between adjacent grid points where samples
will be collected. The probability that a grid point will fall on a hot spot is easily obtained from a geometrical
argument since at least one grid point must fall in any square of area D 2 centered at the center of the hot
spot. From this concept, it follows that the probability of sampling a hot spot P(H/E) is given by:
P(H/E) = (nR*yir ifflSD/2
= {R2!* - 2 arc cos (D/(2R))] + (D/4)V(4R2- D^J/D2 if D/2 < R < D/
= 1
where the angle D/(2R) is expressed in radian measure, H is the case that a hot spot is found, and E is the
case that a hot spot exists.
An example is if the grid spacing is D « 2R, then the probability of a hit is ic/4 « 0.785, which
implies that the probability that this grid spacing would not hit a hot spot if it exists is 0.215.
No Hot Spot Exists: Example*2
This set of formulas addresses the probability that no hot spot exists (given that none were found). This
argument requires the use of a subjective probability, P(E) (where P(E) is the probability that a hot spot
exists), based on historical and perhaps geophysical evidence. Then, if E is the case that there are no hot
spots at the study site and if H is the case that no hot spot is found in the sample, Bayes formula gives:
P(E I H) - P(H I E) P(E)/[P(H I E) P(E) + P(H I E) P(E)J
- P(H I E) P(E) / [P(H I E) P(E) +P(E)J.
For the case where D - 2R, it was found from Example 1 that P(HIE) =0.215. Therefore, if one is given that
the chance P(E) of a hot spot is thought to be 0.25 prior to the investigation, the probability of a hot spot
existing if the study does not find one is:
P(E I no hit) = 0.215 (0.25) / [0.215 (0.25) + 0.75] = 0.067.
Hence, the probability that no hot spot exists is (1-0.067) = 0.933.
Source: Adapted from EPA 1989C.
-------
Appendix IV (continued)
Number of Samples Required in a One-Sided One-Sample t-Test to Achieve a Mini-
mum Detectable Relative Difference at Confidence Level (1-a) and Power of (1-|3)
Coefficient
of Variation
(%)
10
15
20
Power
(%)
95
90
80
95
90
80
95
90
80
Confidence
Level
(%)
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
99
95
90
80
Minimum Detectable
Relative Difference (%)
5
66
45
36
26
55
36
28
19
43
27
19
12
145
99
78
57
120
79
60
41
94
58
42
26
256
175
138
100
211
139
107
73
164
101
73
46
10
19
13
10
7
16
10
8
5
13
8
6
4
39
26
21
15
32
21
16
11
26
16
11
7
66
45
36
26
55
36
28
19
43
27
19
12
20
7
5
3
2
6
4
3
2
6
3
2
2
12
8
6
4
11
7
5
3
9
5
4
2
19
13
10
7
16
10
8
5
13
8
6
4
30
5
3
2
2
5
3
2
1
4
3
2
1
7
5
3
2
6
4
3
2
6
3
2
2
10
9
5
4
9
6
4
3
8
5
3
2
40
4
3
2
1
4
2
2
1
4
2
2
1
5
3
3
2
5
3
2
1
5
3
2
1
7
5
3
2
6
4
3
2
6
3
2
2
Source: EPA 1989c
B2i-ooe-ao.i
237
-------
Appendix IV (continued)
Number of Samples Required in a One-Sided One-Sample t-Test to Achieve a Mini-
mum Detectable Relative Difference at Confidence Level (1-
-------
APPENDIX V
"J" DATA QUALIFIER SOURCE AND MEANING1
Appendix V lists the parameters and criteria that produce a "J" flag in accordance with the
National Functional Guidelines for Organic Data Review (EPA 1991e) and Laboratory Data Validation
Functional Guidelines for Inorganics Analyses (EPA 1988e) as applied to data from the Contract
Laboratory Program. The appendix also indicates the likely implication of this flag on the associated
result(s).
The criteria listed in this guidance should be used to flag CLP data as "J," or "estimated
concentration" (the associated numerical value is an estimate of the amount actually present in the
sample). With proper interpretation, the results of analytes which are flagged "J" can often be used in
making decisions.
Data flagged with "UJ" indicates that the value is undetected and quantitation limit may be
imprecise. Data flagged with "NJ" indicates that the value is tentatively identified and confirmation is
needed in future sampling efforts.
PARAMETER CRITERIA
ACTION
LIKELY
IMPLICATION1
ANALYSIS: Organic (3/90) VOA & BNA
Holding times 14 < VOA < 30 days Associated samples
7 < BNA < 22 days (+ results)
Low
Mass Calibration
Ion Abundance
Several data elements
in expanded window
All associated data
No generalization
Precision
Calibrations
~ initial
— continuing
Average RRF < .05
%RSD > 30%
RRF < .05
%D between initial
and continuing
calibration > 25%
Compound specific (+ results) Low
Compound specific (+ results)
Compound specific (+ results) Precision
Compound specific (+ results)
Blanks
If associated result is
between detection limit
and CRQL
Compound specific
High
239
-------
APPENDIX V (CONTINUED)
PARAMETER CRITERIA
ACTION
LIKELY
IMPLICATION1
Surrogates
If surrogate Fraction specific (+ results)
recoveries are low but (negative results are flagged
> 10% w/sample quantitation limit as
estimated (UJ»
Low
Any surrogate in a
fraction shows
< 10% recovery
If surrogate
recoveries are high
Fraction specific (+ results)
Fraction specific (+ results)
Internal standards If an IS area count is Associated compounds
outside -50% or (+ results) (non-detects flagged
+100% of the w/sample quantitation limit - UI)
associated standard
Low
High
No generalization
TICs
None
All TIC results - (NJ)
No generalization
ANALYSIS: Pesticides (2/88)
Holding Times 7 < PEST < 22
days
Associated positive results
(negative results - UJ)
Low
Instrument
Performance
DDT breakdown
> 20%
Endrin breakdown
>20%
Associated positive DDT Low
results (J)
Results for DDD and/or
DDE (NJ)
Associated positive Endrin results Low
(J); Results for Endrin Ketone (J)
240
-------
APPENDIX V (CONTINUED)
PARAMETER
Calibrations
— initial
— continuing
CRITERIA
ACTION
If criteria for linearity Associated positive results
not met
Surrogates
Compound
Quantitation and
Detection Limits
If %D between
calibration factors
> 15% (20% for
compounds being
confirmed)
If low surrogate
recoveries obtained
Quantitation limits
affected by large, off-
scale peaks
Associated positive results
LIKELY
IMPLICATION1
No generalization
No generalization
Associated results
Low
Estimated quantitation limit (UJ) No generalization
ANALYSIS: Inorganic (3/90)
Holding Times/
Preservation
Calibrations
- ICV or CCV
- ICS (for ICP)
Exceeded
Associated samples > IDL
[ IDL
< 0.995 [
120%
Associated samples
Associated samples > IDL
Associated samples > IDL
Low
No generalization
Precision
Low/High
High
241
-------
APPENDIX V (CONTINUED)
PARAMETER CRITERIA
ACTION
LIKELY
IMPLICATION1
If ICS recovery falls
between 50-79%
Associated samples > IDL
[ IDL
[ 2xCRDL
and 10% reported
concentration of the
affected element
Associated samples
High
LCS (Aqueous)
Recovery within
range 50-79% or
> 120%
Associated samples > IDL
[ IDL
Low/High
Recovery lower than Associated samples [ IDL
Matrix Spike Recovery > 125% or Associated samples > IDL Low/High
Sample < 75%
Recovery within
range 30-74%
Associated samples [< IDL (UJ)] Low
AA Post
Digestion Spike
Duplicate injection
outside + 20%
RSD (or CV) and
sample not rerun once
Associated data > IDL
Precision
242
-------
APPENDIX V (CONTINUED)
PARAMETER CRITERIA
Rerun sample does
not agree within
+ 20% RSD (CV)
ACTION
Associated data > IDL
LIKELY
IMPLICATION1
Precision
Post digestion spike
recovery < 40%
even after rerun
Associated data > IDL
Low
Post digestion spike Associated data [ < IDL (UJ)J
recovery > 115% or
< 85%
High/Low
If sample absorbance Associated samples > IDL
is < 50% of post [ < IDL (UJ)J
digestion spike
absorbance and if
furnace post digestion
spike recovery not
within 85- 115%
Low/High
MSA not done
Any samples run by
MSA not spiked at
appropriate levels
Associated data > IDL
Associated data > IDL
Precision
No generalization
MSA correlation
coefficient < 0.995
Associated data > IDL
No generalization
ICP Serial
Dilution
Criteria not met
Associated data > IDL
Precision
243
-------
APPENDIX V (CONTINUED)
Selected Acronym Key
BNA - Base/neutral/acid or semivolatile
CRDL - Contract required detection limit (inorganics)
CRQL - Contract required quantitation limit (organics)
CV - Coefficient of variation
ICS - Interference check sample
ICV - Initial calibration verification
IDL - Instrument detection limit
IS - Internal standard
PEST - Pesticide
RRF - Relative response factor
RSD - Relative standard deviation
TIC - Tentatively identified compound
VGA - Volatile
2 Implication Key
Low: The associated result may underestimate the true value.
High: The associated result may overestimate the true value.
Precision: The associated result may be of poor precision (high variability).
No generalization: No generalization can be made as to the likely implication.
244
-------
APPENDIX VI
"R" DATA QUALIFIER SOURCE AND MEANING1
Appendix VI lists the parameters and criteria that produce an "R" flag in accordance with the
National Functional Guidelines for Organic Data Review (EPA 1991e) and Laboratory Data Validation
Functional Guidelines for Inorganics Analyses (EPA 1988e) as applied to data from the Contract
Laboratory Program. The appendix also indicates the likely implication of this flag on the associated
result(s).
The criteria listed in this guidance should be used to flag CLP data as "R," or "unuseable." If
the flagged analytes are of interest, then resampling or reanalysis is necessary.
PARAMETER
CRITERIA
ACTION
LIKELY
IMPLICATIONS1
ANALYSIS: Organic (3/90) VOA & BNA
Holding times
Grossly exceeded
Professional judgment Low
(non-detects)
Mass Calibration
In error
Associated samples Unuseable
Ion Abundance
Outside expanded
windows
Associated samples
Unuseable
Calibrations
MeanRRTor
RRF < 0.05
Compound specific
(non-detects)
Low
Blanks
Gross contamination Compound specific High
(saturated peaks) (associated samples)
Surrogates
< 10% Recovery
Entire fraction
(negative results)
Low
Internal Standards
Extremely low area
counts; Major abrupt
drop off
Associated compounds Low
(non-detects)
TICs
Suspected artifacts
Professional judgment Unuseable
245
-------
APPENDIX VI (CONTINUED)
PARAMETER
CRITERIA
ACTION
LIKELY
IMPLICATION*
ANALYSIS: Pesticides (2/88)
Holding Times
Grossly exceeded
Professional judgment
(non-detects)
Low
Instrument
Performance
DDT
Retention
Time
Inadequate separation Affected compounds
Unuseable
RT
Peaks of concern
outside windows
Professional judgment
(positive results and
quantitation limits)
Unuseable
DDT/Endrin Not detected and
Degradation breakdown
concentrations
positive
Samples following last
in-control standard
(quantitation limit - DDT
and Endrin)
Low
Retention
Time Check
DEC > 2.0%
(packed)
> 0.3% (narrow-
bore)
> 1.5% (wide-bore)
Professional judgment Unuseable
Surrogates
Not present
Suggested (negative
results)
Low
Compound
Quantitation and
Detection Limits
Large off-scale peaks Quantitation limits
Unuseable
246
-------
APPENDIX VI (CONTINUED)
PARAMETER
CRITERIA
ANALYSIS: Inorganic (3/90)
Holding Times
Calibrations
- ICV or CCV
ICS (for ICP)
LCS (Aqueous)
Grossly exceeded
Minimum number of
standards not used;
Not calibrated daily
or each time
instrument set up
%R outside of 75-
125% (CN, 70-130;
Hg, 65- 135%)
Al, Ca, Fe or Mg in
samples .<. ICS and
ICS <50%
Results - 2xIDL for
elements which are
not present in the
EPA-provided
solution and levels of
Al, Ca, Feor Mg>
50% of levels found
in ICS, and estimated
interferences due to
Al, Ca, Fe or Mg
> 90%
Recovery < 50%
Matrix Spike Sample Recovery < 30%
AA Post Digestion Recovery < 10%
Spike
ACTION
Professional judgment
(Results < IDL)
Professional judgment
(associated samples)
Associated samples
Affected analytes
Affected analytes
Affected analytes
LIKELY
IMPLICATION1
Low
Precision
Low/High
High
High
Low
Affected samples (results Low
< IDL)
Affected samples (results Low
< IDL)
247
-------
APPENDIX VI (CONTINUED)
1 Selected Acronym Key
AA — Atomic absorption
BNA - Base/neutral/acid or semivolatile
CCV - Continuing calibration verification
DBC - Dibutyl chlorendate
ICP — Inductively coupled plasma
ICS — Interference check sample
ICV - Initial calibration verification
IDL - Instrument detection limit
LCS - Laboratory control sample
RRF — Relative response factor
RT - Retention time
TIC - Tentatively identified compound
VGA - Volatile
2 Implication Key
Low: The associated result may underestimate the true value.
High: The associated result may overestimate the true value.
Precision: The associated result may be of poor precision (high variability).
No generalization: No generalization can be made as to the likely implication.
Unuseable: Data are probably unuseable without resampling and reanalysis.
248
-------
APPENDIX VII
SUMMARY OF COMMON LABORATORY CONTAMINANTS, CONCENTRATION
REQUIREMENTS, AND RISK ASSESSMENT IMPLICATIONS
Appendix VII lists common organic laboratory contaminants that may appear in blanks.
The purpose of this appendix is to inform the reader of chemicals that may appear in analyses
but may not be present at the site. Analytes with values above instrument detection limits are
reported by laboratories. Some sample concentrations may not be reported through the review
process, as explained below, but if they are reported, possibilities of false positives exist. The
implications for risk assessment are included.
Common Laboratory
Contaminants
Concentration Requirements
Risk* Assessment
Implications
Target Compound
Methylene Chloride
Acetone
Toluene
Sample concentrations less than
lOx that detected in method
blanks will be reported as
undetected (or flagged B).
Sample concentrations less than
lOx that detected in method
blanks will be reported as
undetected (or flagged B).
Sample concentrations less than
lOx that detected in method
blanks will be reported as
undetected (or flagged B).
Include analyte if
concentration is greater
than lOx blank.
Include analyte if
concentration is less than
lOx greater than blank
concentration and multiple
chlorinated volatile analytes
are detected.
Exclude analyte in all other
situations.
Include analyte if
concentration is greater
than lOx blank.
Include analyte if
concentration is less than
lOx greater than blank
concentration and multiple
ketones are detected.
Exclude analyte in all other
situations.
Include analyte if
concentration is greater
than lOx blank.
Include analyte if
concentration is less than
lOx blank concentration
and multiple aromatic or
fuel hydrocarbons are
detected.
Exclude analyte in all other
situations.
249
-------
APPENDIX VII (CONTINUED)
Common Laboratory
Contaminants
Concentration Requirements
Risk Assessment
Implications
2-Butanone (methyl
ethylketone)
Sample concentrations less than
lOx that detected in method
blanks will be reported as
undetected (or flagged B).
Phthalates (i.e., dimethyl
phthalate, diethyl
phthalate, di-n-butyl
phthalate, butylbenzyl
phthalate, bis(2-
ethylhexyl) phthalate, di-
n-octyl phthalate)
Tentatively Identified
Compounds
Carbon dioxide
Diethyl ether
Sample concentrations less than
I Ox that detected in method
blanks will be reported as
undetected (or flagged B).
Not reported if present in the
method blank.
Not reported if present in the
method blank.
Hexanes
Not reported if present in the
method blank.
Include analyte if
concentration is greater
than lOx blank.
Include analyte if
concentration is less than
lOx blank concentration
and multiple ketones are
detected.
Include analyte if
concentration its greater
than lOx blank.
Exclude analyte in all other
situations.
o Exclude analyte in all
situations.
o Include analyte if
concentration is greater
than lOx blank.
o Exclude analyte in all other
situations.
o Exclude if analyte
concentration is not lOx
method blank.
o Exclude if analyte
concentration is not lOx
field blank (EPA
definition).
o Exclude if sample is not
analyzed within seven days.
250
-------
APPENDIX VII (CONTINUED)
Common Laboratory
Contaminants
Concentration Requirements
Risk Assessment
Implications
Freons (e.g., 1,1,2-
trichloro-1,2,2-
trifluoroethane, fluorotri-
chloromethane)
Not reported if present in the
method blank.
Solvent preservative
artifacts (e.g.,
cyclohexanone,
cyclohexenone,
cyclohexanol,
cyclohexenol,
chlorocyclohexene,
chlorocyclohexanol)
Aldol reaction products of
acetone (e.g., 4-hydroxy-
4-methyl-2-pentanone, 4-
methyl-penten-2-one,
5,5-dimethyl-2(5H)-
furanone)
Not reported if present in the
method blank.
Not reported if present in the
method blank.
Exclude if analyte
concentration is not lOx
method blank.
Exclude if analyte
concentration is not lOx
field blank (EPA
definition).
Exclude if sample is not
analyzed within seven days.
Exclude if artifact
concentration is not lOx
method blank.
Exclude if artifact
concentration is not lOx
field blank (EPA
definition).
Exclude if sample is not
analyzed within seven days.
Include analyte if
concentration is greater
than lOx blank.
Include analyte if
concentration is less than
lOx greater than blank
concentration and multiple
ketones are detected.
Exclude analyte in all other
situations.
251
-------
APPENDIX VHI
CLP METHODS SHORT SHEETS
TITLE: USEPA CONTRACT LABORATORY PROGRAM
STATEMENT OF WORK FOR ORGANIC ANALYSIS
MULTI-MEDIA, MULTI CONCENTRATION
DOCUMENT NUMBER:
DOCUMENT DATE:
EFFECTIVE DATES:
CONCENTRATION:
DATA TURNAROUND:
MATRICES:
OLM01.0
Not Applicable
September 28, 1990 through February 1994
Low to Medium
14 Days or 35 Days
Aqueous/Soil/Sediment*
SIGNIFICANT FEATURES
• The compounds include volatiles, semivolatiles, and pesticide/PCBs.
Volatiles and semivolatiles are analyzed by GC/MS; pesticides/PCBs are analyzed by GC/ECD.
Major Tentatively Identified Compounds (TICs) are reported for GC/MS analyses.
• Second column confirmation by GC/ECD is required for all pesticides/PCBs. Pesticides/PCBs which
are identified by GC/ECD at concentrations above 10 ng/uL are confirmed by GC/MS analysis.
REVISIONS/MODIFICATIONS
The following is a list of the significant changes from the 2/88 SOW that are incorporated in the
OLM01.0SOW:
• Selected volatile CRQLs have been raised; pesticide/PCB low soil CRQLs have been lowered; and
selected pesticide/PCB aqueous CRQLs have been changed.
• Target Compound List (TCL) changes include the elimination of vinyl acetate from the volatile TCL,
the elimination of benzyl alcohol and benzole acid from the semivolatile TCL, the addition of
carbazote to the semivolatile TCL, and the addition of endrin aldehyde to the pesticide TCL. The
semivolatile TCL compound bis(2-chtoroisopropyl)ether was renamed 2,2'oxybis(l-chloropropane).
• A new method for analysis of pesticides/PCBs is used. Changes include the use of wide bore capillary
columns, new surrogates, and new calibration techniques.
• Pesticide/PCB quantitation is performed using both the primary and secondary columns. The lower
value is reported by the laboratory.
The only significant change in the OLM01.1 (December, 1990) and OLM01.1.1 (February, 1991)
revisions to the OLM01.1 through OLM01.0 SOW was the lowering of selected semivolatile CRQLs. The
significant changes in the OLM01.1 through OLM01.7 revisions to the OLM01.0 SOW were the lowering
of selected semivolatile CRQLs and options for either a 14 day or 35 day data turnaround.
RECOMMENDED USES
This Routine Analytical Services (RAS) method is recommended for broad spectrum analysis to
define the nature and extent of potential site contamination during SSI, LSI, and RI/FS activities. This
method is suitable when a 14 day or 35 day turnaround for results is adequate. It is recommended for
samples from known or suspected hazardous waste sites where potential contamination may be present at
significant risk levels.
* Sediment samples with high moisture content should be solicited as RAS + S AS (Special
Analytical Service) in order to achieve the CRQLs.
COMPOUNDS AND CRQLs
The Target Compound List compounds included in the analysis and their Contract Required
Quantitation Limits (CRQLs) are listed in Attachment 1.
253
-------
TITLE: USEPA CONTRACT LABORATORY PROGRAM
STATEMENT OF WORK FOR ORGANIC ANALYSIS
MULTI-MEDIA, HIGH CONCENTRATION
DOCUMENT NUMBER:
DOCUMENT DATE:
EFFECTIVE DATES:
CONCENTRATION:
DATA TURNAROUND:
MATRICES:
Not Applicable
September 1988
June 7, 1989 through December 26, 1991
High: Greater than 20 ppm
35 Days
Liquid/Solid/Multi-phase
SIGNIFICANT FEATURES
• No holding times are designated for high concentration samples.
• The analyses are suitable for highly contaminated samples (>20mg/Kg).
• The analyses are acceptable for liquid, solid, or multi-phase samples. Multi-phase samples are
separated into water miscible liquid, water immiscible liquid, or solid phases. Each phase is analyzed
separately.
• Volatile, extractable (semivolatiles and pesticides), and multkomponent extractable (Aroclors and
Toxapbene) compounds are included.
Volatiles and extractables are analyzed by GC/MS; Aroclors and Toxapbene are analyzed by GC/ECD.
• Second column confirmation by GC/ECD is required for Aroclors and Toxapbene.
• Major Tentatively Identified Compounds (TICs) are reported for GC/MS analyses.
REVISIONS/MODIFICATIONS
The 1/89 and 4/89 revisions to the 9/88 SOW do not significantly affect data useability.
RECOMMENDED USES
This Routine Analytical Services (RAS) method is recommended for pie-remedial, remedial, or
removal projects where high concentrations of organic contaminants (greater than 20 rag/Kg) are suspected
and a 35 day turnaround for results is adequate. It is recommended for samples obtained from drummed
material, waste pits or lagoons, waste piles, tanker trucks, onsite tanks, and apparent contaminated soil
areas. The waste material may be industrial process waste, byproducts, raw tnatm^i^ intermediates and
contaminated products. Samples may be spent oil, spent solvents, paint wastes, metal treatment wastes,
and polymer formulations.
The method is suitable for solids, liquids, or multiphase samples, a phase being either water
miscible liquid, water immiscible liquid, or solid. Various methods of phase separation may be utilized
depending on the number and types of phases in a sample.
COMPOUNDS AND CRQLs
The Target Compound List compounds included in the analysis and their Contract Required
Quantitation Limits (CRQLs) are listed in Attachment 1.
254
-------
TITLE: USEPA CONTRACT LABORATORY PROGRAM
STATEMENT OF WORK FOR INORGANIC ANALYSIS
MULTI-MEDIA, MULTI CONCENTRATION
DOCUMENT NUMBER:
DOCUMENT DATE:
EFFECTIVE DATES:
CONCENTRATION:
DATA TURNAROUND:
MATRICES:
ILM01.0
Not Applicable
September 7, 1990 through September 26, 1993
Low to Medium
35 Days
Aqueous/Soil/Sediment*
SIGNIFICANT FEATURES
• The analyses are suitable for aqueous, soil, or sediment samples at low to medium concentration levels.
• This Statement of Work includes the midi distillation for cyanide analysis and the microwave digestion
for GFAA and ICP analyses. These two sample preparation procedures require less sample volume
than the traditional Statement of Work sample preparation procedures.
REVISIONS/MODIFICATIONS
None to date
RECOMMENDED USES
This Routine Analytical Service (RAS) method is recommended for broad spectrum analysis to
define the nature and extent of potential site contamination during SSI, LSI, and RI/FS activities. This
method is suitable when a 35 day turnaround for results is adequate. It is recommended for samples from
known or suspected hazardous waste sites where potential contamination may be present at significant risk
levels.
* Sediment samples with high moisture content should be solicited as RAS + SAS (Special
Analytical Service) in order to achieve the CRQLs.
ANALYTES AND CRQLs
The Target Analyte List analytes included in the analysis and their Contract Required
Quantitation Limits (CRQLs) are listed in Attachment 2.
255
-------
TITLE: USEPA CONTRACT LABORATORY PROGRAM
STATEMENT OF WORK FOR INORGANIC ANALYSIS
MULTI-MEDIA, HIGH CONCENTRATION
DOCUMENT NUMBER:
DOCUMENT DATE:
EFFECTIVE DATES:
CONCENTRATION:
DATA TURNAROUND:
MATRICES:
IHC01.2
Not Applicable
May IS, 1991 through November 30, 1993
High
35 Days
Liquid/Solid/Multi-pbase
SIGNIFICANT FEATURES
• The analyses are suitable for highly contaminated samples.
• The analyses are acceptable for liquid, solid, or multi-phase samples. Multi-phase samples arc
separated into water miscible liquid, water immiscible liquid, or solid phases. Each phase is analyzed
separately.
• The analyses include conductivity and pH; potassium is not included.
REVISIONS/MODIFICATIONS
The IHC01.1 and IHC01.2 revisions to the IHC01.0 SOW do not significantly affect data
useability.
RECOMMENDED USES
This routine Analytical Service (RAS) method is recommended for pre-remedial, remedial, or
removal projects where high concentrations of inorganic contaminants are suspected and a 35 day
turnaround for results is adequate. It is recommended for samples obtained from drummed material, waste
pits or lagoons, waste piles, tanker trucks, onsite tanks, and apparent contaminated soil areas. The waste
material may be industrial process waste, byproducts, raw materials, intermediates, and contaminated
products. Samples may be spent oil, spent solvents, paint wastes, metal treatment wastes, and polymer
formulations.
The method is suitable for solids, liquids, or multiphase samples, a phase being either water
miscible liquid, water immiscible liquid, or solid. A phase separation step is applied prior to digestion.
Each phase is analyzed and reported as a separate sample.
ANALYTES AND CRQLs
The Target Analyte List analytes included in the analysis and their Contract Required
Quantitation Limits (CRQLs) are listed in Attachment 2.
256
-------
USEPA Contact Laboratory Program
Statement of Work tor Organic Analyiu
Mult-Madia, Low to Medium and High Concentration
Attachment 1
Target Compound Ifet and Associated CRQU
PoMiee
Compound
Chtoromethane
Bromornetfiane
Vinyl Chloride
CNonetiene
Mettytene Chloride
Acetone
Carbon Otsuffioe
1>Dfchtoroatiene
1.1-OiehloroatMne
1.2-Oichtoroethene (total)
Chtorotorm
1.2-Dtchtaroetiane
2-Butanone
1.1.1-Trtahtoroattane
Carbon TeWjchtonde
Vinyl Ac*tatt
Sranndciiloranvtwoti
Ui-Trichtoroetvane
Benzene
ifotnotoim
2-Hexanone
Tetrachtoroetwne
Toluene
l . i 3. .2-1 atrBcnwroanene
Chtorobenzene
Ethyl Benzene
Slyrene
Xylenei (total)
LowtoHt
Aquiou*
COOL
10
10
10
10
10*
to
10*
to-
10"
1(T
10*
10*
10
10-
10*
—
10-
to-
10*
10*
10-
10*
10*
10*
10-
10
10
10*
10*
10-
10*
10-
10*
1(T
vHumflJ)
Low Sol
COOL
(ugfk^ppb)
10
10
10
10
10*
to
10*
10*
10*
10*
104
to-
10
10-
10*
—
10*
10*
10*
10*
10*
10-
10*
10*
10*
10
10
10"
10-
10*
10*
10*
10"
10-
mgbComBntntionM
"s^r^r
5.0
5.0
5.0
5.0
2.5
5.0
2,5
2.S
2.5
2.5
2.5
2.S
5.0
2.5
2.5
5.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
5.0
5.0
2.S
2.5
2.5
2.S
2.5
2.5
2.5
• CHQL* previouiry 5 uo/L and 5 U9*g in 2«a SOW
Note:
1 The iample-*pacifc CRQLt for toil lampkn will be adjusted lor percent moisture and will be higher than thow
listed above.
2 MedXim level soil CROL x 120 x Aqueous CRQL reported in ugfeg.
3 All CROLs are based on ««t weight and apply to solid and liquid sarnples.
4 Results tor both soW and liquid samples are reported aa moAg. wet weight
21-002470
257
-------
USEPA Contract Laboratory Program
Statement o< Work lor Organic Analysis
Mut>Madia, Low to Medwrn and High Concantration
Attachment 1 (Confd)
Target Compound (Jat and Asaoclatad CRQLs
SMrf-ltaMiM
Compound
i i •nanJJIi«laiM>»
MCMifininMiiBj
2T4-0*nftraph*nol
J kJBllMMtfaaMVil
Oftanahnn
2.4-omin)lakiana
rM^M«jb^^t*^ABrfa«
UMwijnyMiinMin
40*>rophanyHH«ny«har
Duorana
44«raanllna
< Drofnophanyt^tMnytaltMf
HMMCntofOOWlaMfM
PMuoiuupfMnu
Phan.v«tvm
Anttmcw
Carbuda
OMtbulylpMhiMa
Fknranlhana
Dwa^Mav
ryfwnp
am»«janiy|pMhalat»
U--OfcMorebsn*dlna
Banzo(a)antrncana
Ctayaana
hte/3_ntmlhanfnrtftfti^Ma
r*r«)clylpr«halat«
Banzo(t>)luoranUiana
9anzo0c)fluQranthana
Banzo(a)pyrana
lndano(U.3-anpyrana
OtMnaXa,h)anthracana
Banzo(0AI)p*iyl*n*
SM^-VUWto*^^
Lo»toll»dkm
Aqu.au*
COOL
(uoA,ppb)
10
2F
2P
10
10
to
10
to
2S*
2F
to
10
to
2P
to
10
10
10
10
to
to
10"
10
10
to
10
10
10
to
10
to
10
LowSoi
COOL
(00*0. ppt»
330
toy
too-
330
330
330
330
330
800-
800*
330
330
330
aoo*
330
330
330
330
330
330
330
330"
330
330
330
330
330
330
330
330
330
330
CATa^Ha^OTi^ \^9f
Uqu&Scm)
20
100
too
20
20
20
20
20
100
100
20
20
20
100
20
20
-
20
20
20
20
40
20
20
20
20
20
20
20
20
20
20
CRQU pravtoMly S uQ/L and 5 uo/Kg In 2/M SOW
~ CRQUpavfously20ug/Land600ugAgki2/88SOW
2 Madkiml*val«ollCRQL'120xAquaou§CROlraportadlnuB/l«.
3 AJCRCXaarabaaadonwatMigMamtapftytoiolklaniliquUsantp^
4 BwuHi lor both loHd and liquid tamptai ara raportad at mo/kg, w* waight.
21-002-07B.1
258
-------
USEPA Contract Laboratory Program
Statement of Work tor Organic Analyst
Attachment 1 (Confd)
Target Compound Uat and Associated CRQLa
Compound
Phenol
Mp-CMomelnyQctier
2-CMorophenol
1.4-OJcMorabeniMM
1^-Dtehtorooenzene
2-Metiylphenol
i2'-oxyt)l«(1-Chloropropei>e)
HencNoroMhm
NWrobenzene
ISDpnorone
mt(z-e
OknelhylphthaWe
Acenaphfialene
Z6-DW»oWuer»
3-NHroenWrw
LowtoMeefiMf
A?ueouf
CflCM.
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
to
10
10
25-
10
25*
10
10
10
25
Low Sol
COOL
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
330
800-
330
800*
330
330
330
800-
High ConowMnMM
(**)
UquWSotMMV-
(mg/Kg.ppm)
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
100
20
100
20
20
20
too
• C«X»pravk>u*y Sup/Land 5 ug/Vg In 2/86 SOW
1 Tlw umpto-ipacfffc CFKXs for K>« tampM w«l be •4u*«d tor pwtant m^
thoee Hitod atxwe.
2 Medkim level «o«CRCM.« 1000 xAqueou«CHCX reponex) In upA»
3 MCRQLtmbaartonwMwtf^tandappVtotolldindllquklMrnpto*.
4 Bewjflt tar boti eoNd and KquM simple* are reported a* mp/kg, wet weight
21 -O02'079(2Z1 -002-070,2
259
-------
Attachment 1 (Cont'd)
Target Compound List and Associated CRQLs
Swrtf-tfofeWM
Compound
alpha-BHC
beta-BHC
delta-BHC
nmmmm. CUJf* 11 m*imnm\
1 lans^niilnr
nvfMMnior
AWrin
Heptachtoreponde
EndosuKanl
DMdrin
4,4'-DDE
Endrin
Endoeutfanll
4,4'-DDO
Endosulfan surtate
4.4'-DDT
sndiNi ktttocw
alpha-Chtordane
|avrvrMi*Chtofctan0
SMnt-Votita**
LowtoMfdlum
Aqu»oiH
CRQL
(ug^ppb)
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.5
0.10
0.10
0.05*
0.05*
lotrSotf**
CRQL
(ugfcg.ppb)
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
3.3
3.3
3.3
3.3
3.3
3.3
3.3
17.0
3.3
3.3
1.7
1.7
ExtrmetMbt»f (1,2)
High Coneentntton
IJquid/SoikVMulti-Ph»»*
CRQL (mg/kg, ppm)
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
-
20
20
Note:
1 AICRQU am bM»d on wet weight «nd apply to soM and KqukJsainplM.
2 Results for both solid and liquid samples are reported as moAg, wet weight
Aqueous CRQLs changed from 2/88 SOW to the following:
• Aqueous CRQLs (ug/L) - alpha- and gamrm-Chlordane from OS to 0.05.
Al low soil CRQLs changed from 2/88 SOW to the fotowing:
" Low Soil CRQLs (ug/kg):
alpha-BHC through Endosulfan I from 8.0 to 1.7;
DwWrin through 4.4'-DOT and Endrin ketone from 16.0 to 33;
Methoxychlor from 80.0 to 17.0;
alpha- and gamma-CNordane from 80.0 to 1.7.
21-002-079J
260
-------
Attachment 1 (Cont'd)
Target Compound List and Associated CRQLs
Compound
Butyl alcohol
Benzole acid
Monochtorobfphanyl
Dfchtorobiphflnyl
Trtchtoroblphenyl
Tetrachtorobiphenyl
Hexachtorobiphenyl
Pentachkxobfphenyl
Octachtorobiphenyl
Nonachtorobiphenyl
Decachtorobiphenyl
Heptachtorobiphenyl
Toxaphene
Aroctor-1016
Aroctor-1221
Aroclor-1232
Aroctof-1242
Aroctor-1248
Arocter-1254
Aroctor-1260
Seml-Volmtllet
Low to Medium
Aqueous
CRQL
(ug/L, ppb)
-
-
-
-
-
-
-
~
-
~
-
-
5.0*
1.0*
2.0*
1.0*
1.0'
1.0*
1.0
1.0
Low Soil"
CRQL
(ugftg. ppb)
-
-
-
-
-
-
-
-
-
-
-
-
170.0
33.0
67.0
33.0
33.0
33.0
33.0
33.0
ExtnctMblee (1,2)
High Concentration
LiquioySolWMulti-Phase
CRQL (mg/kg. ppm)
20
100
100
100
100
100
100
100
200
200
200
100
50
10
10
10
10
10
10
10
Note:
1 All CRQLs are based on wet weight and apply to solid and liquid samples.
2 Results for both solid and liquid samples are reported as mg/kg, wet weight.
Aqueous CRQLs changed from 2/88 SOW to the following:
* Aqueous CRQLs (ug/L) -
Toxaphene from 1.0 to 5.0;
Aroclors-1016,1232.1242. and 1248 from 0.5 to 1.0;
Aroctor-1221 from 0.5 to 2.0.
All low soil CRQLs changed from 2/888 SOW to the following:
Low Soil CRQLs (ug/kg):
Toxaphene from 160.0 to 170.0;
Aroctor-1016,1232.1242, and 1248 from 80.0 to 33.0;
Aroclor-1221 from 80.0 to 67.0;
Aroclor-1254 and 1260 from 160.0 to 33.0.TCL Ex
21-002-070,4
261
-------
USEPA Contract Laboratory Program
Statement of Work for Organic Analysis
Multi-Media, Multi-Concentration and High Concentration
Attachment 2
Target Analyte List and Associated CRQLs
Anulytt
Aluminum
Antimony
AiMnie
Berium
DMyNuu
Cadmium
Calcium
Chromium
Cobalt
Coooer
""fr
Iran
LMd
Magnesium
Mengineee
Mercury
I***)
Potassium
Selenium
Silver
Sodium
ThOium
Venedium
ZkK
Cyenide
pH
Conductivity
UM Ut f*
Aoufout
CRQL
(ug^pfb)
200
60
10
200
5
5
5000
10
SO
25
100
3
5000
15
02
40
5000
5
10
5000
10
50
20
10
-
-
Low Soil
COOL
(uptog.pf*)
40
12
2
40
1
1
1000
2
10
5
20
0.6
1000
3
0.1
8
1000
1
2
1000
2
10
4
2
-
-
l«rj/| dmnj-lraflllll /•»•»!
UOjubySaMMuW-ftiMe
caOL(mgav.ppm)
80
20
5
80
5
10
80
10
20
40
20
10
80
10
0.3
20
-
5
10
80
20
20
10
1.5
N/A
3.0 (umhoft/cm)
Note:
1 The sample-ipecific CHQLs tor soil samples win be adjusted for percent mobture end wiH be higner ttiwi thom listed
above.
2 Medium level soil CRQL- 120 x Aqueous CRQL reported in ug/Kg.
3 Results tor both soM and liquid samples are reported as mg/kg, w«t weight
21-002-079.S
262
-------
APPENDIX IX
EXAMPLE DIAGRAM FOR A CONCEPTUAL MODEL FOR RISK ASSESSMENT
This appendix provides a schematic example of a conceptual site model. This example is
a copy of Figure 2-2 of Guidance for Conducting Remedial Investigations and Feasibility
Studies Under CERCLA (EPA 1989i).
263
-------
Glossary
Accuracy. The degree of agreement of a measured value with the true or expected value of the quantity of concern.
Analvte. The chemical for which a sample is analyzed.
Analvte Sneciation. The ability of an analyte to exist in, or change between, chemically different forms (e.g.,
valence state, complexation state) depending on ambient conditions.
Anthropogenic Background Levels. Concentrations of chemicals that are present in the environment due to human-
made, non-site sources (e.g., industry, automobiles).
Audit Sample. A sample of known composition provided by EPA for contractor analysis to evaluate contractor
performance.
Average. The sum of a set of observations divided by the number of observations. Other measures of central
tendency are median, mode, or geometric mean.
Background Sample. A sample taken from a location where chemicals present in the ambient medium are assumed
due to natural sources.
Bias. A systematic error inherent in a method or caused by some artifact or idiosyncrasy of the measurement
system.
Biased Sampling. A sampling plan in which the data obtained may be systematically different from the true mean.
Biased sampling protocols are appropriate for certain objectives (e.g., clustering of samples to search for hot spots).
Biota. The plants and animals of the study area.
A clean sample that has not been exposed to the analyzed sample stream in order to monitor contamination
during sampling, transport, storage, or analysis.
Broad Spectrum Analysis. An analytical procedure capable of providing identification and quantitation of a wide
variety of chemicals.
Calibration. The comparison of a measurement standard or instrument with another standard or instrument to report
or eliminate, by adjustment, any variation (deviation) in accuracy of the item being compared. Toe levels of
calibration standards should bracket the range of levels for which actual measurements are to be made.
Cancer Slnpe Factor. A plausible, upper-bound estimate of the probability of cancer response in an exposed
individual, per unit intake over a lifetime exposure period.
Chain-of-Custndv Records. Records that contain information about the sample from sample collection to final
analysis. Such documentation includes labeling to prevent mix-up, container seals to detect unauthorized tampering
with contents and to secure custody, and the necessary records to support potential litigation.
Chemical nf Potential Cnnrqn A chemical initially identified or suspected to be present at a site that may be
hazardous to human health.
Classical Model. A statistical description of experimental data that assumes normality and independence.
265
-------
Confidence. Statistically, a measure of the probability of taking action when action is required or that an observed
value is correct A confidence limit is a value above or below a measured parameter that is likely to be observed at a
specified level of confidence.
Contract laboratory Program (CLP). Analytical program developed for analysis of Superfund site samples to
provide analytical results of known quality, supported by a high level of quality assurance and documentation.
Contract Required Quantitation Limit (CROP The chemical-specific quantitation levels that the CLP requires to
be routinely and reliably quantitated in specified sample matrices.
Data Assessment. The determination of the quantity and quality of data and their useability for risk assessment.
Data Quality Indicator (DPI). A performance measure for sampling and analytical procedures,,
Data Quality Objectives fDQQsl Qualitative and quantitative statements that specify the quality of the data
required to support decisions. DQOs are determined based on the end use of the data to be collected.
Data Review. The evaluation process that determines the quality of reported analytical results. It involves
examination of raw data (e.g., instrument output) and quality control and method parameters by a professional with
knowledge of the tests performed.
Data IJseahilitv. The ability or appropriateness of data to meet their intended use.
Data Validation. CLP-specific evaluation process that examines adherence to performance-based acceptance criteria
as outlined in National Functional Guidelines for Organic (or Inorganic) Data Review (EPA 1991e, EPA 1988e).
Detection Limit. The minimum concentration or weight of an analyte that can be detected by a single measurement
above instrumental background noise.
Dilution. Adding solvent to a sample, with an analyte concentration higher than the standard calibration curve, to
bring the analyte concentration into a quantifiably measurable range.
Dissolved Metals. Metals present in solution rather than sorbed on suspended particles.
Domain. A mappable subset of the total area containing the populations, after which distinct statistical properties
can be described.
Dose-Response Evaluation The process of quantitatively evaluating toxicity information and characterizing the
relationship between the dose of a contaminant administered or received and the incidence of adverse health effects
in the exposed populations.
Duplicate. A second sample taken from the same source at the same time and analyzed under identical conditions to
assist in the evaluation of sample variance.
Exposure Area. The area of a site over which a receptor is likely to contact a chemical of potential concern.
Exposure Assessment. The determination or estimation (qualitative or quantitative) of the magnitude, frequency,
duration, and route of exposure.
Exposure Pathway. The course of a chemical or physical agent from a source to a receptor. Each exposure pathway
includes a release from a source, an exposure point, and an exposure route.
266
-------
Extraction. The process of releasing compounds from a sample matrix prior to analysis.
False Negative (type II or beta error'). A statement that a condition does not exist when it actually does.
False Positive (type I or alpha errort. A statement that a condition does exist when it actually does not.
Field Analyses. Analyses performed in the field using sophisticated portable instruments or instruments set up in a
mobile laboratory on site. Results are available in real time or in several hours and may be quantitative or
qualitative.
Field Portable. An instrument that is sufficiently rugged and not of excessive weight that can be carried and used by
an individual in the field.
Field Screening. Analyses performed in the field using portable instruments. The results are available in real time
but are often not compound-specific or quantitative.
Fixed Laboratory Analyses. Analyses performed in an off-site analytical laboratory.
Frequency of Occurrence. The ratio of occurrence of a chemical existing at a site compared to occurrence at all sites
or compared to the frequency at which the chemical was tested for.
fleographiral Information System CG1S). A computerized database designed to overlay multiple information
elements such as maps, annotations, drawings, digital photos, and estimated concentrations.
Oeostatistical Model. A statistical or mathematical description of experimental data with special attention to spatial
covariance or temporal variation.
fleostatisrics A theory of statistics mat recognizes observed concentrations as dependent on one another and
governed by physical processes. Geostatistical methods consider the location of data and the size of the site for
calculations.
Heterogeneous Distributing Sample property that is unevenly distributed in the population.
Historical Data. Data collected before the remedial investigation.
Holding Time. The length of time from the date of sampling to the date of analysis. CLP designates the holding
time as the date from laboratory receipt of sample until date of analysis.
Homogeneous Distribution. A sample property that is evenly distributed over the population.
Hot Spot. Location of a substantially higher concentration of a chemical of concern than in surrounding areas of a
site.
Hydrocarbon. An organic compound composed of carbon and hydrogen.
Identification. Confirmation of the presence of a specific compound or analyte in a sample.
Instrument Detection Limit ODD. The lowest amount of a substance that can be detected by an instrument without
correction for the effects of sample matrix, handling and preparation.
267
-------
Intake. A measure of exposure expressed as the mass of a substance in contact with the exchange boundary per unit
body weight and unit time.
Integrated Risk Information System (IRK). An EPA database containing verified RfDs, RfCs, slope factors, up-to-
date health risks and EPA regulatory information for numerous chemicals. IRIS is EPA's preferred source for
toxicity information for Superfund.
internal .Standard A compound added to organic samples and blanks at a known concentration prior to analysis. It
is used as the basis for quantitation of target compounds.
Judgmental/Purposive Sampling. The process of locating sampling points based on the investigator's best judgment
from historical data of where the sample should be taken.
Kriging. A procedure utilizing a spatial covariance function and known values at sampling locations to estimate
unknown values at unsampled locations. For each estimate, an error of estimate is generated.
Limit of Detection (LOP). The concentration of a chemical that has a 99% probability of producing an analytical
result above background "noise" using a specific method.
Limit nf fhiantitatinn fLOQV The concentration of a chemical that has a 99% probability of producing an analytical
result above the LOD. Results below LOQ are not quantitative.
Linearity. The agreement between an actual instrument reading and the reading predicted by a straight line drawn
between calibration points that bracket the reading.
Lowest-Observable.Adverse-Effect-Level ff.OAF.LV In dose experiments, the lowest exposure level at which there
are statistically or biologically significant increases in frequency or severity of adverse effects between toe exposed
population and its apparent control group.
Mass Spectrum- A characteristic pattern of ion fragments of different masses resulting from analysis that can be
compared with a mass spectral library for analyte identification.
Matrix/Medium. The predominant material comprising the sample to be analyzed (e.g., drinking water, sludge, air).
Measurement Error. The difference between the true sample value and the observed measured value.
Measurement Variability. The difference between an observed measurement and the unknown true value of the
property being measured.
Media Variability Variability attributed to matrix effects.
Method Blank Performance A measure mat defines the level of laboratory background and reagent contamination.
It is determined by analyzing a method blank consisting of all reagents, internal standards, and surrogate standards
that are carried through the entire analytical procedure.
Method Detection Limit (MpLV The detection limit that takes into account the reagents, sample matrix, and
preparation steps applied to a sample in specific analytical methods.
Minimum Detectable Relative Difference. Percent difference between two concentration levels that can be detected
in analyses.
268
-------
Modeling. A mathematical description of an experimental data set.
Natural Variation. Variation in values or properties of a parameter that are primarily determined by natural forces or
conditions (e.g., variation in background levels of a chemical of potential concern in soils at a site).
Normal Distribution. A probability density function that approximates the distribution of many random variables
and has the form generally called the "bell-shaped curve."
Null Hypothesis. For risk assessment, statistical hypothesis that states on-site chemical concentrations are not
higher than background.
Paniculate. Solid material suspended in a fluid medium (air or water).
Performance Evaluation Sample. A sample of known composition provided for laboratory analysis to monitor
laboratory and method performance.
Performance Objectives. Statements of the type and content of deliverables and results that are necessary to assess
the useability of data for risk assessment. For example, documentation (chain-of-custody records) must be available
to relate all sample results to geographic locations.
Population Variability The variation in true pollution levels from one population unit to the next Some factors that
cause this variation are distance, direction, and elevation.
Power. A parameter used in statistics that measures the probability that the result from a specified sampling or
analytical process correctly indicates that no further action is required.
Practical Qiiantitatinn Limit (PQIA The lowest level that can be reliably achieved within specified limits of
precision and accuracy during routine laboratory operating conditions.
Precision. A measure of the agreement among individual measurements of the same property, under prescribed
similar conditions.
Preliminary Remediation Goals rPRCkV Initial clean-up goals that 1) are protective of human health and the
environment and 2) comply with ARARs. They are developed early in the process based on readily available
information and are modified to reflect results of the baseline risk assessment. They also are used during analysis of
remedial alternatives in the remedial investigation/feasibility study (RI/FS)
Preservation. Treatment of a sample to maintain representative sample properties.
Qualifier. A code appended to an analytical result that indicates possible qualitative or quantitative uncertainty in
the result.
Qualitative. An analysis that identifies an analyte in a sample without numerical certainty.
Quality Assurance Project Plan (QAPiPV An orderly assembly of detailed and specific procedures which delineates
how data of known and accepted quality is produced for a specific project.
Onantitation Limit The lowest experimentally measurable signal obtained for the actual analyte using a particular
procedure.
269
-------
Quantitative. An analysis that gives a numerical level of certainty to the concentration of an analyte in a sample.
Random Sampling. The process of locating sample points randomly within a sampling area.
Range of Linearity. The concentration range over which the analytical curve remains linear. The limit within which
response is linearly related to concentration.
Reasonable Maximum Exposure CRMKV The maximum exposure that could reasonably be expected to occur for a
given exposure pathway at a site. The RME is intended to account for both variability in exposure parameters and
uncertainty in the chemical concentration.
Receptor. An individual organism or species, or a segment of the population of the organism or species, that is
exposed to a chemical.
Recovery. A determination of the accuracy of the analytical procedure made by comparing measured values for a
spiked sample against the known spike values.
Reference Concentration (RfO. An estimate, with uncertainty spanning an order of magnitude, of continuous
exposure to the human population (including sensitive subgroups) through inhalation that is likely to be without
appreciable risk of deleterious effect during a lifetime.
Reference Dnse (RfDV An estimate (with uncertainty spanning an order of magnitude or more) of a daily exposure
level for a human population, including sensitive subpopulations, that is likely to be without an appreciable risk of
adverse health effects over the period of exposure.
Relative Percent Difference fRPDV A measure of precision which is based on the mean of two values from related
analyses and is reported as an absolute value.
Relative Response Factor fRRF>. A measure of the relative mass spectral response of an analyte compared to its
internal standard. RRFs are determined by the analysis of standards and are used in the calculation of concentration
of analytes in samples.
l Tnvp.sripatinn ran A process for collecting data to characterize site and waste and for conducting
treatability testing as necessary to evaluate the performance and cost of the treatment technologies and support the
design of selected remedies.
Representativeness. The degree to which the data collected accurately reflect the actual concentration or
distribution.
Retention Time. The length of time that a compound is retained on an analytical column (common in GC, HPLC
and 1C).
Risk* Assistant A software developed for EPA which provides analytical tools and databases to assist exposure and
risk assessments of chemically contaminated sites.
Risk rharactf>ri*aHcm The process of integrating the results of the exposure and toxkity assessments (i.e.,
comparing estimates of intake with appropriate lexicological values to determine the likelihood of adverse effects in
potentially exposed populations).
Routine Method. A method issued by an organization with appropriate responsibility. A routine method has been
validated and published and contains information on minimum performance characteristics.
270
-------
Sample Integrity. The maintenance of the sample in the same condition as when sampled.
Sample Qiiantitation Limit (SOL). The detection limit that accounts for sample characteristics, sample preparation
and analytical adjustments, such as dilution.
Sampling and Analysis Plan (SAP). A document consisting of a quality assurance project plan, and the field
sampling plan, which provides guidance for all field sampling and analytical activities that will be performed.
Sampling Variability. The variability attributed to various sampling schemes, such as judgmental sampling and
systematic sampling.
Sensitivity. The capability of methodology or instrumentation to discriminate between measurement responses for
quantitative differences in a parameter of interest.
Simple Random Sampling. A sampling scheme where positions, times, or intervals are based on a randomized
selection.
Slope Factor. A plausible upper-bound estimate of the probability of a response per unit intake of a chemical over a
lifetime. The slope factor is used to estimate an upper-bound probability of an individual developing cancer as a
result of a lifetime exposure to a particular level of a potential carcinogen.
Solvent A liquid used to dissolve and separate analytes from the matrix of origin.
Spatial Variation The manner in which contaminants vary within a defined area. The magnitude of difference in
contaminant concentrations in samples separated by a known distance is a measure of spatial variability.
Spike. A known amount of a chemical added to a sample for the purpose of determining efficiency of recovery; a
type of quality control sample.
Split. A single sample divided for the same measurement by two processes for the purpose of monitoring precision,
accuracy or comparability of two analyses.
Standard Deviation The most common measure of the dispersion of observed values or results expressed as the
magnitude of the square root of the variance.
Standard Operating Procedures CSOPsV A written, document which details an operation, analysis, or action whose
mechanisms are thoroughly prescribed.
Stratified Random Sampling. A sampling scheme where the target population is divided into a certain number of
non-overlapping parts for the purpose of achieving a better estimate of the population parameter.
Stratified Systematic Sampling. A sampling scheme where a consistent pattern is apportioned to various subareas or
domains.
Stratify. To divide a physical volume or area into discrete units (strata) which are assumed to have different
characteristics; a numeric procedure to subdivide a set or sets of data.
Surrogate Standard A standard of known concentration added to environmental samples for quality control
purposes. A surrogate standard is not likely to be found in an environmental sample, but has similar analytical
properties to one or more analytes of interest.
271
-------
Surrogate Technique. The use of surrogate analytes to assess the effectiveness of an analytical process (i.e., the
ability to recover analytes from a complex environmental matrix).
Systematic Random (Grid) Sampling. A random sampling plan utilizing points predefined by a geometric pattern.
Target Compound/Analvte. The compound/analyte of interest in a specific method. The term also has been used in
the Federal Register to denote compounds/analytes of regulatory significance.
Temporal Variation. Variation observed in chemical concentrations that is dependent on time.
Tentatively Identified Compound (TIC). Organic compounds detected in a sample that are not target compounds,
internal standards or surrogates.
Toxicitv Assessment. The toxicity assessment considers the following: 1) the types of adverse health effects
associated with chemical exposures; 2) The relationship between magnitude of exposure and adverse effects; and 3)
related uncertainties such as the weight of evidence of a particular chemical's carcinogenicity in humans.
Toxicolopical Threshold. The concentration at which a compound exhibits toxic effects.
Turnaround Time. The time from laboratory receipt of samples to receipt of a data package by the client.
Uncertainty The variability in a process that may consist of contributions from sampling, analysis, review, and
random error.
95% Upper ronfidence Limit (\ TPT .v A value that, when calculated repeatedly for different, randomly drawn
subsets of site data, equals or exceeds the true mean 95% of the time.
nsefnl Range That portion of the calibration curve that will produce the most accurate and precise results.
Variance. A measure of dispersion. It is the sum of the squares of the differences between the individual values and
the arithmetic mean of the set, divided by one less than the number of values.
Viscosity. The physical property of a fluid that offers a continued resistance to flow.
Volatile Orpanics. The solid or liquid compounds that may undergo spontaneous phase change to a gaseous state at
standard temperature and pressure.
Wavelength. The linear distance between successive maxima or minima of a wave form.
Weipht-of-Evidence Classification An EPA classification system for characterizing the extent to which available
data indicate that an agent is a human carcinogen. Recently, EPA has developed weight-of-evidence systems for
other kinds of toxic effects, such as developmental effects.
272
-------
References
Aitchison, J. and Brown, J.A.C. 1957. The Lognormal Distribution with Special Reference to its Uses in
Economics. Cambridge University Press.
American Society for Testing and Materials (ASTM). 1979. Sampling and Analysis of Toxic Organics in the
Atmosphere. ASTM Symposium. American Society for Testing and Materials. Philadelphia, PA.
Baudo, R., Glesy, J., and Muntan, H., eds. 1990. Sediments: Chemistry and Toxicity ofln-Place Pollutants. Lewis
Publishers, Inc. Ann Arbor, MI.
Caulcutt, Roland. 1983. Statistics for Analytical Chemists. Chapman and Hall. New York.
Clesceri, et al., eds. 1989. Standard Methods for the Examination of Water and Wastewater. 17th Edition.
American Public Health Association. Washington, DC.
Dragun, J. 1988. The Soil Chemistry of Hazardous Materials. Hazardous Materials Control Research Institute.
Silver Spring, MD.
Eckel, William P., Fisk, Joan F., and Jacob, Thomas A. 1989. Use of a Retention Index System to Better Identify
Non-Target Compounds. Hazardous Materials Control Research Institute 1989, Proceedings of the 10th National
Conference, pp. 86-90.
Environmental Protection Agency (EPA). 1983. Methods for Chemical Analysis of Water and Wastes (EPA 200
and 300 Methods). Environmental Monitoring Systems Laboratory. Las Vegas, NV. EPA/600/4-83/020.
Environmental Protection Agency (EPA). 1984. Methods for Organic Chemical Analysis of Municipal and
Industrial Wastewater (EPA 600 Methods) as presented in 40 CFR Pan 136, Guidelines Establishing Test
Procedures for the Analysis of Pollutants under the Clean Water Act.
Environmental Protection Agency (EPA). 1985. Methodology for Characterization of Uncertainty in Exposure
Assessment, Office of Research and Development. EPA/600/8-85/009.
Environmental Protection Agency (EPA). 1986a. Guidelines for Carcinogenic Risk Assessment. 51 Federal
Register 33992 (September 24,1986).
Environmental Protection Agency (EPA). 1986b. Test Methods for Evaluating SolidWaste (SW846): Physical/
Chemical Methods. Third Edition. Office of Solid Waste.
Environmental Protection Agency (EPA). 1987a. Data Quality Objectives for Remedial Response Activities:
Development Process. EPA/540/G-87/003 (NTIS 9B88-131370).
Environmental Protection Agency (EPA). 1987b. Field Screening Methods Catalog. Office of Emergency and
Remedial Response.
Environmental Protection Agency (EPA). 1987c. A Compendium ofSuperfund Field Operations Methods. Office
of Emergency and Remedial Response. EPA /540/P-87/001. (OSWER Directive 9355.0-14).
Environmental Protection Agency (EPA). 1988a. Review of Ecological Risk Assessment Methods. Office of Policy
Analysis. EPA/230/10-88/041.
273
-------
Environmental Protection Agency (EPA). 1988b. Superfund Exposure Assessment Manual. Office of Emergency
Response. EPA/540/1-88/001. (OSWER Directive 9285.5-1).
Environmental Protection Agency (EPA). 1988c. Geostatistical Environmental Assessment Software (GEOEAS)
(database).
Environmental Protection Agency (EPA). 1988d. Methods for the Determination of Organic Compounds in
Drinking Water (EPA 500 Methods). Environmental Monitoring Systems Laboratory. Las Vegas, NV. EPA/600/4-
88/039.
Environmental Protection Agency (EPA). 1988e. Laboratory Data Validation: Functional Guidelines for
Evaluating Inorganics Analysis. Office of Emergency and Remedial Response.
Environmental Protection Agency (EPA). 1989a. Risk Assessment Guidance for Superfund, Volume I: Human
Health Evaluation Manual, Part A. Office of Solid Waste and Emergency Response. EPA/540/1-89/002. (OSWER
Directive 9285.7-01 A).
Environmental Protection Agency (EPA). 1989b. Risk Assessment Guidance for Superfund, Volume II:
Environmental Evaluation Manual. Office of Solid Waste and Emergency Response. EPA/540/1-89/001.
Environmental Protection Agency (EPA). 1989c. Ecological Assessment of Hazardous Waste Sites: A Field and
Laboratory Reference. Environmental Research Laboratory. EPA/600/3-89/013.
Environmental Protection Agency (EPA). 1989d. Integrated Risk Information System (IRIS) (data base). Office of
Research and Development.
Environmental Protection Agency (EPA). 1989e. Methods for Evaluating the Attainment of'Cleanup Standards,
Volume 1: Soils and Solid Media. Office of Policy, Planning and Evaluation. EPA/230/2-89/042.
Environmental Protection Agency (EPA). 1989f. Soil Sampling Quality Assurance User's Guide. Environmental
Monitoring Systems Laboratory. Las Vegas, NV. EPA/600/8-89/046.
Environmental Protection Agency (EPA). 1989g. Office of Water Regulations and Standards/Industrial
Technology Division (ITD) Methods (EPA 1600 Methods). Office of Water.
Environmental Protection Agency (EPA). 1989h. Data Use Categories for the Field Analytical Support Project. In
Draft. Office of Solid Waste and Emergency Response.
Environmental Protection Agency (EPA). 1989i. Guidance for Conducting Remedial Investigations and Feasibility
Studies under CERCLA, Interim Final. Office of Solid Waste and Emergency Response. EPA/540/G-89/004.
(OSWER Directive 9355.3-01).
Environmental Protection Agency (EPA). 1990a. Health Effects Assessment Summary Tables. Hirst and Second
Quarters FY 1990. Office of Research and Development (OERR 9200.6-303).
Environmental Protection Agency (EPA). 1990b. Geostatistics for Waste Management (GEOPACK) (database).
Environmental Protection Agency (EPA). 1990c. A Rationale for the Assessment of Errors in the Sampling of Soils.
Office of Research and Development. EPA/600/4-90/013.
274
-------
Environmental Protection Agency (EPA). 1990d. Contract Laboratory Program Statement of Work for Inorganic
Analysis: Multi-Media, Multi-Concentration. Document No. ILMO 1.0. Office of Emergency and Remedial
Response.
Environmental Protection Agency (EPA). 1990e. Contract Laboratory Program Statement of Work for Organic
Analysis: Multi-Media, Multi-Concentration. Document No. OLM01.0. Office of Emergency and Remedial
Response.
Environmental Protection Agency (EPA). 1991a. ECO Update. Office of Emergency and Remedial Response.
Publication No. 9345.0-051.
Environmental Protection Agency (EPA). 19915. Risk Assessment Guidance for Superfund, Volume I: Human
Health Evaluation Manual, Part B. Office of Solid Waste and Emergency Response. EPA/540/1 -89/002.
(OSWER Directive 9285.7-01 A).
Environmental Protection Agency (EPA). 199 Ic. Role of Baseline Risk Assessment in Superfund Remedy Selection
Decision. Office of Solid Waste and Emergency Response. (OSWER Directive 9355.0-30).
Environmental Protection Agency (EPA). 199 Id. Human Health Evaluation Manual Supplemental Guidance:
Standard Default Exposure Factors. Office of Solid Waste and Emergency Response. (OSWER Directive 9285 6-
03).
Environmental Protection Agency (EPA). 199 le. National Functional Guidelines for Organic Data Review.
Office of Emergency and Remedial Response.
Finkel, A.M. 1990. Confronting Uncertainty in Risk Management: A Guide for Decision-Makers. Center for Risk
Management Washington, DC.
Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. VanNostrand. New York, NY.
Keith, L.H. 1987. Principles of Environmental Sampling. American Chemical Society. Washington, DC.
Keith, L.H. 1990a. Environmental Sampling and Analysis. In Print American Chemical Society. Washington,
DC.
Keith, L.H. 1990b. Environmental Sampling: A Summary. Environmental Science and Technology. 24:610-615.
Koch, GeorgeS. and Link, Richard F. 1971. Statistical Analysis of Geological Data. Dover Publications. 0-486-
64040-X.
Krige.D.G. 1978. Lognormal de Wysian Geostatisticsfor Ore Evaluation. South Africa Institute of Mining and
Metallurgy Monograph Series.
Manahan, S.E. 1975. Environmental Chemistry. Willard Grant Press. Boston, MA.
Neptune, D£., Brandy, E.P., Messner, M., and Michael, D.I. 1990. Quantitative Decision Making in Superfund.
Hazardous Materials Control, pp. 18-27.
National Research Council (NRC). 1983. Risk Assessment in the Federal Government: Managing the Process.
National Academy Press. Washington, DC.
275
-------
Seicbel, H.S. 1956. The Estimation of Means and Associated Confidence Limits for Small Samples from Lognormal
Populations. A Symposium on Mathematical Statistics and Computer Applications in Ore Valuation. South Africa
Institute of Mining and Metallurgy, pp. 106-122.
Taylor, J.H. 1987. Quality Assurance of Chemical Measurements. Lewis Publishers, Inc. Ann Arbor, MI.
Thistle Publishing 1991. Risk*Assistant (software). Hampshire Research Institute. Alexandria, VA.
276
-------
Index
Accuracy See Data quality indicators (DQIs)
Analytical
base/neutral/acid (SNA) 39
iron 1, 52, 53
oil and hydrocarbons 45, 51, 52, 84, 106,119
polycyclic aromatic hydrocarbons 45,119
phthalates and non-pesticide chlorinated com-
pounds 52
volatile organics (VOAs) 55,57,78, 80,113
Analytical methods 13, 21, 22, 25, 26, 29, 30, 33,
41,45,47,57, 59,63, 64, 78, 83, 89,99, 100,
117,118.120
atomic absorption (AA) 47, 55, 58
gas chromatography-mass spectrometry (GC-
MS) 41,45,46,52,53
gel permeation chromatography (GPC) 39
inductively coupled plasma (ICP) 52,53,55,58,
101
X-ray fluorescence (XRF) 57
Analytical services 3,21,28, 29, 83
field analyses 2,21,28,29,57,58,84,88,89
fixed laboratory analyses 21,29,54, 57,58,84,
89,100
quick turnaround method 28
special analytical services (SAS) 29
Automated data review 35
B
Background sampling 29,50, 75,119
anthropogenic 2,75,119,120
sampling 29,50
Baseline human health risk assessment 1, 3,4,7
Biota sampling 39,83
Chain-of-custody 29,101
Chemical intake 14,15
Chemicals of potential concern 1,4,25,26,29,30,
35,40,41,46,47.50, 52,53, 55,63-65,72-74,
77,78, 80,83,84,87, 88, 117-120
Comparability See Data quality indicators (DQIs)
Completeness See Data quality indicators (DQIs)
Concentration of concern 10,33, 34,47,48,83
Conceptual model 11,18,22,28
Contract Laboratory Program (CLP) 2,29,41,49,
58, 83. 84,87, 100,103, 105, 106,113
Contract required detection limit (CRDL) 49
Contract required quanb'tation limit (CRQL) 49
Corrective action 4,22,36,88,95,97,100,101,
106
D
Data
assessment 11, 21, 22, 95, 100-103, 105, 107,
109, 111, 113, 114,116
collection 1-4,7, 11, 18, 20, 25, 29-31, 33, 34,
36, 37, 50, 51, 63, 81, 101, 106-112, 116
qualifiers 4, 100, 113
review 2, 3,4, 20, 22, 23, 25, 29, 34, 35, 89,99-
103,105, 107,117-119
sources 1, 2,3, 26,28,29,99,101,111
Data quality indicators (DQIs) 3, 29, 31, 76,103,
121
accuracy 25,29, 31,33, 34. 39,49, 51, 55, 58,
99,101,102, 105-107, 112, 113,116-118
comparability 33,57,76,78,80,99,105,107,
108,112,114,116,121
completeness 76-78,99,100,102,105-107, 114,
116-118, 120,121
precision 29, 34,49,99-102,105-107,109,111-
113,116-118
representativeness 76,99, 105,107-109, 114,
116,117,121
Data quality objectives (DQOs) 2,11,13, 31, 34,
63,100,110, 111
Data useability criteria 3,25,26,99,117,121
Design decisions 81,89
Detection limits 2,25.28,30,33,37.45-48,54,55,
77,83.84,87,89,117-120
contract required detection limit (CRDL) 113
contract required quantitation limit (CRQL) 113
instrument detection limit (IDL) 47,48
limit of quantitation (LOQ) 49,50
method detection limit (MDL) 2,47,48,49,50,
102,113
practical quantitation limit (PQL) 49
sample quantitation limit (SQL) 2,22,23,48,49,
50,84
Exposure 95,97,101,105,107,108,112
area 4,11,13,18,20,25,26, 33, 54, 55,63,65,
72, 74, 77, 78, 80,89,120,121
assessment 4,7,13,14,15,17,18,101,102,108
pathway 11, 13-15.17,18. 33. 58,63,80, 89,
117,120,121
False negatives 11,13,18,25,35,40,41.47,48,
50, 58,64, 75, 76,101,105,108.113.116-118
277
-------
False positives 11,13, 25, 30, 35,41,45,47, 48,
50,53,64,76,101,105, 113, 117-119
Field analyses See Analytical Services
Field records 29
Fixed laboratory analyses See Analytical Services
G
Geographical Information System (CIS) 18,72
H
Hazard Ranking System (HRS) 13,26
Health Effects Assessment Summary Tables
(HEAST) 15
Historical data 11,18,26,28,41,45,52,73, 74,
78,119
Hot spots 13,33,51,54,57,66,73-76,78,89
I
Integrated Risk Information System (IRIS) 15
L
Laboratory performance 25,33,58,59,88,107,
HI
Land use alternatives 78
Linearity
limit of linearity (LOL) 50
range of linearity 47
M
Measurement error 33,37,38,50,76,109,111
Media variability 51,74
N
National Priorities List (NPL) 50
Natural variation 38
Performance evaluation 33,39,58,87.88,116
Performance measures 63,76,80,88,110
Performance objectives 4,25,29, 33,50,97, 105,
111
Precision See Data quality indicators (DQIs)
Preliminary remediation goals (PRGs) 2,48
Q
Qualified data 2,23,105,106, 113
Qualitative/quantitative analysis 57
Quality assurance (QA) 2,18,20,22,29,39,58,
76,100,101
Quality assurance project plan (QAPjP) 2,20,29,
33
Quality control (QC) 2,29, 33, 34, 37, 50,58, 59,
88, 100-103, 105,107, 108, 111, 113, 116, 118,
119
Reasonable maximum exposure (RME) 13, 14, 17,
55,66,105,107,109, 116
Reference concentrations (RfCs) 15,17
Reference doses (RfDs) 15,17
Remedial investigation (RI) 1-4, 11, 18, 20, 21, 25,
26,28, 29,63, 65, 81, 95, 100, 105
Remedial project manager (RPM) 1,4,11, 18,20-
23,25, 29,30, 34-37, 39,41,45^7, 51,53, 58,
59.63-65,72, 77,78,80,81,84.85,87-89,95,
113
Representativeness See Data quality indicators
(DQIs)
Resource issues 88
Risk Assessment Guidance for Superfuod
(RAGS) 1-3, 7.13-15, 17,18, 102, 114,119
Risk assessor 1-4,7,14,15,18,20-23,25,50,52-
55,58, 63-65. 77, 78.80,81,84,87-89,95,97,
100-102,106-108,110, 111, 113.114, 116
s
Sample
preparation 47,49,51,54,55,77,88,108,112
preservation 76,108,116
Sampling and analysis plan (SAP) 1,20-22,25,63,
74,88, 97,100,107,110,112, 113
Sampling design methods
classical model 65,72,78,88
geostatistical model 65,66,73-75,78,88
judgmental/purposive model 65, 73, 74
systematic grid sampling 65,66,72,75,78,88
Sampling Design Selection Worksheet 63,65,72,
80,83,89
Sampling variability 64,65,74, 77,108,109,110,
116
Scheduling 21
Scoping 11,25,28,29,41,88, 105
Site
concentrations 11,13,25,63-66,72-77,80,95,
101,107-109,116,119,120
inspections 3,18,26,73
Soil 4,37,38,41,50,51,55,119,120
data collection 1,4,63,77,78,80,81,117,119,
121
location of hot spots 66,73-75,78
sampling depth 78,80
characteristics 11,80
Soil Depth Sampling Worksheet 37,63
278
-------
Standard operating procedures (SOPs) 29, 31, 100,
101
T
Target compound list 4
Tentatively identified compounds (TICs) 41, 45, 52
Toxicity assessment 4, 7,15, 17,22
Turnaround time 2, 29, 54, 58. 83, 84, 87, 89
u
Uncertainty 1-4, 7, 10, 11, 14, 15, 17, 18,25, 33,
37, 38, 50, 51, 55, 63, 76, 80, 81, 89, 95, 97,
102,105,107,111,114,117,121
analytical 7,10,14,15,17, 18,80
sampling 63,76,77,80,81,89,118
total 76,77
279
------- |