United totes
Environmental Protection
Environmental Monitoring
Systems Laboratory
P.O. lex 18027
Las Veg«» NV Ml 14-8027
>EFV\
Sediment Sampling
Quality Assurance
User's Guide
u s. Environmental Protection AI«K>
Region 5, library (PL-12J) 13thna-
77 West Jacteon Boulevsra, lea> "•*
Chicago. IL 60604-3590
-------
PROJECT SUMMARY
SEDIMENT SAMPLING QUALITY ASSURANCE USER'S GUIDE
by
Delbert S. Barth and Thomas H. Starks
Environmental Research Center
University of Nevada, Las Vegas
Las Vegas, Nevada 69154
Cooperative Agreement
CR 810550-01
Kenneth W. Brown, Project Officer
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
Office of Research and Development
Las Vegas, Nevada
May 1985 u S. Environmental Protection Agtnc*
Region 5, Library (PL-12J)
77 West Jackson Boulevard, IZtn
Chicago. IL 60604-3590
-------
SEDIMENT SAMPLING QUALITY ASSURANCE USER'S GUIDE
PROJECT SUMMARY
U.S. Environmental Protection Agency (USEPA) quality
assurance policy requires that every monitoring and measurement
project must have a written and approved quality assurance (QA)
project plan. Among the sixteen elements which must be contained
i.\ all QA project plans are the following:
o Project description
o 1QA objectives for measurement data in terms of
precision, accuracy, completeness, representativeness,
and comparability.
o Data analysis, validation, and reporting
o Specific routine procedures used to assess data
precision, accuracy, and completeness
This report, which is a companion to an analogous document
on soil sampling quality assurance, addresses selected factors
associated with the application of quality assurance/quality
control (QA/QC) guidelines to sediment sampling. In order to
make this report more self-contained, chapters from the companion
soil report covering such topics as sample handling, analysis of
QA/QC data, and system audits, which are equally applicable to
sediment sampling, are contained verbatim in the appendices.
The most important consideration for sediment sampling is
the objective for which the sampling is being done. The
statement of objectives should contain clear answers to the
following questions:
o How will the resulting data be used to draw
conclusions?
1
-------
o What actions may be taken as a result of those
conclusions?
o What are the allowable errors in the results?
Once answers to these questions are available an appropriate
statistical design for the sampling and analysis program, to
include an adequate and verifiable QA/QC project plan for the
study, can be devised.
Prior to the establishment of an adequate, cost-effective
QA/QC plan for sediment monitoring programs, a decision-making
official, after careful analysis of the consequences, must
specify allowable Type I and Type II errors in the results. A
Type I error, for a situation in which a measured population mean
is being compared to either an action level or a control level,
ie committed when it is concluded that the population mean
exceeds the action or control level when in fact it does not.
For the same situation, a Type II error is committed when it is
concluded that the population mean does not exceed the action or
control level when in fact it does. The desired minimum
detectable difference between a measured population mean and
either an action, or a control level must also be specified.
The goal of this document is to provide a flexible, but
technically sound, framework within which the user can devise a
QA/QC plan consistent with the specific objectives of any
sediment monitoring program. The document has been developed to
serve as a user's guide for anyone designing, implementing,
or overseeing sediment monitoring programs.
The extent to which adequate field-validated models exist
for describing sediment transport and deposition has a direct
bearing on the design of cost-effective sediment monitoring
programs. Generally, when adequate models exist, fewer
monitoring measurements are required to assess pollutant levels
and their significance. Accordingly, this report presents a
brief review of some available sediment transport models after
first providing some background definitions and discussions.
-------
The models range from simple, steady state, dissolved oxygen
relationships to very complex models describing the
interrelationships among pollutant additions and removals,
organic matter concentrations, and life processes occurring in
aquatic environments. Many pollutants can be transported in
suspended solid form or adsorbed on suspended particulates.
Unfortunately, the dynamics of the movement of pollutants
adsorbed on sediments is not well understood.
Sediments play an important role in the transport of
pollutants as well as in the transport of nutrients. Both the
pollution and nutrient aspects must be considered. Sediments can
overwhelm bottom fauna, but the nutrients they carry can give
rise to new biota.
In choosing an appropriate model, a comparison should be
made of available models. A model should be fitted to the
problem and not vice versa. If complete validated models are not
available for the pollutants and other site-specific conditions
of a problem, it still nay be possible to use portions of
available models, or other empirical field experience in the
cost-effective design of sediment campling programs.
The responsibilities of National Program Managers in the
USEPA Mandatory Quality Assurance Program include ensuring that
data quality acceptance criteria and QA Project Plans are
prepared for all data collection projects sponsored by their
offices.
This requires the development of data quality objectives
(DQOs). DQOs are qualitative and quantitative statements
developed by data users to specify the quality of data needed
from a particular data collection activity.
DQOs are the basis for specifying the quality assurance and
quality control activities associated with the data collection
process. QA Project Plans clearly describe what will be done at
each stage of data collection (i.e., cample cite celection,
sample collection, sample handling and analysis, and data
handling and analysis) and include instructions or standard
3
-------
operating procedures for each field and laboratory activity.
Some possible objectives for sediment sampling are:
o Determining the extent to which sediments act as either
sources or sinks for water pollutants,
o Determining presence and distribution of selected
pollutants in sediments in both space and time,
o Determining the risk to human health and/or the
environment from sediment contamination by selected
pollutants, and
o Taking measurements for validation of sediment
transport and deposition models.
Under most circumstances, background data will not be
available for a given monitoring location. These data must be
acquired before, or preferably during, any sediment monitoring
program. The intensity of the background sampling that is
undertaken depends upon the pollutants being measured, the
sediment characteristics and variability, the levels of pollutant
likely to be found in the study area and the purpose of the study.
QA/QC procedures are just as critical for the background
measurements as they are for the study area measurements.
When sediments are contaminated, drinking water or human
foods, contaminated directly or indirectly through contact with
sediments, may be unfit for human consumption. As the hazardous
constituents move through different trophic levels, substantial
biomagnification of contaminants may take place.
The steps outlined below are designed to provide a sediment
monitoring effort with minimal needed sample precision and
representativeness.
o Determine the components of variance that should be
built into the statistical design.
o Choose the allowable probabilities for Type I and Type
II errors and the difference in means considered to be
significant. (These are the DQOs and they are needed
together with an estimate of the coefficient of
-------
variation to determine the number of samples required
in each stratified region.)
o Obtain sampling data from studies with similar
characteristics to the one of interest. (Estimates of
coefficients of variation are of particualr
importance.)
o Calculate the mean and note the range of each set of
duplicates (co-located independent samples).
o Using results from previous studies, develop a table of
critical difference values for duplicate sample results
for various concentrations that span the range of
concentrations of interest. Use this table to accept
or reject sets of duplicates.
Suggestions for additional elements of a more complete QA/QC
plan are provided in the text.
The DQO guidelines below are suggested for the indicated
operational situations.
Confidence Power Relative
Level (1-6) Increase *
(l-o)
Preliminary Site
Investigation 70-80% 90-95% 10-20%
Emergency Cleanup 80-90% 90-95% 10-20%
Planned Removal and
Remedial Response
Activities 90-95% 90-95% 10-20%
* Relative Increase from Background or an Action Level to be
Detectable with Probability (1-6)
-------
Statistical sampling plans are based on assumptions
concerning the probability distributions of the measurements to
be made. The properties of a normal distribution are so
desirable that, if the data are not normally distributed, a
transformation is sought to convert the existing distribution
into a new distribution which is approximately normal.
The maximum probability allowed for a Type I error is called
the significance level of the test of hypothesis and is commonly
denoted by alpha (a). The probability of a Type II error is
usually denoted by beta (8) and is typically a function of a ,
sample size, and the size of the deviation from the null
hypothesis. The probability that the alternative hypothesis will
be accepted when it is true is called the power of the test and
may be denoted by (1-6). Typically, the experimenter will
specify the smallest deviation from the null hypothesis that he
considers to be scientifically, economically, or environmentally
important to detect and then specifies the power of the test that
he wants for that specific alternative.
The Quality Assurance Officer, supported by a qualified
statistician, should be intimately involved in the review of the
experimental or sampling design proposed by the" investigator. He
should insure that the information obtained provides measures of
the components of variance that are identified in the field.
Composite samples provide only an estimate of the mean of
the population from which the samples forming the composite are
drawn. Mo estimate of the variance of the mean, and hence, the
precision with which the mean is estimated can be obtained from a
composite of samples. Since the primary purpose of QA/QC is to
measure the precision of the samples obtained, the compositing of
samples should be avoided if at all possible.
Split samples, spiked samples and blanks are used to provide
a measure of the internal consistency of the samples and to
provide an estimate of the components of variance and the bias in
the analytical process. The number of QA/QC samples needed is
suggested as one out of every twenty samples for most categories
6
-------
of samples. In some instances this guideline may not be adequate
while in others it may provide more samples than are necessary.
It is good practice to perform an initial exploratory study
in which, among other things, QA/QC samples in excess of the
guideline recommendations are collected and analyzed. Analysis
of the resulting data will provide a better estimate of the
optimum required number of QA/QC samples of different types.
Typically, one wishes to estimate the concentration of
measured pollutants in the sediments and to indicate the
precision of these estimates. To indicate precision of an
estimate, one may provide the standard error or a confidence
interval for the expected value of the concentration. The
confidence interval is bounded by confidence limits. Confidence
limits are bounds of uncertainty about the average caused by the
variability of the experiment.
Prediction limits are similar to confidence limits but are
used to identify an interval into which a randomly chosen future
sample value should fall. Equations for both confidence and
prediction limits are provided along with an example calculation.
A problem that is particularly prevalent in data obtained
from field samples is that of outliers. The cause of the outlier
may be an error of procedure in sampling, subsampling, chemical
analysis, or the transcribing of data; or it may be due to an
anomaly that would indicate that a change is required in the
assumed model for the process. Guidelines are provided for
rejecting outliers, however, there are many problems with outlier
tests. If at all possible, prior to rejecting values as
outliers, repeat measurements should be made on the same or
nearly identical samples.
Once objectives have been defined which involve the need for
sediment sampling, the next step is to develop a total study
protocol including an appropriate QA/QC project plan. The
recommended approach is to conduct an exploratory study first
that includes both a literature and information search along with
-------
selected field measurements made on the basis of some assumed
transport model.
To provide a framework for the discussion, a hypothetical
situation involving an abandoned hazardous waste site is
described. The established objective for this hypothetical
situation is to conduct an environmental assessment of the site
and its environs to determine whether a short or long term hazard
to man or the environment exists. If a hazard exists, its nature
and extent must be defined and appropriate recommendations made
to bring the hazard under control A study team is organized to
address the problem and the sediment study group's task is to
identify and make an assessment of potential problems associated
with sediments in a nearby river and estuary.
Questions which must be answered, at least in part, by the
exploratory study include:
o What wastes have been placed at the disposal site over
what time periods?
o What chemicals in what amounts have escaped from the
site via what transport routes and what is the present
geographical extent of these chemicals?
o What adverse effects on human health or the environment
have been reported in the site vicinity?
o What is an appropriate background region to use for the
study?
Before taking any field measurements, a comprehensive literature
and information search should be conducted to determine what
information may already be available. The results of the
exploratory study will provide information and field data that
will serve as the basis for the design of a more definitive
•onitoring study. Thus, any field measurements taken should
include appropriate QA/QC measures to determine the quality of
the data.
The hypothetical case study is developed step by step. Data
quality objectives are identified, a grid system is defined, the
study area is stratified, a background region is selected, number
8
-------
and locations of sites for sampling are determined, and an
appropriate QA/QC project plan is prepared.
In general, the simplest sampling tool deemed to be adequate
should be used. The advantages and disadvantages of some bottom
samplers and some coring devices are presented in tables.
One of the possibilities for error during the sampling
process is discarding non-sediment material collected with the
sediment samples prior to analysis. It is suggested that all
such discarded material be retained. Ten percent of these
samples should be sent to the analytical laboratory for analysis
with the remainder being archived.
If the exploratory study is conducted well, it will provide
seme data for achieving the objectives of the study; it will
provide data concerning the feasibility and efficacy of most
aspects of the study design including the QA/QC plan; it will
serve as a training vehicle for all participants; and it will
pinpoint where additional measurements need to be made.
Following analysis and interpretation of the information and
data resulting from the exploratory study, the next step is the
design of the final definitive study. Any problems with the
QA/QC plan noted should be solved by appropriate modifications of
the plan. The procedure is illustrated by extending the
hypothetical case study based on assumed data obtained from the
exploratory study.
In view of conclusions reached on the basis of the assumed
data, the following questions which should be answered in the
definitive study are identified:
o How far down the stream are the sediments significantly
contaminated?
o What are the relative contributions of surface water
and groundwater to the contamination of sediments?
o How are the sediment levels changing as a function to
time?
o What levels of contamination in human foods are derived
directly or indirectly through contact with sediment?
9
-------
o What is the impact of contaminated sediments on aquatic
biota?
o How should the study area be stratified in the
definitive study?
A table is provided giving the number of samples required in
a one-side, one-sample t-test to achieve a minimum detectable
relative difference at confidence level (1-a) and power (1-8).
In this table the coefficient of variation varies from 10 to 35%,
the power from 80 to 95%, the confidence level from 80 to 99%,
and the minimum detectable relative difference from 5 to 40%. An
equation is provided to calculate values not included in the
table.
The required frequency of sampling depends on the objectives
of the study, the sources and sinks of pollution, the
pollutant(s) of concern, transport rates, and disappearance rates.
Assessment of trends in time will establish whether sediment
concentrations are increasing, decreasing, or remaining fairly
level. Evaluations of these trends will be important to
selection of appropriate remedial response measures.
The analysis and interpretation of QA/QC from the more
definitive study should show how all aspects of the total QA/QC
plan combine to give an overall level of reliability for various
aspects of the resulting data. Another goal nay be to determine
whether all QA/QC procedures used were necessary and adequate.
It is desirable to provide summarized tables of validated QA/QC
data in the final report. Prom such tables it is possible to
determine bias; precision; component random errors associated
with reproducibility, extract matrix, sample matrix, and sample
homogeneity; interlaboratory precision; and uncertainty.
Presentation of QA/QC data also contributes to the building of a
body of data in the literature which allows comparisons to be
•ade between and among studies.
Data from the more definitive study describing variations in
sediment concentrations with depth will show how effective
dredging to different depths might be in the removal of the
10
-------
contamination. If dredging is even contemplated, safe and
effective methods for disposing of the dredge spoil must be
available.
11
-------
SEDIMENT SAMPLING QUALITY ASSURANCE USER'S GUIDE
by
Delbert S. Barth and Thomas H. Starfcs
Environmental Research Center
University of Nevada, Las Vegas
Las Vegas, Nevada 89154
Cooperative Agreement
CR 810550-01
Kenneth W. Brown, Project Officer
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
Office of Research and Development
Las Vegas, Nevada
May 1985
-------
ABSTRACT
This report is intended to serve as a companion to an
analogous document on soil sampling quality assurance. Prior to
the design of an adequate QA/QC plan for sediment sampling there
must be agreement on the objectives of the sampling program.
Clear answers to the following questions should be available:
How will the resulting data be used to draw conclusions? What
actions may be taken as a result of those conclusions? What are
the allowable errors in the results? Once answers to these
questions are available an experimental protocol may be prepared
with an appropriate statistical design and QA/QC plan.
An overview of selected sediment models is presented to
serve as a foundation for stratification of study regions and
selection of locations for sampling sites, methods of sampling,
and sample preparation and analyses. Discussions of situations
relating to rivers, lakes, and estuaries are included.
Objectives of QA/QC plans are presented against a backdrop of
objectives for sediment sampling. A suggested minimal QA/QC plan
for sediment sampling is presented. In relation to different
operational situations suggested guidelines are given for Type I
and Type II errors and minimal relative differences from
background or action levels to be detected.
Statistical considerations presented include experimental
statistical designs to enable AMOVA to be accomplished,
discussion of Type I and Type II errors, numbers and locations of
sampling sites, bias, confidence and prediction limits, outliers,
and testing of hypotheses. Some examples are given to illustrate
the principles. The importance of an exploratory study to the
cost-effective achievement of the overall objectives of a
sediment sampling program is emphasized. A hypothetical case
study related to an abandoned hazardous waste site is defined.
Study objectives are presented. An exploratory study is
designed, implemented and hypothetical data presented. The
hypothetical data are then used to design a final more
definitive study to achieve the objectives.
iii
-------
NOTICE
The information in this document has been funded wholly or in
part by the United States Environmental Protection Agency under
Cooperative Agreement CR 810550-01 to the Environmental Research
Center. It has been subject to the Agency's peer and adminis-
trative review, and it has been approved for publication as an
EPA document.
ii
-------
5. Exploratory Study 70
Introduction 70
Number and Location of Sites 72
Sampling and Sample Handling 75
Analysis and Interpretation of Soils 80
6. Final Definitive Study 81
Introduction 81
Selection of Numbers of Samples and Sampling Sites . 84
References 96
Appendices
A. Percentiles of the T Distribution 100
B. Sample Handling and Documentation 101
C. Analysis and Interpretation of QA/QC Data 108
D. System Audits and Training 115
-------
TABLE OF CONTENTS
Notice ii
Abstract iii
Figures vi
Tables vii
1. Sediment Sampling Quality Assurance User's Guide . . 1
Introduction 1
Background 3
Objectives 5
Audience 6
Approach 6
2. Modeling Sediment Transport and Deposition 8
Introduction 8
Transport and Sedimentation 10
Rivers 13
Lakes 15
Estuaries 17
Modeling Theories , 20
Conclusions 27
3. Objectives of Quality Assurance-Quality Control Plans 28
Introduction 28
General Identification of the Objectives 34
Objectives for Background Monitoring 37
Specific Objectives for Monitoring in Support of
CERCLA 38
Preliminary Site Investigation 41
Emergency Cleanup 42
Planned Removal and Remedial Response Studies ... 43
Monitoring or Research Studies 44
4. Statistical Considerations 45
Introduction 45
Distribution of Sediment Sampling Data 45
Statistical Designs 47
Type I and Type II errors 48
Number and location of samples 52
Role of quality assurance in experimental design . 52
Components of variance 53
Compositing of samples 55
Split samples, spiked samples and blanks 56
Data Analysis 58
Bias 58
Confidence and prediction limits €2
Outliers 64
Testing of hypothesis 66
Statistics associated with biological monitoring. . 69
iv
-------
TABLES
number Page
1. Wentworth particle size scale 9
2. Overview of selected water quality models 21
3. Analysis of variance of nested sediment sample . 54
4. QA/QC procedures for sediment samples 59
5. PCS study to determine contamination of an area 67
(hypothetical data)
6. Common measurements for surface water, aquatic
organisms and sediment sampling 76
7. Comparison of bottom grabs/samplers 77
8. Comparison of coring device . . , 78
9. Summary of selected hypothetical results from the
exploratory study 82
10. Number of samples required per stratified region as
a function of indicated parameters 84
11. Number of samples required in a one-sided one-sample
t-Test to achieve a minimum detectable relative
difference at confidence level (1-a) and power of
(1-6) 85
12. New stratified regions for the more definitive study 87
13. Sampling containers, preservation requirements, and
holding times for sediment samples 90
B-l. Sampling containers, preservation requirements, and
holding times for soil samples 103
B-2. Accountable document control requirements .... 106
vii
-------
FIGURES
Number Page
1. Block diagram of a braided stream and adjacent
environments 14
2. Block diagram of a meandering stream showing major
depositional environment 14
3. Four stage diagram showing stratification and
overturn periods for a dimictic lake 16
4a. Acceptance region for H:yo«30.0 51
4b. Type II or 8 error 51
5. Sketch map of river showing stratified regions and
sampling points 74
VI
-------
CHAPTER 1
SEDIMENT SAMPLING QUALITY ASSURANCE USER'S GUIDE
INTRODUCTION
•• U. S. Environmental Protection Agency (USEPA) quality
as trance policy requires that every monitoring and measurement
prc.ect must have a written and approved quality assurance (QA)
project plan (USEPA, 1980). The sixteen elements which must be
contained in all QA project plans are listed below with some
brief explanatory notes.
(1) Title page with provision for approval signatures.
(2) Table of Contents. (This must include a serial listing
of each of the 16 QA project plan components.)
(3) Project description. (A general description of the
project should be provided together with the intended
end use of the acquired data.)
(4) Project organization and responsibility. (List the key
individuals, including the QA officer, who are
responsible for ensuring the collection of valid
measurement data and the routine assessment of
measurement systems for precision and accuracy.)
(5) QA objectives for measurement data in terms of
precision, accuracy, completeness, representativeness,
and comparability. (For each major measurement
parameter list the QA objectives for precision, accuracy
and completeness. All measurements must be made so that
-------
results are representative of the media and conditions
being measured.)
(6) Sampling procedures. (For each major measurement
parameter(s), including all pollutant measurement
systems, provide a description of the sampling
procedures to be used.)
(7) Sample custody. (Where samples may be needed for legal
purposes, "chain-of-custody" procedures will be used.)
(8) Calibration procedures and frequency. (Information
should be provided on the calibration standards to be
used and their source(s).)
(9) Analytical procedures. (Describe the analytical
procedures to be used for each major measurement
parameter.)
(10) Data analysis, validation and reporting. (This will
include the principal criteria that will be used to
validate data integrity during collection and reporting
of data as well as methods used to treat outliers.)
(11) Internal quality control checks. (Examples of items to
be considered include: replicates, spike samples, split
samples, control charts, blanks, internal standards,
•pan gases, quality control samples, surrogate samples,
calibration standards and devices, and reagent checks.)
(12) Performance and systems audits. (Each project plan must
describe the internal and external performance and
systems audits which will be required to monitor the
capability and performance of the total measurement
system(s).)
(13) Preventive maintenance. (This should include a schedule
of important preventive maintenance tasks as well as
inspection activities.
(14) Specific routine procedures used to assess data
precision, accuracy and completeness. (These procedures
should include the equations used to calculate
precision, accuracy and completeness, and the methods
2
-------
used to gather data for the precision and accuracy
calculations.)
(15) Corrective action. (This must include the predetermined
limits for data acceptability beyond which corrective
action is required as well as specific procedures for
corrective action.)
(16) Quality assurance reports to management. (These reports
should include a periodic assessment of measurement data
accuracy, precision and completeness as well as an
identification of significant QA problems and
recommended solutions. USEPA, 1980)
In this report some of the factors associated with the
application of these general guidelines to sediment sampling will
be addressed.
BACKGROUND
This report is intended to serve as a companion to an
analogous document on soil sampling quality assurance (Barth and
Mason, 1984). While considerable effort is expended to make this
report self-contained, it is not considered desirable to repeat
all the applicable detailed discussions and explanations
contained in the soil sampling report.
The most important consideration for sediment sampling, as
for sampling any other media, is the objective for which the
sampling is being done. The statement of objectives should
contain clear answers to the following questions:
o Bow will the resulting data be used to draw conclusions?
o What actions may be taken as a result of those
conclusions?
o What are the allowable errors in the results?
Once answers to these questions are available, an
appropriate statistical design for the sampling and analysis
program must be devised. This statistical design should yield
3
-------
data from which an analysis of variance components may be done.
The analysis of variance should identify components of variance
associated with sampling, sample preparation, extraction, and
analysis.
The statistical design of the experiment should incorporate
an adequate and verifiable quality assurance/quality control
(QA/QC) program for the overall study. Control is defined as the
system of activities required to provide a quality product,
whereas quality assurance is the system of activities required to
provide assurance that the quality control system is performing
adequately. It cannot be overemphasized that an adequate QA/QC
program cannot be tailored for a study until a clear statement of
monitoring objectives, together with allowable errors, has been
provided.
Often actions may not be taken on the basis of monitoring
measurements in a single medium such as sediments. If one is
concerned about risks to human health or the environment, for
example, concentrations of hazardous substances in sediments may
not provide sufficient information on which to base the magnitude
and extent of necessary control actions. For such a risk
analysis it nay be necessary in addition to measure
concentrations of hazardous substances in surface waters,
groundwater, and foodstuffs to obtain some measure of the
biological availability of the hazardous substances in sediments
which can be related to potential exposures via various routes.
In cases in which sediment sampling is only a part of the total
monitoring program, it is mandatory to modify the QA/QC program
to cover all aspects of the total program to ensure that the
total combined errors in the final results will not exceed
allowable errors (McNeils et al., 1984).
Prior to engaging in a more detailed discussion of QA/QC
aspects for sediment sampling, it is desirable to present and
discuss some possible sediment monitoring objectives.
Objectives of sediment sampling may include:
-------
o Determining the extent to which sediments act as either
sources or sinks for water pollutants,
o Determining presence and distribution of selected
pollutants in sediments in both space and time,
o Determining the risk to human health and the
environment from sediment contamination by selected
pollutants, and
o Obtaining measurements for validation of sediment
transport and deposition models.
Further discussion of these objectives in Chapter 3 includes
some hypothetical examples related to different environmental
protection laws.
To establish an adequate, cost-effective QA/QC plan for a
sediment monitoring program, it is necessary for a
decision-making official after careful analysis of the
consequences, to specify allowable Type I and Type II errors in
reaching conclusions based on sample data. A Type I error, for a
situation in which a measured population mean is being compared
to either an action level or a control level, is committed when
it is concluded that the population mean exceeds the action or
control level when in fact it does not. For the same situation,
a Type II error is committed when it is concluded that the
population mean does not exceed the action or control level when
in fact it does. See Chapter 4 for additional discussion of Type
I and Type II errors. The political, social, and economic
consequences of making either a Type I or Type II error must be
weighed before a decision-making official can establish allowable
frequencies for each type error.
OBJECTIVES
This document is intended to serve as a user's guide that
identifies and explains selected principles and applications of
the methods and procedures for establishing an adequate QA/QC
program for sediment sampling aspects of environmental monitoring
5
-------
programs. It is not intended to serve as a guide for identifying
all sediment sampling equipment or to serve as a sediment
sampling protocol. Similarly, it is not intended to provide
•cook book" type details for the development and implementation
of a universal QA/QC plan for all sediment monitoring programs.
The goal is to provide a flexible, but technically sound,
framework within which the user can devise a QA/QC plan
consistent with the specific objectives of any sediment
monitoring program.
Mo detailed treatment of analytical quality assurance
procedures is given since that important aspect of the overall
problem has been adequately treated elsewhere (USEPA, 1982;
DStfPA, 1984). It should be noted, however, that in a QA/QC sense
sampling procedures are not fully separable from analytical
procedures. This is particularly true for sample collection and
handling procedures. Thus, sediment sampling QA/QC procedures
presented here should be viewed as important integral elements of
the overall QA/QC plan.
AUDIENCE
This document has been developed to serve as a user's guide
for anyone designing, implementing, or overseeing sediment
monitoring programs. It is especially applicable for personnel
responsible for regulatory programs involving sediment monitoring.
Special attention is given to sediment sampling examples related
to CERCLA since such applications are deemed of high priority for
sediment sampling programs. Many of the principles and
procedures discussed, however, are applicable to other situations
as well.
APPROACH
In Chapter 2 a brief overview of models describing the
dynamics of sedimentation in different bodies of water is
6
-------
presented. Knowledge of sediment dynamics provides a firmer
foundation for the design of sediment monitoring programs and
associated QA/QC plans and assists in the interpretation and
evaluation of the resulting data. Chapter 3 provides examples of
some hypothetical sediment monitoring situations together with
discussions of required QA/QC plans. Chapter 4 contains selected
applicable statistical methodology.
The role of an exploratory or preliminary study prior to the
performance of the definitive study is described in Chapter 5.
Chapter 6 describes how to determine for the final definitive
study the required number of sediment samples and sampling sites
consistent with established allowable probabilities for Type I
ard Type II errors and the desired minimum detectable difference
between means and either control levels or action levels.
Chapter 6 also discusses sediment sample collection, sample
handling, and analysis and interpretation of QA/QC data.
The subjects of systems audits and training are not
addressed in this document. The treatment of these subjects in
the companion volume (Barth and Mason, 1984) is considered to be
equally applicable to sediment sampling. In order to make this
report more self-contained, the entire chapters on sample
handling and documentation, analysis and interpretation of QA/QC
data, and systems audits and training from the companion soil
document are included in Appendices B, C and D, respectively.
-------
CHAPTER 2
MODELING SEDIMENT TRANSPORT AND DEPOSITION
INTRODUCTION
In determining the appropriate model to use in describing
the role of sediments in the transport and fate of hazardous
substances, one must have a definition of sediments along with
site-specific characteristics for sites of interest. For areas
of concern, i.e., rivers, lakes, and estuaries, sediments and
related data of importance will have general (geological
strata, soil type, climate, etc.) as well as specific (flow
rate, bed load, water pH, etc.) characteristics. The term
sediment is defined as any particulate matter which can be
moved by water, to or from a land surface and into or through
the waterways of a river basin, a lake system or an estuary
(Leytham and Johanson, 1979). Particulate sediment matter is
usually partially made up of once-living organic material in
various degrees of decomposition with particle sices ranging
from colloidal humus to large pieces of material. Sediments
normally contain some mineral particles. These may include any
of the three major rock types: igneous, metamorphic or
sedimentary rocks. The size of these particles can range from
that of clays through silts and sands to large boulders. A
size classification scheme has been developed by Wentworth and
is shown in Table 1.
Total sediments are the sum of suspended and bed-load
sediments. Suspended sediments occur mainly in slower moving
waters of sluggish rivers, lakes and estuaries. Suspended
sediments may have more long-term adverse effects on ecosystems
8
-------
Table 1. WENTWORTH PARTICLE SIZE SCALE
121
Unr
Immll
Snull
-II
-10
- 7
- 5
- 3
- 2
- I
0
4 I & C"~
42 290
49 129.
CotIK
Medium
Fine
Vv>finr
VV)CMTK
'to
4 3
4 •
4 •
4 9
II
F«e
V«>ftDf
•ouldert
O
*
m
Cnnuln
Smd
1
o
Source: Davis, 1983
-------
than bed-load sediments. These sediments can increase
turbidity of the water and therefore decrease sunlight
availability to the primary producers, as well as limit
visibility of predators. They can also clog filtering devices
of molluscs and fish (Farnesworth, et al., 1979).
Bed-load sediments are more significant in the faster
moving waters of river systems. These sediments can scour,
abrade and bury all or part of the benthic organisms, thus
modifying the food chain (Farnesworth, et al., 1979). They can
even modify the habitat structure. The effects of sediments in
general can be propagated throughout an ecosystem and may
result in the mass movement of organisms out of an area. This
is not to say sediments are always negative factors to an
ecosystem; sediments may carry nutrients into an area, thereby
increasing biological productivity. Most negative sediment
impacts are observed after runoff episodes associated with
storms or snow melt.
Sediments may readily adsorb pollutants. The dynamics of
pollutant movement on adsorbed sediment are not well
understood; however, research is ongoing to elucidate such
transport. Some of the factors involved include concentration
of the dissolved pollutants, flow velocity of the water,
kinetic adsorption coefficients, and depth of flow (Krenkel and
Novotny, 1980).
The process of adsorption-desorption of pollutants on
sediments has a direct effect on the transport processes and on
the bioavailability of the pollutants (OECD, 1981). Sediments
will have varying reaction phases with pollutants, depending
upon the sediment's chemical makeup and certain environmental
factors (temperature, pressure, water flow rate, etc.).
TRANSPORT AND SEDIMENTATION
The first factor to consider is the texture of the
sediments. Sediment texture has a number of characteristics.
10
-------
Particle size of the sediments is important; sediments can
either be homogeneous or heterogeneous with regard to particle
size. The particle's shape and surface characteristics are
important in determining whether and to what extent pollutants
are adsorbed. Porosity and permeability are two important
properties of sediments.
Sedimentation processes include: 1) Biological processes,
2) Organism-enhanced sedimentation and 3) Physical processes.
In biological processes, two important factors predominate.
They are degradation, which is the working and reworking of the
sediment by biological organisms, and pelletization, which is
the accumulation of biological excrement. In organism-enhanced
sedimentation, it is the bottom-rooted plant life that promotes
trapping and deposition of sediments. Physical processes are
by far the most important. These include in particular fluid
flow characteristics in relation to the settling of different
type and size particle. In fluid flow, there are two different
types of flow: 1) laminar flow and 2) turbulent flow. Either
the Reynold's number or Froude's number may be used to
characterize the flow as laminar or turbulent (Davis, 1983).
Stoke's Law of settling identifies and relates the
different variables involved in the settling of particles
(Davis, 1983). Unfortunately, Stoke's Law tends to be valid
for only a single particle, and concentrations of sediment tend
to retard the total settling.
For a specific sized particle of a specific shape and
density, there is a minimum fluid velocity needed to move that
particle. This minimum velocity is known as the threshold
velocity. There are several important mechanisms involved in
the movement of sediment particles in fluids. Traction defines
the mechanism whereby particles may slide or roll over the
substrate, and is particularly important on the bottom where
particles are in contact with one another. Saltation is
transport whereby the grains bounce or hop along the substrate
11
-------
and it usually accompanies traction processes. Both traction
and saltation processes contribute to the bed load. Bed load
nay be defined as the sediment load that moves by traction
and/or saltation along the bed as the result of shearing at the
boundary of flow (Dav.is, 1983). Suspended sediment load is
comprised of particles in the main flow of the current that
move significant distances without contact with the bottom or
side substrata. Maximum transport of sediments occurs mainly
during turbulent flow, such as that which occurs during storm
or snow-melt periods.
The sediment texture (or particle size distribution) is
directly related to the hydraulics of the system. The most
prominent cause contributing to observed sediment texture is
a change in the competence or capacity of a stream, which
causes sediment particles to come to rest. The coarsest
particles are present in the traction population of sediments.
The saltation sediment population contains the bulk of the
sediments with the particles therein being well sorted. The
sorting is due to the differential efficiencies of continued
suspension and redeposition while particles bound along. The
suspended load of a sediment sample shows considerable
variation due to both the intensity of turbulence and the
original characteristics of source sediments, such as cohesion
and flocculation. Sorting within this population is poor.
Turbidity currents occur when fluid turbulence causes
sediments to become suspended. Turbidity currents can occur in
deltaic regions and also in estuaries. Liquified sediment
flows occur when sediment is supported by upward-flowing fluid
as particles settle. Debris flows are a mixture of fine
sediments and fluid which support larger particles. These
usually occur off mountain sides. A slump occurs when masses
of soil move along shear planes. These often occur on the
sides of rivers and also are types of •mud" flows.
12
-------
Rivers
Two main types of rivers are found in the world today.
One type, the braided stream (stream will be used synonymously
with river), is or has been a predecessor to the second, the
meandering stream.
A braided stream has numerous channels that are separated
by bars and small islands. The deposition of sediment is
characterized by the shifting of the channels and bar
aggradation. These types of streams have an overabundance of
sediments. Streams are braided due to the inability of the
stream to move the coarse component of its load (Davis, 1983).
K wever, during floods, all sized particles are moved. There
i. ,
9. - four types of events in which sedimentation occurs in
br *ded streams: 1) flooding, 2) lateral accretion - side or
point bars develop, 3) channel aggradation - due to the waning
energy of the stream and 4) reoccupation of an older channel
causing cut and fill. Examples of braided streams include the
Trollhiem River in California, the Platte River in Nebraska and
the Bijou Creek in Colorado. Models (geologic) have been based
on these rivers. Figure 1 shows a block diagram of this type
of stream.
A meandering stream is a single channeled stream that
displays a relatively ordered condition of riverine and
sediment accumulation processes. These are commonly situated
downstream from braided streams. They lack gravel, have a
modest suspended load and have a broadly meandering pattern.
These types of streams are commonly found on coastal plain
regions flowing more or less perpendicular to the coast. They
have specific sedimentary deposits such as levees, floodplain
and point bar deposits. These streams are characterised by
turbulent flow, and sediment is transported in both bed load
and suspended load. Sediment is commonly eroded from one bank
and accreted on another downstream. Examples of meandering
13
-------
Figure 1. Block diagram of a braided atreair. and adjacent
environments. Source: Davis, 1983,
After Williams and Rust, 1969
Figure 2. Block diagram of a meandering stream shoving
major depositional environment. Source: Davis, 1963
14
-------
streams are the Mississippi River, the Ohio River, and the
Colorado River. The Colorado is an excellent example of a
braided stream becoming a meandering stream. Figure 2 is a
block diagram of this stream type.
Deltas are accumulations of sediment at the end of a
river channel where it discharges into a standing body of water.
Deltas can occur in oceans, lakes and estuaries. Erosion of a
delta can be dominant at times, with the primary agents being
waves and/or currents. The processes that act upon a marine
delta are riverine processes and marine processes.
In riverine processes, three primary forces are generally
dominant: 1) inertia, 2) bed friction and 3) buoyancy.
Circumstances leading to the formation of deltas occur in
lakes, estuaries, and enclosed seas in which there are broad,
flat, offshore slopes.
In marine processes there are three dominant forces: 1)
tides, 2) waves, and 3) coastal currents. The Mississippi
delta is a major example for which a model has been developed.
Lakes
Lakes occur throughout most climatic belts of the world
and receive large volumes of sediments. Most lake studies
emphasize the biological, chemical and physical aspects of the
environment. Only relatively recently have lake sediments been
given the major consideration due them.
Depending upon a variety of environmental factors, lakes
may stratify in the summer and in the winter. Figure 3
illustrates the process and the mechanism whereby mixing may
occur in spring and fall months.
The Great Lakes are so large that the circulation caused
by the cooling and sinking of maximum density water, which is
replaced by deeper water, is not sufficient to cool the whole
15
-------
!•)
wwr feting •v*nu<>
fe) Stf«tif«d wfir-
I
*C
fc)
Wind'
Ml
Figure 3. Four stage diagram shoving stratification and
and overturn periods for a dinictic lake.
Source: Davis. 1983, After: Hough. 1958
-------
lake body to maximum density, and hence they never completely
freeze over (Garrels et al., 1975). Stratification in large
lakes such as the Great Lakes occurs only in the summer.
During stratification, if enough organic material exists
in deep water, oxygen can disappear completely. This produces
changes in the bottom fauna and promotes production of gases
such as hydrogen sulfide (H2S) and methane (CB4>. Shallow
lakes are stirred by wind and waves, thereby minimizing
stratification, but lakes of intermediate depth are very
susceptible to stratification and oxygen deficiency. Excessive
plant nutrients promote plant macrophyte growth which aids in
the deoxygenation process in small lakes by reducing wave
action and thus mixing. This can lead to a lake being
overwhelmed by organic material.
There are two main types of sediments other than organic
material found in lakes. One, terrigenous sediments, can
originate from two main sources, either from the edge of the
lake itself or from being transported in by other means, i.e.,
rivers and waste water. The second sediment type is composed
of chemical precipitates and comes from the water constituents
themselves. There are two categories of lakes based on their
chemical constituencies: 1) saline lakes and 2) carbonate
lakes. Waste water can add chemicals to the water of either
category and form various types of precipitates.
Estuaries
There is a wide variety of morphology, hydrodynamics and
sediment distribution in estuaries. Four main morphological
types of estuaries are known: 1) drowned river valleys, 2)
fjords, 3) bar-built estuaries and 4) tectonically produced
estuaries. Widely distributed, irregularly shaped estuaries
are common along coastal plains as a result of drowned river
17
-------
valleys from the sea level rising in the Holocene period.
Chesapeake Bay and Delaware Bay are examples of this type.
This estuary type is characterized by rapid sediment
accumulation. Fjords are deep, steep-sided estuaries carved by
glaciers and are characterized by poor circulation and slow
sediment accumulation. Fjords tend to be small and are
typically developed on tectonically active coasts. As
estuaries develop and coastal processes transport sediment
along, it is common to develop spits and barriers that can
partially or completely close the mouth of the estuary, with
tidal inlets interrupting an otherwise continuous barrier. The
large estuaries behind the outer banks of North Carolina
represent this type. Tectonically produced estuaries are
generally confined to leading-edge coasts where faulting and
subsidence create embayments. San Francisco Bay is such an
estuary (Davis, 1983).
Two main processes are found to be of great importance to
sediment accumulation in estuaries. Tidal currents, which
constitute the first process, are directly related to tidal
range in most instances. The size of the inlet into an estuary
is also important, as well as the speed of the current into and
out of the inlet. Sediment from freshwater runoff as well as
from oceanic processes must be considered. A sudden decrease
in the current speed at the landward or seaward side can cause
rapid accumulation of sediment. Riverine processes constitute
the second main factor contributing to sediment accumulation.
Since an estuary is a standing body of water, a delta can form
at a river's mouth. If the tidal currents are not sufficient
to remove the sediment, accumulation occurs. It is due to
these two processes that estuaries are generally short-lived
geologically.
Estuary circulation is primarily based upon the sone in
which freshwater comes into contact with seawater. There are
three estuary types based upon the nature and distribution of
18
-------
this cone. A highly stratified or salt wedge estuary is one in
which there is little nixing of the waters and a density
stratification occurs. River discharge must be the dominant
process in the formation of this estuary (Pritchard, 1955).
Mixing only occurs by vertical advection in the shear zone
between the two opposing masses (Biggs, 1978). Sediment
carried to the estuary from the stream may settle into the
salt-wedge layer and be transported to the landward tip for
deposition. Well-stratified estuaries display a complicated
circulation which is related to the Coriolis effect. During
flood tides, the interface of the water masses is tilted up on
the right side of the estuary in the northern hemisphere as one
looks landward, and in ebb tide it is tilted to the left side
k,
(Davis, 1983). This results in a circular flow component in
whic*h the'center is a null point. Partially mixed estuaries
are ones in which tidal influence is dominant in determining
circulation and mixing of waters. Turbulence created by tidal
action causes downward movement of freshwater as well as upward
movement of seawater (Pritchard, 1955). This results in a
gradual increase of salinity from top to bottom. Suspended
sediment tends to concentrate in the area of maximum turbidity
which is located just downstream from the landward limit of
seawater intrusion. When riverine and tidal processes are
equal in importance, a totally nixed estuary will result. The
Coriolis effect also plays a role in circulation and
sedimentation of these estuaries. These estuaries are
vertically homogeneous. Sediment will follow the pattern
provided by the Coriolis effect with marine sediments
concentrating on the right (looking landward from the sea),
and river sediments concentrating on the left (Biggs, 1978).
The nodels reviewed in the next section will demonstrate
general principles and how they apply to sediment campling.
Pew nodels are based on the sediments alone; most include the
system as a whole.
19
-------
MODELING THEORIES
Mathematical models of systems are often a'useful net hod of
generating and evaluating the various outcomes. A model,
however, should not be considered valid until it has been
substantiated by field and/or laboratory measurements (Krenkel
and Novotny, 1980). Table 2 presents an overview of some
commonly used models. The range and choice of available models
is clearly quite broad.
The following guidelines have been taken from Grimsrud
et al., 1976 on the selection and use of models:
1) Define the problem and determine what information is
needed and what questions must be answered.
2) Use the simplest methods that can provide the answers to
your questions.
3) Use the simplest models that will yield adequate
accuracy.
4) Do not try to fit the problem to a model but select a
model to fit the problem.
5) Do not confuse complexity with accuracy.
6) Always question whether increased accuracy is worth the
increased cost and effort.
7) Do not forget the assumptions underlying the model used,
and do not read more significance into the simulation results
than are actually there.
Stream (river) as well as lake and estuary models tend to
be based upon a one-dimensional approximation of the flow,
momentum and mass conservation equations. These models put
more emphasis on convective transport of pollutants than on
dispersion. The models range from simple, steady state,
dissolved oxygen relationships to very complex models
describing the interrelationships among pollutant additions and
removals, organic matter concentrations, and life processes
occurring in aquatic environments (Krenkel and Novotny, 1980).
20
-------
Table 2. Overview of Selected Mater Quality Models
tWMM
ITOtM «r« tn fd
U«rt tuw> Am>
DOSAG 1cu>»H Or.
II EPA
•WMM tE W« Kn Ul CPA
CCI\
O Ot«tr«
MR •DM*'**
B 0«ran.«Mr* tor
NSr II THAN
MEL OCAL
M I T N«i MIT
• I T
1 1
H.IMI
D
i
i
DO.*"
•Mf*«cMir hMfMMMri tWrAfee C*
Dt»i
TX
MMHrkiMii>lMiiUM«rT«ctaol«t Dtp ^ Cml Itfmtntt Ci
Source: Krenkel and Novotny, 1980
21
-------
Sediments in some instances are considered pollutants.
Discharge limitations have been imposed for suspended solids.
Many pollutants can be transported in suspended solid form or
adsorbed on suspended part iculates . Unfortunately, the
dynamics of the movement of pollutants adsorbed on sediments is
not well understood.
The description and solution of the hydrodynamic behavior
of surface or groundwater systems are essential parts of every
water quality model. Basic hydrodynamic laws which must be
included in descriptions of water quality systems are: 1) the
water conservation equation (the equation of continuity) and 2)
the momentum conservation equation (equation of motion)
(Krenkel and Novotny, 1980). The water conservation equation
states that the difference of the flow entering and leaving a
control volume must equal the rate of storage in the volume.
The applicable partial differential equation is:
- q.
3t
where
A is the cross-sectional area
t is time
Q is the flow
x is the direction of flow
Si is the lateral inflow into the control volume per
unit path length in the direction of flow.
If one multiplies each term in this equation by a unit of path
length in the direction of flow, it can be seen that
•|£ represents rate of storage, ~ outflow rate, and q±
Q t wX
lateral inflow rate; or, rate of storage * lateral inflow rate
- outflow race.
The momentum conservation equation is based upon Newton's
22
-------
second law of notion which states that the rate of change of
momentum equals the sum of external forces acting on the
control volume. The applicable partial differential aquation
is as follows:
+ J_ C(U) (UH)] + gH JH - gH(S0-Sf )
at ax x
where
U is flow velocity
t is time
x is the direction of flow
g is gravity acceleration
H is the depth
S0 is the bottom slope
Sf is the energy (friction) slope of the flow (may be
obtained from semiempirical flow formulas)
If one multiplies each term of the equation above by the water
density p and Ax / the terms in the equation have the following
meaning:
3_ (pOH) Ax * rate of change of momentum in a control volume
3t
pg _ [(U)(UB)lAX « difference between rate of momentum entering
3x and that leaving a control volume
pgH BH AX • net hydrostatic pressure of the surrounding
dx water on the control volume
pgHS0 AX » gravity force due to the weight of the control
volume
p gHSf Ax » friction shear resistance force
23
-------
In words, the equation states that for a control volume of
water the difference between the rate of momentum entering and
leaving plus the rate of change of momentum inside the control
volume is equal to the sum of the external forces acting on the
control volume.
Suspended particles originate from soil erosion, bank
erosion, urban solids, washload and organic life processes
(KrenJcel and Novotny, 1980). The channel phase of sediment
transport can be divided into the suspended fraction and the
fraction of sediments contained by moving streambeds. In
suspended sediment transport analysis, it is important to
determine where and when a particle will settle or when and
where the bed particles will be resuspended. Stoke's Law is
the general basis for sedimentation.
The equations of continuity and motion remain the same in
any sediment transport model. The mass balance equation for
pollutants (i.e., phosphorous, heavy metals, adsorbed
pesticides) must be coupled with sediment transport since
adsorption or release may take place between the adsorbed and
dissolved pollutant phases. The adsorbed component moves with
the sediment and is therefore subject to any processes that may
influence the sediment. The exchange of matter between the
bottom deposits and overlying water is governed by adsorption
equilibrium and limited by the diffusion velocity through the
bottom boundary layer. Two phases described when giving
general mass balance equations for adsorbed pollutant movement
are the free phase and the sorbed phase. The coupled equations
for each are as follows (KrenJcel and Novotny, 1980):
Free phase: ^--U^C-P^Sl1!*- KflC
at ax at
Sorbed phase: $S - K8(Se-S) - KSSS * M)/H
at
24
-------
where
C is the concentration o5. the dissolved pollutant
-------
the rate of change in concentration of a dissolved pollutant -
rate of loss + rate of loss + rate of gains + rate of
by flow to adsorption or losses from loss of the
(convective on suspended sources or pollutant
transport) solids sinks from the
respectively system by
processes
not
otherwise
accounted
for
In words, the equation for the sorbed phase states that
the rate of change in concentration of the pollutant adsorbed
on suspended solids ~
Rate of gain of adsorption + Rate of loss due
(driven by the difference to settling + rate
between the adsorption of loss due to the
equilibrium concentration scour rate of the
and the actual adsorption pollutant adsorbed
concentration) on the sediments.
Use of the cited equations plus others is very important
when developing a model of sediment/pollutant relationships.
The development of the CHANL model by the U.S. Environmental
Protection Agency (USEPA) has demonstrated the process.
The basic equations of any model must all be defined.
Also, exact definition of the solution being sought is needed
before an appropriate model can be selected to solve the
problem.
26
-------
CONCLUSIONS
When using or developing mathematical models all the
parameters must be chosen carefully. Sediment, in this case,
is very important but is linked to many other parameters.
Knowledge of these parameters is imperative when deciding which
model is to be used and how the results will be displayed.
Sediment plays an important role in the transport of
pollutants as well as in the transport of nutrients. Both the
pollution and nutrient aspects must be considered. Sediments
can overwhelm bottom fauna, but the nutrients they carry can
give rise to new biota. By the same token, sediments can
transport pollutants that are hazardous to some life forms of a
particular waterway.
In choosing an appropriate model, a comparison should be
made of available models. A model must be fitted to the
problem and action taken accordingly. Many good models exist,
but only the ones which contain sediment factors will be
adequate for our needs here.
27
-------
CHAPTER 3
OBJECTIVES OF QUALITY ASSURANCE-QUALITY CONTROL PLANS
INTRODUCTION
USEPA Order 5360.1 establishes the responsibilities of
National Program Managers in the Agency's Mandatory Quality
Assurance Program. These responsibilities include ensuring
that "data quality acceptance criteria* and QA Project Plans
are prepared for all data collection projects sponsored by the
office. In a memorandum of April 17, 1984 accompanying the
issuance of Order 5360.1, Deputy Administrator Aim identified
two steps that must be taken to ensure that all data collected
by USEPA are suitable for their intended use:
"...the user must first specify the quality of data he
needs; then the degree of quality control necessary to assure
that the resultant data satisfy his specifications must be
determined."
The first step is accomplished through the development of
Data Quality Objectives (DQOs). Data Quality Objectives are
qualitative and quantitative statements developed by data users
to specify the quality of data needed from a particular data
collection activity (USEPA Draft, 1984).
DQO development is an iterative process involving both
decision makers and technical staff. DQOs, which are
statements of the quality of data needed to support a specific
decision or action, are developed in three general stages.
First, the decision maker and the technical staff discuss the
problem being addressed, the resource and time constaints for
addressing the problem, and the information needed. Second,
28
-------
the decision maker and the technical staff discuss specific
questions developed by the staff to clarify what information is
needed, how the information will be used, and what limitations
of the information will be acceptable. Third, the technical
staff develops possible approaches for collecting the necessary
data and determines the quality of the data that can be
expected from each approach. The outcome of the third stage is
the decision maker's selection of the specific approach that
will be used and the statement of the DQOs for that approach.
The quality of a data set is represented in terms of five
characteristics of the data: precision, accuracy, represent-
ativeness, completeness, and comparability.
The objectives of a study or monitoring program should
include the following concepts:
o What information is needed and what function the
information serves in addressing the problem;
o How the information will be used, in terms of the
types of conclusions that are anticipated from the
data and the criteria that will be used to make
decisions;
o The limitations and applicability of the data, in
terms of the universe to which the conclusions and
decisions will apply;
o How conclusions based on the data can be in error and
what level of risk of making incorrect or questionable
• decisions is acceptable;
o The time and resource constraints for data collection.
The study or monitoring objectives are the input for stage
three of the DQO development process.
DQOs are the important starting point for the detailed
design of a data collection effort and are the basis for
specifying the quality assurance and quality control activities
associated with the data collection process. QA Project Plans
are required of all USEPA data collection activities. Such
plans clearly describe what will be done at each stage of data
29
-------
collection (i.e., sample site selection, sample collection,
handling and analysis, and data handling and analysis) and
include instructions or standard operating procedures for each
field and laboratory activity.
During the detailed planning and preparation of technical
guidance for data collectors, DQOs are used as the starting
point for developing explicit, quantitative statements of the
type of errors that will be controlled, the level to which
these errors will be controlled, and the information that will
be collected in order to characterize all the known sources of
error. These quantitative statements are known as data quality
indicators. Data quality indicators are needed in order to
select appropriate methods for sample collection, laboratory
analysis and statistical data analysis. They are also the
basis for 'selecting QA and QC procedures (USEPA Draft, 1984).
In the remainder of this report the general guidance
provided above will be applied to selected aspects of sediment
sampling programs. The cogent relationship among the
objectives for sediment sampling, the DQOs, and the QA/QC plan
should constantly be kept in mind.
In Chapter 1 some possible objectives of sediment sampling
were identified as:
o Determining the extent to which sediments act as
either sources or sinks for water pollutants,
o Determining presence and distribution of selected
pollutants in sediments in both space and time,
o Determining the risk to human health and/or the
environment from sediment contamination by selected
pollutants, and
o Taking measurements for validation of sediment
transport and deposition models.
Bach of these objectives will now be examined identifying
possible actions which might be taken once the objectives have
been achieved.
30
-------
In essence, the mission of the USEPA is to control
environmental pollutants and to abate potential adverse effects
on man and/or the environment. Complying with this mission
requires identifying significant sources of pollutants of
concern, and linking these source emissions via exposure of
important receptors to adverse effects. Thus, to carry out the
intent of, for example, the Clean Water Act, concentrations of
hazardous pollutants in waters should not be allowed to exceed
levels established as being adequately protective of man and
the environment when the intended uses of the waters are taken
into consideration. Identification of the sources of the
pollutant of concern should not only include the present
emissions but also an assessment of likely future emissions.
«, ,
Fc: example, one needs to establish the role of sediments as
sources c*r sinks for selected water pollutants and how that
role may change in time and space, and also the effect of
such physical parameters as water temperature, depth, pH, and
flow rates, suspended solids, bedload, and geological factors
on that role. Biological factors may also be involved in the
degradation or transformation of pollutants into different
substances.
If, for example, significant quantities of the pollutants
of concern become essentially permanently attached to the
sediments and remain biologically unavailable, the sediments
stay constitute a sink for the selected pollutants. Control
needs for these selected pollutants may be reduced by the
amounts which the sediments remove in the sense described
above, provided that no harm from the added load of pollutants
comes to the biota dwelling in the sediments. Underestimates
of the ability of sediments to act as a sink might lead to
source control requirements more stringent than necessary,
whereas overestimates might lead to less stringent control
requirements than necessary.
Bowever, one should use sediments as a sink for
contaminants with caution. When the sediments become
31
-------
contaminated, dredging as a clean up measure is a complicated
proposition. It involves extensive testing of the sediment and
proposed disposal options to determine which one will have the
least environmental impact. With a badly contaminated sediment
one ends up with the problem of what to do with the material
once it has been dredged.
If significant quantities of the selected pollutants are
found to be associated with sediments initially and then
released slowly over relatively long periods of time, the
sediments in essence act as a pollutant source. In this
instance, to keep concentrations of the pollutants below
acceptable levels in downstream waters, it may be necessary to
either over-control industrial, municipal, or non-point
sources, or remove some or all of the polluted sediments by
dredging. Underestimation of the extent to which sediments act
as sources might lead to insufficient controls of other
sources, whereas overestimation might lead to controls more
stringent than necessary and perhaps even to the institution of
expensive dredging operations to a greater degree than
necessary.
The determination of the presence and distribution of
selected pollutants in sediments in both space and time is
necessary to achieve source or sink monitoring objectives.
One possible action which might be taken on the basis of the
mere presence of selected pollutants without regard to whether
the sediments act as a source or as a sink is related to a case
covered under the hazardous wastes regulations (CERCLA or RCRA).
If the selected pollutants are constituents being stored,
treated, or disposed of at a permitted haxardous waste
facility, and there is probable cause that they have originated
from this facility, there nay be grounds for revoking the
permit of the facility. Reporting the pollutants present in
the sediments when they are not there would be a Type I error
and might lead to the revoking of a hazardous waste facility
permit when the facility is not in violation. Failing to
report the pollutants present in the sediments when they are
32
-------
there would be a Type II error and would lead to allowing a
hazardous waste facility to continue operations when it is in
violation of its permit.
Determination of risk to human health and the environment
from contaminated sediments involves several steps. What is
ultimately required are exposure distributions to the most
sensitive population of receptors of concern via all
significant exposure pathways involving sediments. This will
involve concern over possible exposure to water in contact with
the sediments either through ingestion or skin absorption, as
well as concern over possible exposure through ingestion of
food contaminated directly or indirectly through contact with
sediments (crops or domestic animals using water which has been
in contact with the sediments, and/or aquatic foods such as
fish or shellfish contaminated directly or indirectly from the
sediments). It is generally the water in contact with the
sediments which leads ultimately to the exposure of receptors.
Thus, it is important to measure or estimate the extent to
which the sediments act as a source (to contacting waters) for
the pollutant (s) of concern. Knowing the concentration of
pollutants in water originating from contaminated sediments is
not sufficient for estimating exposure. An additional
parameter required is the biological availability of the
pollutant(s) of concern. For example, if pollutants are
not incorporated into the edible parts of seafood, even large
concentrations in the water might not lead to significant human
exposure through ingestion of aquatic food stuffs.
Once desired exposure distributions have been constructed,
comparison to established exposure-response relationships
enables a determination of whether or not the existing risk is
acceptable. Underestimation of the exposures might lead to
accepting an unacceptable risk, whereas overestimation of the
exposures might lead to unnecessary, and possibly costly,
control actions.
The taking of measurements for validation of sediment
transport and deposition models will not normally lead to
33
-------
control actions. Thus, positive or negative errors are
unlikely to lead to corresponding over or under estimates of
control needs. However, errors of unknown direction and size,
if sufficiently large, might seem to validate an erroneous
model or fail to validate an acceptable model. The
consequences of such errors cannot be evaluated without knowing
the purposes for which the model might be used and what actions
might be taken on the basis of conclusions drawn from the
model.
The point to be made is that, prior to undertaking any
sediment sampling program to achieve defined objectives, it is
necessary to establish acceptable levels of precision for end
results. These should be established after due consideration
of the consequences of taking actions which might subsequently
be shown not to be justified on the basis of the available
data.
Once levels of precision have been established, an
experimental protocol should be prepared setting forth what is
to be done for what purpose; and how, when, where and how many
samples will be collected. Also, the protocol should indicate
how the samples will be prepared for analysis and then analyzed
for what substances, and how the resulting data will be
validated, analyzed and interpreted. As part of this protocol,
a complete QA/QC plan must be included covering all aspects of
the experimental program with special attention to sampling
aspects. In the remainder of this report, additional details
will be presented with regard to specific required elements of
the QA/QC plan for various kinds of sediment sampling programs.
GENERAL IDENTIFICATION OF THE OBJECTIVES
Some functional objectives for sediment sampling and
associated QA/QC programs have been identified and discussed.
This material will now be recast for application to problems
related to carrying out the provisions and intent of RCRA and
34
-------
CERCLA. Operational situations in which sediment sampling may
be involved include:
o Preliminary site investigations
o Emergency cleanup operations
o Planned removal operations
o Remedial response operations
o Monitoring
o Research or technology transfer studies
With the possible exception of research or technology transfer
studies, all of the operational situations listed have a
potential for litigation. For this reason, a statistical
experimental design incorporating appropriate QA/QC measures
including "chain-of-custody" procedures should be incorporated
into the sampling program. The total QA/QC plan should require
that the accuracy and comparability of the analytical methods
used, as well as the precision and representativeness of the
sampling, be demonstrated. Generally, the demonstration of
accuracy and comparability will be part of the QA/QC plan for
the appropriate analytical laboratory. Demonstration of the
precision and representativeness of the sampling must be part
of the QA/QC plan incorporated into the sampling protocol.
Precision measures the repeatability of the results obtained
from analyzing the collected sediment samples.
Representativeness of the sample has two components: the
sample taken must reflect what is actually present in the
sediment (this is difficult to quantify) and, the reliability
of the mean and standard deviation as measures of the amount of
a chemical present in a particular area must be established.
Increased sampling intensity, independent sampling, and
sampling audits are examples of techniques that help ensure
that the sample is representative of the condition in the area
under investigation.
The purpose of a preliminary site investigation is to
provide information about a specific site that can be used in
making initial management decisions, and, should further work
35
-------
be necessary, for designing a more detailed and comprehensive
sampling investigation. Since the data collected during the
preliminary study will be used to make important decisions
•bout the site, it is essential that the reliability of the
data be demonstrated through incorporation and implementation
of an adequate QA/QC plan for this investigation. For example,
the preliminary results may indicate that an emergency response
should be initiated. Making an erroneous decision based upon
data of unknown quality concerning such an important matter
could lead to serious consequences.
The purpose of an emergency cleanup operation is to remove
enough of the pollutants as quickly as possible to achieve a
level that is not considered an unacceptable threat to human
health or the environment. The principal role of the QA/QC
plan in this situation is to provide a reliable demonstration
that cleanup operations have been adequate. An emergency
cleanup operation often leads to a requirement for either a
planned removal or a remedial response operation. Thus, any
sediment sampling undertaken during the emergency phase should
have adequate QA/QC measures to ensure that the resulting data
may be used as a foundation for any subsequent investigations.
The purpose of planned removal or remedial response
operations (they differ principally with regard to time scale)
is to provide a more permanent solution to the problem. These
operations may involve extensive sampling and data analysis
programs. Adequate QA/QC measures are essential since
litigation to recover the costs of the operations is a likely
sequel. Consequently, all data collected may well undergo
close scrutiny in court.
Monitoring, or sequential measurements over time, may take
place before, during, or after any of the operational
situations listed above. Whatever trends are measured must be
demonstrated to be reliable in order to serve as a basis for
making decisions that hold up to challenges.
36
-------
The purposes of research or technology transfer studies
vary widely. In any event, the incorporation of adequate QA/QC
plans into these studies is mandatory in order for the results
of the studies to withstand the normal peer review processes
required for publication and/or application of the findings.
In summmary, an adequate QA/QC plan should be part of any
sediment sampling program relevant to any of the operational
situations listed. The only question remaining pertains to the
definition of the word "adequate." That question will be
addressed in a subsequent section of this chapter.
OBJECTIVES FOR BACKGROUND MONITORING
v,,
^Generally the design of sediment monitoring programs
req-.res that the levels of defined hazardous or potentially
hazardous substances and their spatial and temporal trends be
measured for some specific purpose. Often it is critical not
only to quantify levels and trends but also to link the
existing levels to sources. This is necessary to enable
adequate control actions to be taken whenever a situation that
is hazardous to human health, welfare, or the environment is
identified. Often the situation is complicated by the fact
that multiple sources contribute to the measured levels.
The situation is further complicated by the presence of
pollutants of recent origin mixed with pollutants of past
origin. This mixing becomes especially important when the
investigator attempts to trace the migration from source to
receptor and also in predicting what future levels are likely
to be after various proposed control measures are implemented.
Identification of spatial and temporal trends along with
linkage of observed measurements to sources requires that
adequate background or reference or control samples be taken.
In the absence of such background samples, interpretation
of the resulting data may become extremely difficult, if not
impossible. The burden of proof that background samples are
37
-------
not necessary for a particular sediment monitoring study rests
with the principal investigator. In the absence of such proof,
a prudent investigator will ensure that the collection of
adequate background samples is included in the monitoring study
design. Furthermore, some EPA regulations concerning
regulatory monitoring (U. S. Code of Federal Regulations, 1983)
specifically require background sampling.
Since measured levels in presumably higher concentration
areas will be compared to background levels, QA/QC procedures
are just as critical for the background measurements as they
are for the study area measurements. Thus, for background
sampling, a QA/QC procedural umbrella must cover the selection
of appropriate geographical areas, the selection of sampling
sites within the geographical areas, sampling, sample storage
and/or preparation, sample analysis, data reduction, and
interpretation of study results.
Under most circumstances, background data will not be
available for a given monitoring location. These data must be
acquired either before or during the exploratory or preliminary
investigation phase. The intensity of the background sampling
that is undertaken depends upon the pollutants being measured,
the sediment characteristics and variability, the levels of
pollutant likely to be found in the study area and the purpose
of the study.
SPECIFIC OBJECTIVES FOR MONITORING IN SUPPORT OF CERCLA
The principal sampling media now being measured to carry
out the provisions and intent of CERCLA, and RCRA as well, are
• oil and groundwater. What, then, is the proper role for
sediment sampling in support of CERCLA? Hazardous constituents
from a hazardous waste facility may enter sediments through
transport of the constituents from the waste site to sediment,
via either surface water or groundwater flow into receiving
bodies of water. Air transport followed by rainout or washout
38
-------
will generally be less important than the other two transport
routes. What information can be gained, then, from sediment
measurements which cannot be gained from soil, air, surface
water, or groundwater measurements?
Suppose a situation exists in which hazardous waste
constituents have been leaving a site for a relatively long
period of time and an adjacent body of water has built up a
considerable amount of selected constituents in its sediments.
Further, suppose that the sediments now constitute a source of
the hazardous constituents. At this time, removal of the
hazardous wastes from their original disposal site may still
leave an unsolved significant problem in the form of the
contaminated sediments. Human foods, contaminated directly or
indirectly through contact with sediments, may be unfit for
human consumption. Furthermore, as the hazardous constituents
move through different trophic levels, substantial
biomagnification of contaminants may take place, thereby
increasing the risk to humans consuming foods from higher
trophic levels. Thus, it is conceivable that situations may
exist in which concentrations of hazardous constituents in
sediments may represent a major risk to human health or the
environment. To identify such situations, data from sediment
sampling is an important link in the chain of required
evidence.
The steps outlined below are designed to provide a
sediment monitoring effort with adequate sample precision and
representativeness (USEPA, FR44:233, 1979 and Bauer, 1971).
1. Identify the objectives of the study.
2. Determine the components of variance that should be
built into the statistical design.
3. Choose the allowable probabilities for Type I and Type
II errors and the difference in means considered to be
significant. (These choices together with an estimate of the
coefficient of variation are needed to determine the number of
samples required in each stratified region.)
39
-------
4. Obtain sampling data from other studies with similar
characteristics to the one of interest. (Estimates of
coefficients of variation are of particular importance. )
5. Calculate the mean and note the range of each set of
duplicates (co-located independent samples).
6. Group the sets of duplicates according to
concentration ranges and by the types of samples believed to be
similar.
7. Calculate the critical difference Rc (number not to be
exceeded to maintain adequate QA/QC) from the formula
3.27 C
n
n
Z
where C « concentration, n * number of duplicate analyses, Ri *
range « Xi ")(Xi+i ,
and Xi » mean * (x± + X£+i)/2.
8. Using results from previous studies, develop a table
of Rc values for various concentrations that span the range of
concentrations of interest. (These data are used to accept or
reject sets of duplicate samples. )
9. Use the preliminary RC table to accept or reject sets
of duplicates. When approximately 15 pairs (USEPA, 1979) of
results from the present study are available, a new table of Rc
values should be constructed based upon the data that have been
accepted .
10. Use data collected during the preliminary or
exploratory site investigation and any emergency response
activity as the data base upon which later studies are
evaluated and/or designed.
Suggestions for additional elements of a more complete
QA/QC plan are provided in subsequent chapters.
The specific goals for each type of study will determine
the allowable probabilities of Type I and Type XI errors and
the minimum relative difference between sampled population mean
and either background mean, or designated action level that is
40
-------
considered important to detect. Suggested guidelines are given
below for the operational situations listed previously.
PRELIMINARY SITE INVESTIGATION
The preliminary or exploratory investigation is the
foundation upon which other studies in hazardous waste site
assessments should be based. As part of this study, it is
essential to determine whether or not sediments are sample
media of importance to the total assessment. The total
assessment must draw conclusions with regard to whether or not
there is imminent and substantial danger to human health
requiring emergency action and whether there is an unacceptable
long term risk to man or the environment. If sediments are
determined to be unimportant in the preliminary study, it is
likely that no further attention will be directed to them. In
view of this, a Type II error is considered to be of greater
importance than a Type I error. Presented below are suggested
guidelines for DQOs that may be used initially.
Confidence Level Power Relative Increase
(1 - a ) (1 - 8 ) over Background
to be Detectable
with a Probability
(1-B)
70-80% 90-95% 10-20%
If resources limit the number of samples that can be taken, the
investigator should determine, for the number of samples that
can be collected, value-judgment based optimum values for
confidence level, power, and detectable relative difference.
If these values are deemed adequate, the study may proceed.
41
-------
Using five percent duplicate samples may provide adequate
QA/QC for measuring variance between samples (Plumb, 1981).
However, there should be a minimum of two sets of duplicates in
each strata sampled. As data become available, these
assumptions should be checked. This is usually accomplished by
taking and analyzing more duplicates initially, and then
checking to determine the minimum number required for the sites
being sampled and the pollutants being measured.
EMERGENCY CLEANUP
Emergency sampling is designed to identify those areas
in,, which sediments are contaminated to such a degree as to
threaten imminent and substantial endangerment to human health.
The threat may be due to the sediments acting as a source of
hazardous constituents to drinking water or to human foods.
The emergency action in either event is more apt to be
switching to bottled water for drinking and/or taking certain
locally produced human foods off the market than it is to be a
dredging program to remove the contaminated sediments.
Dredging may well be implemented at a later date as part of a
planned removal or a remedial response operation. Of course,
any long term solution to the problem would also have to
address the removal of the primary source of hazardous
substances to the sediments.
For an emergency response operation involving sediments, a
Type II error is considered of greater importance than a Type I
error. Presented below are suggested guidelines for DQOs that
may be used for emergency response operations.
42
-------
Confidence Level Power Relative Increase
(1 - a) (1-B> from Background or
an Action Level to
be Detectable with
Probability (1-B)
80-90% 90 - 95% 10 - 20%
PLANNED REMOVAL AND REMEDIAL RESPONSE STUDIES
These studies are usually continuations of those initiated
during emergency cleanup studies. They should be designed to
provide specific information needed to resolve control option
itsjes. The areas to be surveyed should be stratified and
sarpled according to a design that can be used to determine
spatial variability. A suitable statistical design should be
formulated so that components of variance for the study
situation may be identified and evaluated. Appropriate QA/QC
procedures must be formulated and implemented.
If the sampling during exploratory or emergency response
investigations has been done properly, there will be a sound
basis for determining the sample size and sampling site
distributions. The design will have to incorporate information
on the vertical distribution as well as the horizontal
distributions. Measurements of concentration trends with time
may be of critical importance particularly if sediment
concentrations are changing appreciably with time. For
example, sediments may at least partially cleanse themselves
once the primary source of contamination is removed. This
cleansing process, or reduction in concentration of
contaminants in sediments, may be due to a combination of
biotic degradation of the contaminants together with the
addition of uncontaminated sediments.
For a planned removal or a remedial response
operation involving sediments, it is considered that a Type I
43
-------
and a Type II error are of about equal significance.
Furthermore, an attempt at cost recovery which might lead to
mitigation is a likely successor to these studies.
Accordingly, it is important to achieve the highest order of
precision feasible. Presented below are suggested guidelines
for DQOs that may be used for planned removal and remedial
response studies.
Confidence Level Power Relative Increase
(1 - a) (1 - 6 ) from Background or
an Action Level to
Be Detectable with
Probability (1-6)
90-95% 90-95% 10-20%
MONITORING OR RESEARCH STUDIES
The guidelines for these studies for confidence levels,
power, and detectable relative differences should be set on the
basis of the objectives of the studies. As actions which may
be taken on the basis of resulting data become more and more
significant and costly, greater effort should be placed on
achieving an increased level of reliability for the data.
Publication of the results in a peer-reviewed journal will also
usually'require some demonstration that an adequate QA/QC plan
has been incorporated into the experimental protocol.
44
-------
CHAPTER 4
STATISTICAL CONSIDERATIONS
INTRODUCTION
This chapter reviews the role of statistics in the
sediment pollution monitoring process. Statistics is a
science of data collection and analysis to efficiently obtain
information concerning questions of interest. Without
statistics there would be no basis for comparison of sampling
procedures of equal cost. There are numerous texts and
journals dealing with statistics. Some references that relate
to the statistics of sediment sampling are given in this
chapter. The techniques presented in these references will not
be discussed in detail. The user is encouraged to utilize the
referenced materials if additional information is required.
However, in the actual planning of a sediment campling design
the reader is advised to consult a professional statistician.
DISTRIBUTION OF SEDIMENT SAMPLING DATA
Statistical sampling plans are based on assumptions
concerning the probability distributions of the measurements to
be made. These assumptions should be consistent with results
from past surveys taken under similar conditions. The
variability in sample data is a function of the variable being
measured, the analytical procedure, and the sampling procedure.
If the distribution of a measurement is normal, it is symmetric
about its expected value (center of gravity of the probability
distribution) and its variability is uniquely determined by its
variance (variance is the moment of inertia of the probability
45
-------
distribution about its mean when probability is treated as
mass). The symmetry makes the expected value a reasonable
measure of location, whereas in non-symmetric distributions
other measures may be preferred (e.g., the median). Also, the
statistician has means of dividing variance into components
representing various sources of variation. With most other
probability distributions, the variability is only partially
described by the variance. Hence, these properties of
symmetry, and variance representing variation, are two of the
prime reasons for transforming variables so that the new
distributions are approximately normal. Procedures for such
transformations are given in Box and Cox (1964) and in Hoaglin
et al. (1983). A discussion of the importance of the normality
assumption and some possible transformations appears in Scheffe
(1959, Chapter 10). In what follows, we shall assume that the
data have been transformed to near normality.
In the paragraph above, only variables with quantitative
measurements were considered. If the variable of interest has
a count measurement, such as radioactivity or presence or
absence of a pollutant, other statistical methods are required.
These methods are usually denoted qualitative or discrete
statistical methods. Bishop et al. (1975) is a good reference
to these procedures. The methods of this chapter should not be
applied to count data.
The environmental scientist can obtain information on the
distribution of a variable by conducting an exploratory or
pilot study. The exploratory studies conducted during the
initial phases of an investigation can provide an indication of
the site specific probability distribution pattern and the
transformation to normality that may be needed. McKay and
Pater son (1984) discuss the use of the normal, log normal and
Meibull distribtions in environmental studies. The
environmental scientist is interested in finding the location
and amounts of pollutants that emanated from a source;
46
-------
therefore the pilot study should provide information on both
contaminated and background sediment areas.
Additional information about the distributions of
measurements of pollutants may be obtained from EPA's Regional
Offices and Laboratories and EPA's National Enforcement
Investigation Center in Denver, Colorado.
STATISTICAL DESIGNS
The design and method of analysis for the sampling study
must be determined before the sampling is undertaken. Improper
design or analysis may invalidate the resulting conclusions, or
prevent valid conclusions from being made. Care must be taken
not to allow time of sampling to be confounded with an effect
being estimated. Also it is very important that the individual
samples and subsamples be taken in such a way that the
measurements are comparable. Basic ideas of sampling design
may be found in Hansen et al. (1953) and Gy (1982). Two of the
simpler designs are the simple random sampling design and the
stratified random design. In the simple random sampling
design, the n sample points are randomly selected in such a way
that all combinations of n points in the population have the
same chance of being chosen. While the simple random design
allows easy methods for the analysis of data, it is inefficient
in the use of resources and is infrequently used in practice.
The stratified random design is one in which the area under
study is subdivided into smaller areas (strata) that have the
potential of being markedly different in pollutant
concentrations and then simple random sampling is done within
each stratum. This procedure ensures that no large sub-area is
without sample points and thereby helps reduce sampling
variance when there are substantial differences in
concentrations between strata. Methods for optimizing the
choice of the number of strata and number of points within
strata are given in the text by Ransen et al.(1953).
47
-------
There are two basic approaches to the planning and
analysis of sediment sampling. One is the traditional sampling
model approach, found in Bansen et al.(1953), which uses
randomization in the selection of sample points, as a
probability basis for statistical inference, and an
analysi s-of -variance model approach to inference. The second
is a "geostatistical" model approach using the idea that an
underlying random process created the spatial distribution of
the variable. The geostatistical approach involves the
estimation of spatial structure of random functions and kriging
to estimate isopleths of variable values. An introduction to
these procedures may be found in Journel and Huijbregts (1978).
The methods given in this chapter relate to the more
traditional analysi s-of -variance sampling model.
Type I and Type II Errors
The environmental manager may wish to make an informed
decision through a statistical test of hypothesis based on the
sediment samples. For example, he may need to decide whether
the study area is contaminated or not. The hypothesis to be
tested is the 'null* hypothesis of no contamination, which
might be expressed as
R: MS * PB (°r US £
where \i stands for the mean of a population and the subscripts
S and B stand for the study and background populations
respectively. If the test rejects the hypothesis above, then
the alternative hypothesis of study-area contamination
A: vs > ME
is accepted. This test is a one-sided test in that A is uS>yB*
In a two-sided test, the two hypotheses are Htyg^UB' and
48
-------
A: g« g. For example, the two-sided test may be of interest in
determining whether pollutants have caused a change in pH.
A test of hypothesis is basically a decision rule
specifying a test statistic (i.e., a function of the sample
data) and a set of possible values of that test statistic,
called the critical region of the test, such that if the value
of the test statistic for the obtained sample data is in the
critical region, the null hypothesis is rejected and the
alternative hypothesis is accepted. If the value of the test
statistic does not fall in the critical region, the alternative
hypothesis is not accepted. Two types of error are possible.
The acceptance of the alternative hypothesis when the null
hypothesis is true (false positive) is said to be a Type I
e- -or. Failure to accept the alternative hypothesis when it is
tr .* (false negative) is a Type II error. The two types of
error may be equally well defined in terms of acceptance and
rejection of the null hypothesis. Then one would say that if
the value of the test statistic is in the critical region, the
conclusion is to reject the null hypothesis; otherwise one
accepts the null hypothesis. Similarly, one may call the
complement of the critical region, the acceptance region.
Figure 4 illustrates a two-sided test situation where the
acceptance region is the interval below the center of the
density curve and the critical region consists of the two
intervals below shaded tails of the density curve. The maximum
probability allowed for a Type I error in testing a hypothesis
is called the significance level of the test. The significance
level of a test is commonly denoted by the Greek letter alpha
(a), Typical values used for significance levels are 0.001,
0.01, 0.05 and 0.10. The value chosen depends on the
consequences of making a Type I error and is not limited to the
typical values. The diagram below illustrates the
relationships described for Type I and Type II errors.
49
-------
TRUTH
Accept H
DECISION:
Accept A
H
Correct
Type I
Error
Type II
Error
Correct
The probability of a Type II error (i.e., the
probability of accepting the null hypothesis when it is false)
is usually denoted by the Greek letter beta (6) and is
typically a function of a , sample size, and the size of the
deviation from the null hypothesis. The probability that the
alternative hypothesis will be accepted when it is true (i.e.,
the probability that the test statistic will take on a value in
the critical region when the alternative hypothesis is true) is
called the power of the test and may be denoted by (1-6).
Typically, the experimenter will specify the smallest deviation
from the null hypothesis that he considers to be
scientifically, economically, or environmentally important to
detect and then specify the power of the test that he wants for
that specific alternative. Obviously he wants the test to have
high power for the scientifically important alternative and low
significance level. However, it is evident that if one
increases power by increasing the size of the critical region,
one is also increasing significance level. One way to increase
power, without increasing significance level is to increase the
amount of information; that is, increase the sample size.
Figure 4b shows the probability density curve for a test
statistic under the null hypothesis,
B:
30.0
SO
-------
I
ACCEPTANCE REGION
202 *> • SO.O
Rf. Aa Acceptance rrpon for H po ** 30.0.
M-IO.O
T)-pc H or f error.
20.2 *. - 30.0
39J
51
-------
The shaded portion represents the probability of a Type I error
(a). In Figure 4b the left curve represents the probability
density function of the test statistic when y « 10. The shaded
area in Figure 4b represents the probability (6) of a Type II
error in this situation (Juran et al., 1979).
Number and Location of Samples
There are three basic procedures for increasing the
precision of statistical estimators and the power of
statistical tests. They are (i) use more efficient statistical
estimators and tests, (ii) improve the sampling design, and
(iii) increase the sample sizes. Table 11 in Chapter 6 gives
information on sample sizes to use when employing t-tests of
means. Discussion concerning the origin and use of these
tables is also given in Chapter 6. Additional tables for the
determination of sample sizes can be found in Beyer (1968).
The use of t-tests requires some form of random selection
process so that the standard deviation of an observation may be
estimated.
Stratification is a sampling procedure for improving
precision of estimates. This technique makes use of scientific
knowledge that the measurements may be quite different in
different identifiable segments of the area being sampled. A
typical stratification criterion used in soil science is the
soil type. Another criterion that might be useful in sediment
sampling is distance from point sources of pollutants.
Role of Quality Assurance in Experimental Design
The Quality Assurance Officer should be intimately
involved in the review of the experimental or sampling design
proposed by the investigator. He should require that the
information obtained provide measures of the components of
52
-------
variance that are identified in the field. An additional
quality check that should be undertaken as part of the QA
program is the review of the design by qualified sediment
scientists and other peers that are in a position to provide
the necessary oversight of the sampling effort.
Broms (1980) makes the following statement; "There should
be a balance between the soil investigation method, the quality
of the soil samples, and the care and skill spent on the
preparation and the testing of the samples. There is no point
in spending time and money on careful sample preparation and on
testing if the quality of the samples is poor.* This statement
is equally applicable to sediment sampling. The QA program
must address the total flow of information from the design to
tr~ reporting of the results. The sampling design is the
foundation of the whole study, therefore, it should be given
maximum support if the purposes of the sampling effort are to
be met.
Components of Variance
The components of variance analysis, (see Scheffe,
Chapters 7 and 8) provides estimates of the portion of the
total variation coming from each of the sources of variation in
the measurements. Basic assumptions of this procedure are that
the measurements are normal in distribution, independent, and
each source has constant variance. An excellent example of the
use of this technique is provided in a report by the Electric
Power Research Institute (Eynon and Switzer, 1983). An example
presented in Table 3 gives the components of variance for
hypothetical sample data from a stratified random design with
four strata, three random samples per stratum, two subsamples
per sample, and one analysis per subsample. (The stratum
effects are assumed fixed here, so this is really a mixed-model
53
-------
TABLE 3. ANALYSIS OF VARIANCE OF A NESTED SEDIMENT SAMPLING DESIGN.
Stratum Sample Subsample Xjiu X;; X;.. X...
(i) (j) (k) J J
1 1
2
3
2 1
2
3
3 1
2
* , 3
4 1
2
3
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
3.17
2.64
1.79
3.00
2.20
1.95
1.10
2.94
2.77
1.95
2.71
3.00
4.33
4.50
4.25
4.53
3.87
4.79
5.03
4.65
3.95
3.76
4.79
4.63
5.61
4.79
4.15 14.75
4.04
4.72
5.71 14.47
8.83
8.78
8.66 26.27
9.68
7.71
9.42 26.81 82.30
a-4
b-3
n-2
I. C • (X...)2/(«bn) - (82.30)2/24 • 282.2204
II. Total: Zxfjk - C • (3.172+...+4.632) - 282.2204 - 29.8656
III. Strata: EX£2v/bn - C - (14.752+...+26.812)/6 - C • 23.7517
IV. Samples: ZXfj/n - C - (5.812+...+9.422)/2 - C • 26.3109
V. Samples in Strata: IV - III • 2.5592
VI. Analysis of Sample: II - IV - 3.5547
ANOVA TABLE
Source of
Variation
Strata
Samples/Strata
Analysis/Samples/
Strata (error)
Total
Degrees of
Freedom
3
8
12
23
Sum of
Squares
23.7517
2.5592
3.557
29 .8656
Mean
Square
7.9172
0.3199
0.2962
Expected
Mean Square
VA * «VS *
*A * »VS
VA
bnM/3
s2 " 0.2962 estimates V^ or variance due to subsampling and analysis
a2 • (0.3199 - 0.2962)/2 - 0.0118 estimates Vs
where Vg is the variance due to sampling within strata.
M • Sum of squared deviations of stratum means about grand mean.
54
-------
analysis (i.e., some random and some fixed effects), but it
does provide estimates of components of variance from within
stratum sampling and combined subsampling and analytical
errors). The results in Table 3 would indicate that the
experimenter should either have made a greater effort to reduce
subsampling and analytical errors or taken more subsamples
since the error variance is much larger than the variance
between samples within strata.
Compositing of Samples
A technique that is often employed to reduce sample
handling and analytical costs is the compositing of samples.
Combining the samples from several sampling locations reduces
the costs for analysis. This procedure is used extensively by
agricultural workers to determine fertilizer requirements for
farm fields. Peterson and Calvin (1965) make the following
statement about the technique:
"It should be pointed out that the composite samples
provide only an estimate of the mean of the
population from which the samples forming the
composite are drawn. Ho estimate of the variance of
the mean, and hence, the precision with which the
mean is estimated can be obtained from a composite of
samples. It is not sufficient to analyze two or more
subsamples from the same composite to obtain an
estimate of the variation within the population.
Such a procedure would permit the estimation of
variation among subsamples within the composite, but
not the variation among samples in the field.
Similarly, if composites are formed from camples
within different parts of a population, the
variability among the parts, but not the variability
within the parts, can be estimated. If an estimate
of the variability among sampling units within the
55
-------
population is required, two or more samples taken at
random within the population must be analyzed
separately."
Youden and Steiner (1975) caution against the use of the
composite sample for much the same reasons as those outlined
above. Since a prime purpose of QA/QC is to assess and assure
the accuracy (i.e., lack of bias and level of precision) of the
data and of estimates obtained from the data, it is essential
that estimates of the precision be made from the data.
Therefore, the compositing of samples cannot, in general, be
recommended.
Some work on determining the precision of estimates of
the mean from composite samples has been published. Such
estimates of precision usually require some strong assumptions
about variance components and/or the stochastic nature of the
composited samples (see Duncan (1962) and Elder, et al.
(1980)).
Split Samples, Spiked Samples and Blanks
Split samples, spiked samples and blanks are used to
provide a measure of the internal consistency of the samples
and to provide an estimate of the components of variance and
the bias in the analytical process. To obtain an unbiased
measure of the internal consistency of samples and their
analyses, the individual samples should be labeled with a code
number in such a way that the chemist (and preferably also the
laboratory) do not know the relationship between the samples
that he is analyzing. This reduces the chances of conscious or
unconscious efforts to improve the apparent consistency of the
analyses.
Samples can be split to:
o Provide samples for both parties in a litigation or
potential litigation situation.
56
-------
o Provide a measure of the within sample variability
(this is needed in order to determine the
influence of other factors that may be confounded
with sample splitting.)
o Provide materials for spiking in order to test
recovery.
o Provide a measure of the sample bank and extraction
error.
The location of the sample splitting determines the
component of variation that is measured by the split. A split
made in the sample bank measures error introduced from that
level onward. A split made in the field includes errors
associated with field handling. A split or series of
subsamples made in the laboratory for extraction purposes
measures the extraction error and subsequent analytical errors.
Spiked samples are prepared by adding a known amount of
reference chemical to one of a pair of split samples. The
results of the analysis of a split compared with the non-spike
member of the split measures the recovery of the analytical
process and also provides a measure of the analytical bias.
Spike samples are difficult to prepare with sediment
material itself. Frequently the spike solution is added to the
extract of the sediments. This avoids the problem of mixing,
etc. but does not provide a measure of the interaction of the
chemicals in the sediments with the spike, nor does it provide
an evaluation of the extraction efficiency.
Blanks provide a measure of various cross-contamination
sources, background levels in the reagents, decontamination
efficiency and any other potential error that can be introduced
from sources other than the sample. For example, a trip blank
measures any contamination that may be introduced into the
•ample during shipment of containers from the laboratory to the
field and back to the laboratory. A field blank measures input
from contaminated dust or air into the sample. A
decontamination blank measures any chemical that may have been
57
-------
in the sample container or on the tools after decontamination
is completed.
The number of QA/QC samples have been selected by a rule
of thumb that one out of every twenty samples is to be assigned
to each of the categories of samples. This ratio has been used
successfully in several major USEPA studies (USEPA, 1982, 1984).
Table 4 presents the breakdown of QA/QC samples used in these
previously conducted monitoring studies.
DATA ANALYSIS
The topics that follow are designed to provide insight
into the use of statistical techniques for evaluating the data
obtained during an investigation. They are not by any means
exhaustive, but are chosen to provide the basis for designing
the quality assurance portions of a sampling effort and to
provide the basis for obtaining the most benefit from the data
acquired.
Bias
The variation seen in analytical data can be composed of
variation within the sample itself, variation introduced in
sample collection or preparation and variation in the analysis of
the samples. The variation can further be divided into sample
variation and bias. Bias identifies a systematic component of
the error that causes the mean value of the sample data to be
either higher or lower than the true mean value of the samples.
Bias must be due to a fault in the sampling design, sampling
procedure, analytical procedure or statistical sample. An
example of a bias would be the error in analytical results
introduced by an instrument being out of calibration during a
portion of the analysis. Laboratories usually introduce
reference samples into their sample load in order to detect these
changes. Bias in sediment sampling is difficult to detect. The
58
-------
TARE 4. QA/QC PROCEDURES FOR SEDIMENT SAMPLES
Procedure
1. Field Blanka
Coamente
One for each aaapling team
par day. A aample container
filled with dietilled,
da-ionized water, exposed
during aaapling than analyzed
to detect accidental or
incidental contamination.
2. Sample Bank Blanka
The blank, about one for each
40 aamplea, paaaed through
the aaaple preparation
apparatue, after cleaning, to
check for raaidual
contaaination.
3. Decontamination Blanka
4. Reagent Blank
A blank, about 1 for each 20
aaaple a, paeaed over the
aaapling apparatua after
cleaning, to check for
raaidual contamination.
One for each 20 aaaplaa to
check reagent contaaination
level.
5. Calibration Check Standard
6. Spiked Extract
7. Spiked Saaple
One for each 20
check inetruMnt
lea to
libration.
One for each 20 aaaplae to
check for extract matrix
effecte on recovery of known
added analyta.
One for each 20 aaaplaa. A
aeparate aliquot of the aoil
aaaple aplked with NBS Lead
Nitrate to check for aoil and
extract Matrix affacta on
recovery.
59
-------
TABLE 4. CONTINUED
Procedure
8. Total Recoverable
Coemnte
One for aach 40 eeaplee, a
eacond aliquot of the aaaple
la di gee ted by a aore
vigoroua method to check the
efficacy of the protocol
•athod.
9. Laboratory Control Standard
10. Re-extraction
11. Split Extract
12. Triplicate Saaple (Splita)
•13. Duplicate Saaple
One for each 20 aaaplee. A
eaaple of NBS River Sediwnt
carried through the
analytical procedure to
determine overall aethod
biae.
One for eech 20 aaaples. A
re-extraction of the reaidue
from the firat extraction to
determine extraction
efficiency.
One for each 20 aaaplae to
check injection end
inetrtawnt reproducibility.
One for eech 20 aaaplee. The
prepared eaaple ia eplit into
three portione to provide
blind duplicetee for the
analytical laboratory and a
third replicate for the
referee laboratory to
deteimine interlab pracieion.
One for each 20 aaaplaa to
determine total randoa error.
60
-------
presence of bias can be proven by use of one of the techiques
described below. On the other hand it is difficult to prove that
bias is not present because the absence of bias may be the result
of the inability to measure it rather than its actual absence.
Standard Additions-- It is necessary to conduct special
experiments in order to detect bias in the sampling effort. The
major technique used is that of adding known amounts of standard
solutions to the samples: it is recommended that this be done in
the field or in a field laboratory. The main problem encountered
is that mixing sediments to obtain homogeneity is difficult in a
laboratory much less in the field. Several known quantities of
the standard are added to samples taken in the field. The
results should follow the equation for a straight line:
where x is the increase in concentration and y is the value
obtained by the laboratory. Bias is indicated if the data do
not follow the straight line equation, or if a < 0. If the
units of x and y are the same, the value of b, should be unity;
and significant deviations from unity indicate a proportional
bias (Allmaras, 1965).
Internal Consistency-- If several samples of sediments of
different size are analyzed for a constituent, the results
should fit a linear equation of the form:
y - a +
where Z is the quantity of sample analyzed. The amount of
chemical detected should be directly related to the quantity of
the sample analyzed. The plot of the (y,Z) data should be
essentially linear; if not, bias is indicated. The intercept,
a, should be within sampling error of zero and the slope b
61
-------
should represent the concentration of the chemical in the
sediments. A linear graph in which the intercept is definitely
nonzero would indicate an additive bias in the analytical
procedure.
Confidence and Prediction Limits
Typically one wishes to estimate the concentration of
measured pollutants in the sediments and to indicate the
precision of these estimates. To indicate precision of an
estimate one may provide the standard error or a confidence
interval for the expected value of the concentration. Where
statistical designs have been used in the sampling, the
analysis of variance (ANOVA) provides needed information for
calculating standard errors and confidence intervals.
The confidence interval is bounded by confidence limits
(CD. The confidence limits are "the bounds of uncertainty
about the average caused by the variability of the experiment"
(Bauer, 1971). The limits for the mean are defined by the
following equation.
CL * x ± ts//m
where x » sample mean, s » sample standard deviation, m «
number of samples and t » Student's t value at the desired
level of confidence and with degrees of freedom associated with
s in the ANOVA (see Appendix A, for values of t).
Consider again the example of Table 3. If all the strata
represent equal area subdivisions of the study area, the
logical estimate of the expected concentration for the study
area is just the sample mean of the 24 measurements,
x - 82.3/24 - 3.43
which could also be obtained by first finding the average of
each pair of subsamples and then averaging these 12 sample
62
-------
values. The variance of the average over a pair of subsamples
is
When one averages over the 12 samples, a new source of
variation enters in; namely, the samples -wit hi n-st rat a
(samples/strata) variance. Therefore, the variance of the
•ample mean is
[VS + VA/2]/12 - (VA + 2Vs)/24
The quantity,
VA + 2VS
is estimated by the mean square for samples/strata in the ANOVA
table with 8 degrees of freedom. Therefore our estimate of the
standard error of the mean, s/Ym, (s « /0.3199 « 0.5656 and m «
(2)(12) « 24) is
0. S656//24 - 0.115
The table in Appendix A gives t - 2.306 for a two-sided
confidence interval with 95% confidence based on an estimate of
s with 8 degrees of freedom. Hence the 95% confidence interval
in this case is bounded by the confidence limits.
CL • 3.43 + (2.306X0. 115) - 3.16, 3.70.
Prediction limits (PL) (see Hahn, 1969; and Guttman et al.,
1962) are similar to confidence limits in appearance but are
used to identify an interval into which a randomly chosen
63
-------
future sample value from stratum i should fall. The defining
equation for these limits is:
PL «
where xi is the sample mean for stratum i. Hence, one can say
for the above example that if one more sample were randomly
taken from the stratum 1, one would be 95% confident that the
means of the analyses on the two subsamples would give a value
between the prediction limits,
PL « 2.46 + <2.306)<0.5656)/«l/2)+(l/6))
- 2.46 + 1.06
- 1.40, 3.52
Outliers
A problem that is particularly prevalent in data obtained
from field samples is that of outliers (i.e., observations that
are discordant with the rest of the data set). The basic
question is whether it is reasonable to expect such a
discordant observation in the sample; if not, the measurement
is considered an outlier. The cause of the outlier may be an
error of procedure in sampling, subsampling, chemical analysis,
or the transcribing of data; or it may be due to an anomaly
that would indicate that a change is required in the assumed
model for the process (e.g., vegetation that takes up a heavy
metal being measured is not present at one of the sample points
and this causes a much higher measurement at that point than at
the others).
The discordance of an observation depends on the assumed
probability distribution for the variable being measured. A
measurement that is large relative to the other neasurements
may appear discordant to an observer who assumes a normal
distribution for the variable, but not discordant to another
64
-------
observer who assumes that the probability distribution of the
variable is highly skewed to the right. Hence, tests of
hypotheses concerning the presence or absence of outliers are
based on assumptions concerning the underlying probability
distribution. Many tests have been devised for normal, gamma,
and Poisson distributions. A book by Barnett and Lewis (1978)
lists many of these outlier tests and also gives tables of
critical values for the tests.
In environmental monitoring, extremely large measurements
of pollutant concentrations are particularly disturbing. A
test that is good for checking a discordant measurement on the
right of a data set (i.e., the largest measurement) having an
underlying normal probability distribution uses the test
statistic
*
W -
-------
Testing of Hypotheses
The most commonly used test of hypotheses for comparison
between two population means or for comparison of a population
mean with some standard value is a t-test. To compare two
means, using data from simple random samples of the two
populations, the following test statistic is employed:
- x2)/sp/[l/n1)-Kl/n2)l
where, the pooled standard deviation,
and X£, Si, and nj, are the sample mean, sample variance, and
sample size for the ith (i»l,2) sample. In this two-sample
t-test, one is either testing the null hypothesis, R: ^1^2'
against the two-tailed alternative that two means are
different, hs]*L+V2' or against a one-tailed alternative, A:
Pi>u2- For tne two-tailed case, one accepts the alternative
hypothesis only if | ts | >. t, where t is the value found in
the table of Appendix A and listed in the 1-ct column, for
two-tailed tests, and the (ni+n2-2) (df) row. For the
one-tailed alternative, one accepts the alternative hypothesis
only if ts > t, where t is now obtained from the same row of
the table, but from the l-a column for one-tailed tests. Mote,
in the table that, 'confidence level" is one minus significance
level and reflects a correspondence between confidence
intervals and tests for means based on the Student's
t -distribution.
The one-sample t-test which compares a population mean
with a standard value may arise in determining whether the mean
concentration of a pollutant in a study area exceeds a
66
-------
specified action level. The test statistic for this test is
tc » (x - L)(/n)/s
where L is the standard value (action level) and s is the
•ample standard deviation. One-and two-tailed tests are
performed in the same way as described above for the two-sample
test, except that the numbered degrees of freedom is now (n-1).
In dealing with action levels one would be interested in the
one-tailed test.
Example:
A preliminary study is done in an area suspected of being
contaminated with polychlorinated biphenyls (PCB's). Sixteen
sediment samples were collected from both the study area and
from a background area through the use of simple random
sampling. Table 5 lists the data and their summary statistics.
TABLE 5. PCB STUDY TO DETERMINE CONTAMINATION OF AN AREA
(HYPOTHETICAL DATA)
Background Area (ppb)
Study Area (ppb)
XB -
*s -
•cv -
35.8
45.5
35.5
32.0
50.0
39.0
37.0
47.0
40.23 ppb
52.61 ppb
Coefficient
38.5
36.0
40.5
35.5
45.5
37.0
36.0
53.0
8B2 - 36.8825
ss2 • 60.2598
of variation in %
47.0
62.0
47.0
59.5
40.0
57.5
48.5
53.0
nB m 16
nS - 16
50.0
49.6
53.5
68.0
60.0
45.0
42.5
58.7
CVB*-
CVS -
15.1%
14.8%
The test statistic is calculated as follows:
sp - /[1506.8825 + 60.2598)/(16 + 16 - 2» - 6.97
ts - [52.61 - 40.231/16.97/(2/16) - 5.02
67
-------
the critical value t, for a a - 0.01 significance level
one-tailed test with 30 degrees of freedom, is found in the
Appendix A table to be 2.457. The observed value of the test
statistic, 5.02, is much larger than the critical value and so
one would conclude that the mean level of PCB concentration in
the study area is larger than that in the background area.
While the difference in the two sample means was found to be
statistically significant at the 1% significance level, one may
still wonder whether the difference is scientifically
significant in terms of potential health hazard. We can be 99%
confident that the mean concentration of the study area exceeds
that in the background area by
xs - XB - ts /[(l/nB)+(l/nsn
• 52.61 - 40.23 - (2.457)(6.97)/(2/16)
•6.28 ppb.
This is a one-tailed confidence interval; Pg-
The t-tests are based on the assumptions that the data
are independent, normally distributed with equal variances, and
that all observations from the same sample have the same
expected value. In the two-sample t-test the assumptions of
normality and equal variance may be relaxed if sample sizes are
essentially equal. One-tailed one-sample t-tests on data from
a non-normal skewed distribution may have probabilities of Type
I and Type II errors that are considerably different from
those determined on the assumption of a normal distribution.
If the samples are not simple random samples but do have a
random component in their selection such as in stratified
random sampling, then the estimate of standard deviation and
the calculation of degrees of freedom will be affected. One
will use the positive square root of the ANOVA table mean
square for "Samples" as the estimate (s or sp) of standard
68
-------
deviation in the test statistic, and the degrees of freedom for
t will be the degrees of freedom for "Samples" in the ANOVA
table.
Consider again the data in Table 3 as coming from strata
of equal area and suppose the action level is 3.0. The test of
the hypothesis, H: y « 3.0, against the alternative, A: y >
3.0, would have test statistic,
tc «
-------
CHAPTER 5
EXPLORATORY STUDY
INTRODUCTION
Once objectives have been defined which involve the need
for sediment sampling, the next step is to develop a total
study protocol including an appropriate QA/QC program.
Generally, not enough information or data will be available to
proceed directly. The recommended approach is to conduct an
exploratory study first that includes both a literature and
information search along with selected field measurements made
on the basis of some assumed transport model.
In order to provide a framework for the discussion, a
hypothetical situation involving an abandoned hazardous waste
site will be described. In this scenario there is substantial
reason to believe that an abandoned waste site for hazardous
chemicals is leaking chemicals into the surrounding environment
which includes a few scattered farms and a medium size river
which empties into an estuary of the Gulf of Mexico about
twenty kilometers downstream.
The established objective for this hypothetical situation
is to conduct an environmental assessment of the site and its
environs to determine whether a short or long term hazard to
nan or the environment exists. If a hazard exists, its nature
and extent must be defined and appropriate recommendations made
to bring the hazard under control. Assume that a study team is
organized to address this problem and that the sediment study
group's task is to identify and make an assessment of potential
problems associated with sediments in the river and in the
estuary. Other members of the team will be concerned with
70
-------
soil, groundwater, and air pollution problems and their
consequences. All data gathered by specific members of the
team will be shared with the entire team.
Questions which must be answered by the exploratory study
include but are not limited to the following:
o What wastes have been placed at the disposal site
over what time periods?
o What chemicals in what amounts have escaped from the
site via what transport routes and what is the
present geographical extent of these chemicals?
o What adverse effects on human health or the
environment have been reported in the site vicinity?
o What is an appropriate background, or control region,
to use for the study?
Before t'aking any field measurements, a comprehensive
literature and information search should be conducted to
determine what information may already be available. Only
after relevant information has been collected, collated, and
evaluated should any field measurements be taken. The results
of the exploratory study will provide information and field
data that will serve as the basis for the design of a more
definitive monitoring study. Thus, any field measurements
taken should include appropriate QA/QC measures to determine
the quality of the data.
Assume that the information and literature search elicit
the following items. The wastes are from a chemical company
which specialised in petrochemical products. The wastes were
placed at the site beginning about forty years ago and ending
about fifteen years ago when the company went out of business.
Metal drums containing the wastes were covered over with a thin
layer of soil prior to abandonment of the site. Some of the
known constituents of the wastes have been listed as hazardous
by the USEPA. Complaints from nearby residents constitute
strong evidence that some of the hazardous constituents have
escaped from the site in surface waters, and because the
71
-------
groundwater at this site is not very deep, there is reason to
suspect that it too nay be contaminated. No quantitative
information was found on concentrations of the hazardous
chemicals in soil, surface waters, groundwater, air, locally
produced food, or sediments. A few recent studies in varied
locations were found in which measurements for some of the
hazardous chemicals of concern had been made in sediments. The
coefficient of variation for these studies averaged about 30%.
NUMBER AND LOCATIONS OF SITES FOR SAMPLING
The sediment study group concludes that there is
sufficient evidence to warrant conducting an exploratory study
ir. the sediments of the nearby river. Using the guidelines
i
suggested*in Chapter 3, plus information obtained from the
literature search, the following input factors are established
to determine the required number of samples: CV » 30%,
Confidence Level « 80%, Power * 95%, and Minimum Detectable
Relative Difference » 20%. The approximate number of samples
required for a one-sample one-sided t-test of the hypotheses,
B:p*L versus A:y>L may be calculated using the following
formula (Guenther, 1981)
n > [. ZQ)-a, Zg is similarly defined, and
D * (minimum detectable relative difference)/CV.
72
-------
Hence, for this example,
n j> I (0.842+1.645)/<20/30)l2 +0.5(0.842)2
n > 13.917-0.354*14.269
n * 15 (note:always round UP)
For a two-sided one-sample t-test, determine n by replacing Z
in the above formula with Z .2? that is, in the above example
replace 0.842 with 1.282 to obtain n«21.
For a one-sided two-sample t-test, the sample size for
each sample should satisfy the formula,
n > 2t(Z + ZC)/D12 +0.25Z 2.
— ap a
Again to obtain the corresponding minimum number for a
two-sided two-sample t-test, change Z to Z ,_.
To determine the locations of these samples, the following
approach is suggested. Estimate the sampling location(s) on
the river closest to the waste site via the likely surface
water flow. Label this spot zero on a coordinate system
extending down river. Stratify the study region and locate the
sampling points systematically as shown in the following
sketch.
73
-------
Sampling
Transects
Sartpling
Points
Figure 5. Sketch nap of river showing stratified regions and
sampling points.
The first stratum would be from 0 to 1 km, the second from 1 to
3 km, and the third from 3 to 7 km. Locate sampling transects
at 1/4, 1/2, and 3/4 the distance along the river from the
beginning of the stratum to its end. Locate sampling points
along the transects at 1/6, 1/3, 1/2, 2/3, and 5/6 the distance
from bank to bank. This provides 15 sampling points within
each stratified region as required.
It is suggested that a background region be established
approximately 10 km upstream from the 0 point of the river-
based coordinate system and extending about 1 additional km
upstream to define a region the same size as the first study
stratum. The fifteen sampling points in the background regron
would then be located as they are in the first study stratum.
The QA/QC program for the exploratory study Bust be
adequate for the resulting data to serve as a foundation for
further studies. For our hypothetical case, it is suggested
that three duplicate samples be collected from each stratified
study region (to include the background region as well).
74
-------
Also it is suggested that three samples from each stratified
region be split into triplicate samples. It is recommended
that a modest number of additional independent QA/QC sediment
samples be taken at approximate mid-points between selected
sampling points at locations in stratified regions in which the
hypothetical model predicts the highest concentrations will be
found. Data from these additional samples will give some
measure of how well the QA/QC plan is achieving its objectives.
In addition, all normal analytical QA/QC procedures such as
field and trip blanks, etc., should be operative for the
exploratory study.
SAMPLING AND SAMPLE HANDLING
An approved protocol should be followed for sampling,
handling, labeling, transporting, and chai n-of -custody
procedures for sample containers and samples. The possible
presence of volatile pollutants should be considered in the
selection of an appropriate protocol. Sample volumes will be
specified by the analytical laboratory depending on the
analytical methods to be used and the desired sensitivity.
Often, in addition to measurements of principal hazardous
constituents in sediments, other chemical, physical, or
biological measurements will be made for various purposes.
Examples of possible additional desired measurements for either
the exploratory or the definitive study are presented in Table
6.
75
-------
TABLE 6. COMMON MEASUREMENTS POR SURFACE WATER, AQUATIC
ORGANISMS AND SEDIMENT SAMPLING
Chemical
Dissolved oxygen
Phosphate
Physical
Color
Turbidity
Biological
Fish
Benthic Macroin-
Nitrogen series
Alkalinity
Silica
PH
Specific conductance
Solids (TDS,TS,TSS)
Organic matter and
demand
Pesticides
Heavy Metals
Hater temperature
Stream velocity
Water depth
Sediment composition
vertebrates
Periphyton
Phytoplankton
Zooplankton
Macrophytes
Macroalgae
Bacteria
Source: USEPA, 1982a
The sampling device used should be consistent with the
objectives of the final study. In general, the simplest
sampling tool deemed to be adequate should be used. The
advantages and disadvantages of some bottom grabs/sampler and
of some coring devices are presented in Tables 7 and 8,
respectively. It can be seen that all methods of sediment
sampling have disadvantages as well as advantages. When
choosing a sampler, weigh the type of samples needed to achieve
the objectives of the study against the advantages,
disadvantages, and cost of the various alternatives.
Surface sampling should normally be augmented with a
modest number of sediment core samples to determine how the
various measured parameters vary as a function of depth. These
additional samples should be located in areas in which the
highest contamination levels are expected. Data from these
samples will provide information for deciding if more than
76
-------
TMLE 7. OMVUUaON OF BOTTOM GRA£:VSMf>LERS
Devi<
ce
Advantage
DiBadvantagea
Ponar
Tall
Peteisuii
Smith-Mcintyre
Diver
Safe, easy to use, prevent* eacape of
•aterial with end plates, reducea shock
wave, coaDines advantagea of others (
preferred grab in mat cases
Use in soft aediaenta and calm waters,
collects atandaid aiae aaaple
(quantitative), reducea shock wave
Does not loae aadiaent over top; uae in
aoft sediments and caba water, standard
aaaple aize, reducea ahock wave
quantitative Maples in fine aadienta,
good for hatd bottoas and sturdy and
staple construction
Useful in bad weather, reducea
praaature tripping, uae in depths up to
1500 • (3500 ft), flange on jawa
reducea •aterial loss, acreen reduces
aback waves, good in all sediment types
Can detemne anst representative
aaapling point and current velocity
Can become buried in aoft sediments
Not useful in nuah water; not useful if
vegetation on bottoa
Not useful in rough waters, others as for Bonn
My loae aaapled Material, praaature tripping,
not eaay to close; does not sanple constant areas;
limited aaapling capacity
Urge, complicated and heavy, hazardous for
samples to 7 CM depth only, shock wave created
Requires costly equipment and special training
Source: U9B&, 1982a
-------
IMBUE 8. OWARI90N CF CORIIC DEVICES
Device
Disadvantages
CO
Kajak or
K. B. Corer
Moore (Pfleger)
O'Comer
Elgamfc's
Jenkina
Bnequist
Kirpicento
Obea not
pressure
free flew of water, no
, easily applied to large
Valve allows simple to be held
Can sample water with fund bottoms
Saaple easily reaoved, good in soft
suds, easy to collect, easty to remove
aaaple
Good in soft aediaenta and for
collecting an undisturbed
aedimcnt water interface aaaple.
Visual examination of benthic algal
growth and rough eatiaatea of mixing
near the interface after atoms can be
Good in soft/aediuai aediaents, closing
aechaniaa
Soft and hard bottoas, narioua siees,
cloaea autonatically
Careful handling necessary to avoid sediment
rejection, not in soft sediments
Not in deep water
Hot in hard sediments
Complicated
Does not penetrate hard bottom
Hot for stouy bottc
Source: USEPA, 1962a
-------
surface sediments need to be sampled in the final definitive
study.
Additional concerns in sampling design include whether
samples should be composited, frequency of sampling, sample
preparation for analyses, and the QA/QC aspects of all of these
parameters. The exploratory study provides a limited
opportunity to investigate some of the above subject areas.
The major concerns with regard to compositing of sediment
samples are that the samples be representative and that high
concentrations not be cancelled out in the calculation of the
mean by being averaged with too many low-level samples. The
best approach usually is not to composite unless there is
adequate justification for doing otherwise. The exploratory
study cannot be designed to obtain information on temporal
patterns in sediment concentrations since the study must be
completed in a relatively short period of time. Thus, temporal
trends should be addressed in the final study.
Sample preparation for analyses introduces some
possibilities for errors. The sample preparation may involve
drying, grinding, mixing, or sieving. Also, prior to sample
preparation, non-sediment material may be removed from the
collected sediment sample. Any equipment or devices used in
sample preparation must be carefully cleaned between each
sample to avoid cross-contamination. The final rinse fluid
used for cleaning equipment should be sampled to provide a
decontamination sample blank for use in evaluating the
cleanup efficiency. Collection of one sample blank after
processing each 20 samples has been used successfully in some
EPA Studies (USEPA 1982, 19S4).
One of the possibilities for error during the sampling
process is discarding non-sediment material collected with the
sediment sample prior to analysis. It is suggested that all
such discarded material be retained. Ten percent of these
samples should be sent to the analytical laboratory for
analysis with the remainder being archived. Care must be taken
79
-------
in evaluating and interpreting these data as data quality will
be a function of analytical capability.
In order to make this report more self-contained, the
enter chapter on Sample Handling and Documentation from the
companion soil document (Earth and Mason, 1984) is included in
Appendix B.
ANALYSIS AND INTEPRETATION OF DATA
Analysis and interpretation of all information and data
resulting from the exploratory study will provide the basis for
designing the final definitive monitoring study including all
elements of the QA/QC plan. For example, decisions must be
made on whether the selected control area is adequate; whether
the hypothesized model is valid; whether the study area should
be stratified in a different way; what number of additional
samples should be collected at what locations; whether the
QA/QC plan for sampling is adequate; etc. All deficiencies or
errors detected should be corrected in the final study design.
If the exploratory study is conducted well, it will
provide some data for achieving the objectives of the study; it
will provide data concerning the feasibility and efficacy of
most aspects of the study design including the QA/QC plan; it
will serve as a training vehicle for all participants; it will
pinpoint where additional measurements need to be made; and it
will provide a body of information and data for incorporation
into the final report for the total study.
A summary of some assumed results from an exploratory
study for the specific hypothetical case posed in this chapter
will be provided at the beginning of the next chapter. These
results will then be used to indicate corrections and additions
needed for the final definitive study.
80
-------
CHAPTER 6
FINAL DEFINITIVE STUDY
INTRODUCTION
Following analysis and interpretation of the information
and data resulting from the exploratory study the next step is
the design of the final definitive study. Any problems with
the QA/QC plan noted during the exploratory study should also
be solved by appropriate modifications of the plan. The
procedure will be illustrated by extending the hypothetical
case study defined in Chapter 5. To do this it is necessary to
present some assumed summary results from the exploratory study.
Accordingly, Table 9 gives mean values and standard deviations
obtained in the various stratified regions and in the
background, or control region, for the principal hazardous
constituent deemed to be critical in the sense of posing the
greatest potential danger to man or the environment. The units
are parts per billion in the sediments by weight.
81
-------
TABLE 9. SUMMARY OF SELECTED HYPOTHETICAL RESULTS FROM THE
EXPLORATORY STUDY.
Region Background(15)* 1(15) 2(15) 3(15)
(Stratum)
Mean (ppb) 1.24 13.2 15.1 11.5
CV (%) 30.3 45.2 40.7 47.6
Samples taken at different depths in Region 1
Depth Mean (pphm) CV(%)
0-4 in (5) 14.8 48.1
4-8 in (5) 5.21 52.4
8-12 in (5) 1.75 56.7
* Numbers of samples in parentheses.
Assume that three duplicates and three triplicates were
taken in each of the stratified regions as part of the QA/QC
plan for the exploratory study and that the resulting data
confirmed the adequacy of two duplicates and two triplicates
per stratified region. All normal analytical QA/QC procedures
were in force and no problems were identified. Other sampling
efforts confirmed the presence of the contaminant measured in
sediment- in surface water, groundwater, soil and selected
foods, with the largest concentrations observed close to the
hazardous waste site. Analysis of variance of the sediment
data showed that in excess of 70% of the total variance was due
to location.
Returning to an evaluation of the hypothetical results
shown in Table 9 allows certain tentative conclusions to be
drawn.
o Sediments are sufficiently contaminated to be a cause
for concern.
82
-------
o The background area selected is adequate (The mean
determined is close to other reported background
levels).
o The implicit hypothesized model which expected the
highest mean concentration to be in Region 1 is
questionable since Region 2 had a slightly higher
mean.
o The mean value for Region 3 suggests that sediments
farther downstream are likely to be significantly
contaminated.
o The depth measurements taken suggest that only the
top 8 inches of sediments may be contaminated
.. significantly.
In view of these conclusions certain matters will need to be
»
clarified in the definitive study. Some questions which should
be answered include the following:
o Bow far down stream are the sediments significantly
contaminated?
o What are the relative contributions of surface water
and groundwater to the contamination of sediments?
o How are the sediment levels changing as a function of
time?
o What are the levels of contamination in human foods
derived directly or indirectly through contact with
sediments?
o What is the impact of contaminated sediments on
aquatic biota?
o Row should the study area be stratified in the
definitive study?
These questions will be discussed at some length in subsequent
sections of this chapter.
It is likely that for a situation of this type an
emergency action level, as well as a long term residual level,
would be specified by a decision-making official if none exists.
The most likely media for such an action limit would be
83
-------
drinking water and/or foods. Such an approach would require
that a model be available or developed to link contaminant
levels in sediments to drinking water and/or food levels. Such
a derived level in sediments might be used as an operational
action level.
SELECTION OF NUMBERS OF SAMPLES AND SAMPLING SITES
Assume that, after careful consideration of all available
information, a decision official has come to the conclusion
that emergency action is not warranted but a remedial response
operation is called for. Referring back to Chapter 4,
recommended values for confidence level, power, and minimum
detectable relative difference are 90-95%, 90-95%, and 10-20%,
respectively. Table 11 presents the numbers of samples
required to achieve these values for different coefficients of
variation (CV). Table 10 below summarizes the situation over
the range of the recommended values for an assumed average CV
of 25%. This assumes that the CVs measured in the exploratory
study can be reduced by more judicious stratification of the
study region.
Table 10. NUMBER OF SAMPLES REQUIRED PER STRATIFIED REGION AS
A FUNCTION OF INDICATED PARAMETERS.
Confidence Level
95%
95%
90%
90%
Power
95%
90%
95%
90%
Minimum
Detectable Relative
Difference
10%
20%
20%
20%
No. of
Samples
> 69
> 19
> 15
> 12
The decision-making official decides to go with a
confidence level of 90%, a power of 95%, and a minimum
84
-------
TABLE 11. NUMBER OF SAMPLES REQUIRED IN A ONE-SIDED ONE-SAMPLE
t-TEST TO ACHIEVE A MINIMUM DETECTABLE RELATIVE
DIFFERENCE AT CONFIDENCE LEVEL (1-a) AND POWER
OF (1-6).
Coefficent Power Confidence Minimum Detectable
of Level Relative Difference
Variation
10 20 30 40
10 95 99
95
90
80
90 99
95
90
80
86 99
95
90
80
15 95 99
95
90
80
90 99
95
90
80
80 99
95
90
80
20 95 99
95
90
80
90 99
95
90
80
66 99
95
90
80
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
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
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
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
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
85
-------
TABLE 11. CONTINUED.
Coefficent
Of
Variation
(t)
Power Confidence
Level
Minimum Detectable
Relative Difference
10
20
25 95 99
95
90
80
90 99
95
90
80
80 99
95
90
80
30 95 99
95
90
80
90 99
95
90
80
80 99
95
90
80
35 95 99
95
90
80
90 99
95
90
80
80 99
95
90
80
397
272
216
155
329
272
166
114
254
156
114
72
571
391
310
223
472
310
238
163
364
224
164
103
775
532
421
304
641
421
323
222
495
305
222
140
102
69
55
40
85
70
42
29
66
41
30
19
145
99
78
57
120
79
61
41
84
58
42
26
196
134
106
77
163
107
82
56
126
78
57
36
28
19
15
11
24
19
12
8
19
12
8
5
39
26
21
15
32
21
16
11
26
16
11
7
42
35
28
20
43
28
21
15
34
21
15
10
14
9
7
5
12
9
6
4
10
6
4
3
19
13
10
7
16
10
8
5
13
8
6
4
25
17
13
9
21
14
10
7
17
10
7
5
9
6
5
3
8
6
4
3
7
4
3
2
12
8
6
4
11
7
5
3
9
5
4
?
15
10
8
6
13
8
6
4
11
7
5
3
86
-------
detectable relative difference of 20%. Accordingly, a minimum
of 15 samples will be required per stratified region which by
chance happens to be the same number of samples used in the
exploratory study. Additional QA/QC samples necessary have
been indicated in Table 4, Chapter 4. It is suggested that
fifteen additional depth samples be taken in Region 2 in the
same fashion as they were taken in Region 1 in the exploratory
study.
In deciding on how to stratify the study region for
the more definitive study, the information gained in the
exploratory study should be used. Since the means in Regions 1
and 2 for the exploratory were almost equal, it seems justified
to combine them into a single region. Thus, the suggested new
stratified regions are as shown in Table 12 below.
TABLE 12. NEW STRATIFIED REGIONS FOR THE MORE DEFINITIVE
STUDY.
Region A Region B Region C
0-3km 3-9 km 9- 21 km
Note: All regions now extend only from the near bank to the
middle of the river. See discussion below.
Note that the estuary into which the study river flows is 20 km
from the 0 point of the river coordinate system. Thus, Region
C extends 1 km from the mouth of the river into the estuary.
Location of sampling sites within the stratified regions
is the next order of business. Assume that analysis of data
from the exploratory study showed consistently that sampling
points from the middle of the river channel to the far bank
gave much lower levels than the other sampling points. This
finding serves as the basis for altering both the
stratification and the sampling site selection process for the
more definitive study into study regions extending only from
the near bank to the middle of the river channel.
87
-------
Also, note that combining old Regions 1 and 2 into new
Region A means that 12 measurements (the other 18 obtained are
now outside Region 4) are already available in Region A from
the exploratory study. It is recommended that 6 additional
samples be taken in Region A at sites 1/12 and 1/4 the distance
along the three sampling transects used for the exploratory
study. Region B contains 6 measurements from the exploratory
study, but with no measurements beyond kilometer 6. It is
suggested that 4 additional measurements (at sites 1/12, 1/6,
1/4, and 1/3 the distance along the cross-river transects)
be made at kilometers 7 and 8. In addition, 6 additional
samples should be taken in Region B at sites 1/12 and 1/4 the
distance along the sampling transects used for the exploratory
study. This will give a grand total of 20 measurements for
Region B. For Region C it is suggested that 4 samples each be
taken along transects (at sites 1/12, 1/6, 1/4, and 1/3 the
distance across the river) located at kilometers 11, 14, 17 and
20 and that 4 samples each be collected in the estuary at sites
1/12, 1/6, 1/4, and 1/3 the distance from the near shore and
along arcs centered at the mouth of the river and at distances
of 1/2 and 1 km. This will provide a total of 24 samples for
Region C. The plan proposed thus calls for the collection of
44 additional samples. The extra samples suggested for Region
C are to get a better estimate of the contamination of
sediments in the estuary.
Coordination would have to be established with water and
food sampling teams to assure that they direct a portion of
their more definitive study efforts to obtaining measurements
in water and food which might be related to sediment
measurements. It would be particularly important to obtain
samples of seafood harvested in the estuary.
Similarly, coordination would have to be established with
aquatic biologists assessing the impact of sediment
contaminants or aquatic biota. Particular attention should be
paid to assessing effects of the contaminants on juvenile
88
-------
populations of human food species as well as reproductive
success of the same species.
So far no attention has been given to the question
concerning relative contributions of surface water and
groundwater to the contamination of sediments. Perhaps data
obtained by the teams measuring these media close to the
hazardous waste site will provide some important evidence.
Geophysical remote sensing measurement tools may help to
delineate the groundwater hydraulic gradient and patterns of
groundwater flow in the vicinity. Also, estimates of total
contributions to contamination of sediments taken together with
estimates of surface water contributions enable the groundwater
contributions to be estimated by taking the difference between
trese two values. It is particularly important to have an
estimate pf the groundwater contribution and how it varies as a
function of time in order to evaluate the likely success of
different control options.
Sample collection, sample handling, and documentation must
be done in accordance with an established protocol. In this
instance, the same procedures used in the exploratory study
should be applicable to and adequate for the more definitive
study. If problems have been detected in the exploratory
study, appropriate modifications must be made to solve these
problems prior to proceeding with the more definitive study.
Table 13 contains some suggestions for sampling containers,
•
preservation requirements, and holding times for sediment
samples. Audits are perhaps the most effective tool to ensure
that all aspects of sample collection, sample handling and
documentation are being accomplished according to the approved
protocol (See Appendix D and USEPA, 1985).
The required frequency of sampling depends on the
objectives of the study, the sources and sinks of pollution,
the pollutant of concern, transport rates and disappearance
rates (physical, chemical, or biological transformations as
89
-------
Table 13. Sapling Containers, Preservation fbquiremente, end Holding Tie** for Sedieent Samples
CONTAMINANT CONTAINER PRESERVATION HOLDING TIME
'Acidity P.G
Alkalinity P,G
Ammonia P,C
Sulfate P,G
Sulfide P.G
Sulfite P.G
Nitrate P.G
Nitrate-Nitrite P.G
Nitrite P.G
Oil and Creese G
Organic Carbon P,G
Metsls
Chromii* VI P.G
Mercury P.G
Metals except above P.G
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°
Cool, 4°C
Cool, 4°C
Cool, 4°C
Organic Compounds
Extractables (including
phthalatea, nitroaamines
erganochlorine pesticides,
PCB'a nitroarometiea,
iaophorone, Polynuclesr
aromatic hydrocarbons,
haloethere, chlorinated
hydrocarbons and TCOO)
Citractablea (phenols)
Pur gable s (halocarbons
and aromatics)
Purgsblee (ecrolein and
acrylonitrete)
Orthophoaphate
Peaticidee
Phenols .
Phosphorus (elemental)
Phosphorus, total
Chlorinated organic
compounds cap
C, teflon-lined Cool, 4°C
cap
G, teflon-lined
cap
6, teflon-lined
aaptum
G, teflon-lined
aaptum
P.C
C. teflon-lined
cap
P.C
G
P.C
G, teflon-lined
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
Cool, 4°C
14 deys
14 daya
28 days
28 days
28 daye
48 houra
48 hours
2B daya
48 hours
28 days
28 daya
48 hours
28 daya
< Months
7 days (until i
30 days (after
ttrai
Ktn
:tion)
Ktion)
7 days (until extraction)
30 daya (after extraction)
14 days
3 days
46 hours
7 days (until extraction)
30 daya (after extraction)
28 daya
48 houre
28 days
7 days (until extraction)
30 days (after extraction)
Polyethylene(P) er Glaas(G)
Saaple preservation ahould be performed iaavdiately upon aaaple collection. For composite samples
each aliquot ahould be preserved at the tiae of collection. Nhen iepoaaible to preaerve eech
aliquot, then aaaplea aay be preaerved by aaintaining at 4°C until coepoaiting and aaaple splitting
la completed.
Saaplaa ahould be analyzed ae aoon aa possible after collection. The timea lieted are the mexi*nn
times thet eemples aey be held before anelyeis and atill considered valid. Samples may be held for
longer periode only if the analytical laboratory has dsta on file to ahow that the apecific types of
aaaples under atudy are atable for the longer time.
for additional information aae Ford at al. (1983).
90
-------
well as dilution or dispersion by any other means). Sampling
frequency nay be related to changes over tine, season, or
precipitation. Little infornation will be available on
sanpling frequency from exploratory study data. However, these
data will provide baseline information at a given point in time
from which future trends nay be measured. Assessment of future
trends will establish whether sediment concentrations are
increasing, decreasing, or remaining fairly level. Evaluation
of these trends will be important to selection of appropriate
remedial response measures or to the determination that
remedial response measures will not be required.
The recommended procedure for establishing time trends is
to sample monthly for the first year. Evaluation of the trend
of the data will then enable a determination to be made
concerning possible changes in campling frequency. If the only
concern is for time trends in each stratified study region,
then compositing 15 or nore samples from each region for each
monthly sample may be the simplest way to proceed. On the
other hand, if the changing of spatial patterns with time is of
interest, the compositing approach would not be recommended.
In the latter case, the time trends for changes in individual
samples at definite locations would be needed. Thus the
preferred approach would be to repeat the sampling program
previously described at monthly intervals until sufficient data
accumulate to justify changing the sampling frequency intervals.
The major focus should be on the highly contaminated and
immediately adjacent areas.
Quality assurance/quality control procedures for frequency
of sampling validation may be accomplished through techniques
such as trend line or interdiction analysis. Also, the taking
of initial samples on a frequency considered to be more often
than is likely to be required may provide some redundant data
but will assist in verifying the adequacy of the sampling plan.
A comparison between the first samples taken and the most
recently collected samples should show a decrease in pollutant
91
-------
concentrations unless there is a new source of pollutants,
there is migration into the sampled sediments, chere is an
error in the data, or the decrease is not sufficient to be
resolved due to the variability of sample data. This test
becomes a better indicator the longer the study runs.
The analysis and interpretation of QA/QC data from the
more definitive study should show how all aspects of the total
QA/QC plan combine to give an overall level of reliability for
various aspects of the resulting data. Another goal may be to
determine whether all QA/QC procedures used were necessary and
adequate. This entire evaluation must be closely linked to the
objectives of the study. In summary the important questions to
be answered are, "What is the quality of the data?" and "Could
the same objective have been achieved through an improved QA/QC
design which may have required fewer resources?"
It is desirable to provide summarized tables of validated
QA/QC data in the final report. For example, QA/QC data
validation procedures used in a number of soil sampling studies
reported by Brown and Black (1983) included validation of
sample data sets by checking and assessing the accompanying
QA/QC data. In order to make this report more self-contained,
the entire chapter on analysis and interpretation of QA/QC data
from the companion soil document (Barth and Mason, 1984) is
included in Appendix C. This approach is equally applicable
for sediment sampling data. The criteria for QA/QC samples and
procedures used to validate all data should be clearly stated.
From such tables of validated QA/QC data it is possible to
determine bias, precision (total random error), component
random errors associated with reproducibility, extract matrix,
sample matrix, and sample homogeneity, interlaboratory
precision, and uncertainty.
Presentation of QA/QC data allows readers to verify
conclusions drawn as to reliability of the data. Such
presentation also contributes to the building of a body of
QA/QC data in the literature which allows comparison to be made
92
-------
between and among studies. Special emphasis should be placed
on explaining how overall levels of precision and confidence
were derived from the data.
As a final check, the adequacy of all aspects of the QA/QC
plan should be examined in detail with emphasis on defining for
future studies an appropriate minimum adequate plan. Some
aspects of the plan actually used may have been too
restrictive, while others may not have been restrictive enough.
Appropriate analyses and interpretation of the data should
identify the actual situation.
There is insufficient knowledge dealing with sediment
monitoring studies to state with confidence which portions of
the QA/QC plan will be generally applicable to all sediment
monitoring studies and which portions must be varied depending
.on site-specific factors. As experience is gained, it may be
possible to provide more adequate guidance on this subject. In
the meantime, it is recommended that the best approach is to
assume that important factors of QA/QC plans may be site
specific and to conduct an appropriate exploratory study at
each new study site to verify that various aspects of the QA/QC
plan are adequate to meet program objectives prior to
proceeding with the final definitive study.
In lieu of providing hypothetical data resulting from the
more definitive study, a brief general discussion will be
provided indicating possible conclusions which might be drawn
from the data. Comparison of the calculated means and standard
deviations for each stratified study region to any assigned
action level by appropriate statistical methods outlined in
Chapter 4 will establish which stratified regions presently
have concentrations exceeding acceptable limits. If action
levels are only specified for drinking water and foods, then an
estimated comparable action level for sediments must be derived
from an appropriate model.
If time trend analyses indicate that concentrations in
sediments are increasing with time, peak values have not yet
93
-------
been achieved. In this case, available data from the study
teams should be combined with alternative control options and
an appropriate model to predict when and where the maximum
future values will be found as well as their estimated peak
concentrations.
If time trend analyses indicate that concentrations in
sediment are decreasing with time, projected values for the
future should be predicted by combining data from their study
teams with alternative control options and an appropriate model.
If the trends show concentrations decreasing rapidly enough,
there may be no necessity for control actions.
The case in which time trends show fairly constant values,
or sometimes increasing and sometimes decreasing ones, should
be treated similarly to the case in which concentrations are
increasing with time.
For the more definitive study, additional measurements in
sediments over and above the concentrations of the hazardous
waste of concern should include as a minimum the following for
each sampling period:
Depth of the river
Flow rate
Suspended solids
Bed load
PH
Temperature
Living species populations and diversity in sediments
Body burdens of the hazardous waste for selected species
dwelling in sediments
Adverse effects on selected species dwelling in sediments
The purpose of these extra measurements, in addition to their
intrinsic value, is to validate existing sediment transport
models or provide data on the basis of which modifications may
be made in existing models or new models may be developed. The
biological measurements may assist in either defining adverse
94
-------
effects on sediment biota or in providing information for
linking contamination in sediment biota to contamination in
human foods via models.
Data from the more definitive study describing variations
in sediment concentrations with depth will show how effective
dredging to different depths might be in the removal of the
contamination. If dredging is even contemplated, safe and
effective methods for disposing of the dredge spoil must be
available.
95
-------
REFERENCES
Allen, J. R. L. Physical Processes of Sedimentation: An
Introduction. Allen 4 Unwin, London. 1970. 248 pp.
Allmaras, R. R. Bias. In: Methods of Soil Analysis, Part 1,
Physical and Minerological Properties, including Statistics of
Measurement and Sampling. C. A. Black, et al., ed. American
Society of Agronomy, Madison, Wisconsin, 1965. 24-42 pp.
Ball, C. B., L. L. Conquest, R. Pyke and E. P. Smith. Some
Nonparametric Statistics for Monitoring Water Quality Using
Benthic Species Counts. Environmetrics 81: Selected Papers.
ElAM: Philadelphia, 1981. 260 pp.
Burnett, V. and T. Lewis. Outliers in Statistical Data. John
Wi-ey & Sons, New York, NY, 1978. 365 pp.
Barth, D. S. , and B. S. Mason. Soil Sampling Quality Assurance
User's Guide. EPA-600/4-84-043. Environmental Monitoring
Systems Laboratory, Las Vegas, NV. 1984.
Bauer, Edward L. A Statistical Manual for Chemisto. Academic
Press, New York, New York, 1971. 193 pp.
Beyer, W. H. CRC Handbook of Tables of Probability and
Statistics (2nd ed.). The Chemical Rubber Company, Cleveland,
OH, 1968. 642 pp.
Biggs, R. B. Coastal Bays. In: Coastal Sedimentary
Environments, R. A. Davis, ed. Springer-Verlag, New York. 1978.
pp. 69-99.
Bishop, Y. M. M., S. E. Fienberg, and P. W. Holland. Discrete
Multivariate Analysis. The MIT Press, Cambridge, MA, 1975. 557
PP.
Box, G. E. P. and D. R. Cox. An Analysis of Transformations.
Journal of the Royal Statistical Society, Series B. 26(2):211-243.
1964.
Broms, Bengt B. Soil Sampling in Europe: State-of-the-art.
Journal of the Geotechnical Engineering Div., 106:65-98.
1980.
96
-------
Brown, K. W. and S. C. Black. Quality Assurance and Quality
Control Data Validation Procedures Used for the Love Canal and
Dallas Lead Soil Monitoring Programs. Environmental Monitoring
and Assessment, 3:113-112. 1983.
Davis, R. A. Jr. Depositional Systems: A Genetic Approach to
Sedimentary Geology. Prentice-Hall, Inc. 1983. 669 pp.
Duncan, A. J. Bulk Sampling: Problems and Lines of Attack.
Technometric 3 (3);319-344. 1962.
Elder, R. S. , W. 0. Thompson and R. H. Myers. Properties of
Composite Sampling Procedures. Technometrics 22 ( 2 ): 179-186 .
1980.
Eynon, Barry and Paul Switzer. A Statistical Comparison of Two
Studies of Trace Element Composition of Coal Ash Leachates.
Final Report. Electric Power Research Institute, Palo Alto,
California. July, 1983. EA-3181.
Farnesworth, E. G. , M. C. Nichols, C. N. Vann, L. G. Wolf son, R.
W. Bosserman, P. R. Hendrix, F. B. Golley, J. L., Cooley. Impact
of Sediment and Nutrients on Biota in Surface Waters of the
United States. EPA-600/3-79-105. 1979. 314 pp.
Ford, Patrick J., Paul J. Turina, and Dougles E. Seely.
Characterization of Hazardous Waste Sites--A Methods Manual.
Vol. II. Available Sampling Methods. EPA 600/4-83-040. U.S.
Environmental Protection Agency, Environmental Monitoring
Systems Laboratory, Las Vegas, NV. 1983.
Garrels, R. M., F. T. Mackenzie, and C. Bunt. Chemical Cycles
and the Global Environment: Assessing Human Influences. Wm.
Kaufmann, Inc. 1975. 206 pp.
Grimsrud, G. P., E. J. Finnemore, and E. J. Owen. Evaluation of
Water Quality Models: A Management Guide for Planners. 1976.
U.S. Environmental Protection Agency, Office of Research and
Development, Washington, D. C. 1976.
Guenther, W. C. Sample Size Formulas for Normal Theory T Tests.
American Statistician, 35(4):243-244. 1981.
Guttman, I., S. S. Wilks and J. H. Hunter. Introductory
Engineering Statistics (3rd ed.) John Wiley, New York, 1982.
580 pp.
Gy, P. M. Sampling of Particulate Materials. Elsevier
Scientific Publishing Company, New York, NY, 1982. 431 pp.
97
-------
Hahn, G. J. Factors for Calculating Two-Sided Prediction
Intervals for Samples from a Normal Distribution. J. American
Statistical Assoc. 64(327): 878-888. 1969.
Hansen, M. H., W. N. Hurwitz, and W. G. Madow. Sample Survey
Methods and Theory, Vol. I. John Wiley 6 Sons, New York, NY,
1953. 638 pp.
Boaglin, D. C., F. Hosteller, and J. W. Tukey. Understanding
Robust and Exploratory Data Analysis. John Wiley 6 Sons, New
York, NY, 1983. 447 pp.
Hough, J. L. Geology of the Great Lakes. University of Illinois
Press, Urbana, IL. 1958. 313 pp.
Journel, A. G. and Ch. J. Buijbregts. Mining Geostatistics.
Academic Press, New York, NY, 1978. 600 pp.
Juran, J. M., F. M. Gryna, Jr. and R. S. Bingham Jr., eds.
Quality Control Handbook, Third Edition, McGraw Hill. 1979.
Krenkel, P. A. and Novotny, V. Water Quality Management,
Academic Press, New York, p. 671. 1980.
Leytham, K. M. and R. C. Johanson. Watershed Erosion and
Sediment Transport Model. EPA-600/3-79-028. 1979. 355 pp.
McKay, D. and S. Paterson. Spatical Concentration Distributions.
Environmental Science and Technology. 16(7):207A-214A. 1984.
Me Nelis, D., D. Barth, M. Khare, T. LaPoint, and E. Yfantis.
Exposure Assessment Methodologies for Hazardous Waste Sites.
DER-006(84). 1984. 217 pp.
OECD Guidelines for Testing of Chemicals, Organization for
Economic Cooperation and Development, Paris, France. 1981.
Oregon -State University. Methods of Soil Analysis Used in the
Soil Testing Laboratory at Oregon State University. Special
Report 321. Corvallis, Oregon. 1971.
Peterson, R. G. and L. D. Calvin. Sampling. In: Methods of
Soil analysis. Part 1, Physical and Mineralogical Properties,
including Statistics of Measurement and Sampling. C. A. Black,
et al., ed. American Society of Agronomy, Madison, Wisconsin,
1965. 54-71 pp.
Plumb, Russell H. Jr. Procedures for Handling and Chemical
Analysis of Sediment and Water Samples. EPA-48/05-5720-10, U.S.
E.P.A./Corps of Engineers Technical Committee on Criteria for
Dredged and Fill Material, Grosse lie, Michigan, 1981.
98
-------
Pritchard, D. W. Estuarine Circulation Patterns. Amer. Soc.
Civil Eng. Proc. 1955. pp. 717/1-717/11.
Scheffe, H. The Analysis of Variance. John Wiley, New York.
1959. 477 pp.
U.S. Code of Federal Regulations, Land Treatment, 40CFR, Part
264, Subpart M, July 1983.
U.S. Environmental Protection Agency. Standard Operating
Procedures for Conducting Sampling and Sample Bank Audits.
EPA-600/4-85-003, Las Vegas, NV. 1985.
U.S. Environmental Protection Agency. Documentation of EMSL-LV
Contribution to Dallas Lead Study EPA-600/4-84 - 0 1 2 .
Environmental Monitoring Systems Laboratory, Las Vegas, NV.
1984.
U.S. Environmental Protection Agency. Documentation of EMSL-LV
Contribution to the Kellogg Idaho Study. EPA 600/X-84-052 ,
Environmental Monitoring Systems Laboratory, Las Vegas, NV,
1984a.
U.S. Environmental Protection Agency. The Development of Data
Quality Objetives (QAMS). Draft. Washington, D.C. 1984.
U.S. Environmental Protection Agency. Environmental Monitoring
at Love Canal. Volumes I - III. EPA-600/4-82-03 0 a-d,
Washington, D.C. 1982.
U.S. Environmental Protection Agency. Handbook for Sampling and
Sample Preservation of Water and Wastewater. EPA-600/4-82-029,
Cincinnati, Ohio. 1982a.
U.S. Environmental Protection Agency. Interim Guidelines and
Specifications for Preparing Quality Assurance Project Plans.
QAMS-005/80. Washington, D.C. 1980.
U.S. Environmental Protection Agency. Guidelines Establishing
Test Procedures for the Analysis of Pollutants; Proposed
Regulations. Federal Register, 44:233, pp. 69464-69575,
Washington, D.C. 1979.
Williams, P. F. and Rust, B. R. The Sedimentology of a Braided
River. Jour. Sed. Petrology, 39:649-679, 1969.
Youder, W. J. and E. H. Steiner. Statistical Manual of the
association of Official Analytical Chemists. Association of
Official Analytical Chemists, Washington, D.C. 1975. 88 p.
99
-------
APPENDIX A
FEROBNTILES CF THE T DISTRIBUTION
Confidence Level (%):100(l-a) for two-tailed test
20
40
60
Confidence Level (%)
df
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
40
60
120
03
60
.325
.289
.277
.271
.267
.265
.263
.262
.261
.260
.260
.259
.259
.258
.258
.258
.257
.257
.257
.257
.257
.256
.256
.256
.256
.256
.256
.256
.256
.256
.255
.254
.254
.253
70
.727
.617
.584
.569
.559
.553
.549
.546
.543
.542
.540
.539
.538
.537
.536
.535
.534
.534
.533
.533
.532
.532
.532
.531
.531
.531
.531
.530
.530
.530
.529
.527
.526
.524
80
1.376
1.061
.978
.941
.920
.906
.896
.889
.883
.879
.876
.873
.870
.868
.866
.865
.863
.862
.861
.860
.859
.858
.858
.857
.856
.856
.855
.855
.854
.654
.851
.848
.845
.842
80
: lOO(l-a)
90
3.078
1.886
1.638
1.533
1.476
1.440
1.415
1.397
1.383
1.372
1.363
1.356
1.350
1.345
1.341
1.337
1.333
1.330
1.328
1.325
1.323
1.321
1.319
1.318
1.316
1.315
1.314
1.313
1.311
1.310
1.303
1.296
1.289
1.282
90
95
for one-tailed
95
6.314
2.920
2.353
2.132
2.015
1.943
1.895
1.860
1.833
1.812
1.796
1.782
1.771
1.761
1.753
1.746
1.740
1.734
1.729
1.725
1.721
1.717
1.714
1.711
1.708
1.706
1.703
1.701
1.699
1.697
1.684
1.671
1.658
1.645
97.5
12.706
4.303
3.182
2.776
2.571
2.447
2.365
2.306
2.262
2.228
2.201
2.179
2.160
2.145
2.131
2.120
2.110
2.101
2.093
2.086
2.080
2.074
2.069
2.064
2.060
2.056
2.052
2.048
2.045
2.042
2.021
2.000
1.980
1.960
98
test
99
31.821
6.965
4.541
3.747
3.365
3.143
2.998
2.896
2.821
2.764
2.718
2.681
2.650
2.624
2.602
2.583
2.567
2.552
2.539
2.528
2.518
2.508
2.500
2.492
2.485
2.479
2.473
2.467
2.462
2.457
2.423
2.390
2.358
2.326
99
99.5
63.657
9.925
5.641
4.604
4.032
3.707
3.499
3.355
3.250
3.169
3.106
3.055
3.012
2.977
2.947
2.921
2.898
2.878
2.861
2.845
2.831
2.819
2.807
2.797
2.787
2.779
2.771
2.763
2.756
2.750
2.704
2.660
2.617
2.576
100
-------
APPENDIX B
SAMPLE HANDLING AND DOCUMENTATION
INTRODUCTION
The goal is to define the segment of the QA/QC plan dealing
with all aspects of sample handling including the transfer of the
sample from the collecting device to a suitable container,
transportation of the sample, and the preparation of the sample
for analysis. The importance of all these aspects of sample
handling and possible errors introduced thereby will naturally
vary with the sampling methods, monitoring objectives,
characteristics of the soil being sampled and the physical and
chemical properties of the pollutants of concern.
>
CONTAINER PREPARATION, LABELING, PRESERVATION, AND SAMPLE
PREPARATION
The sampling protocol and the QA/QC plan must address the
following factors.
o Type of container material, its size, shape and the type
of lid.
o Cleaning procedures for the containers
o Decontamination procedures for sampling instruments.
o Decontamination procedures for sample bank equipment.
o Labeling scheme and log book entries
o Chain of custody procedures
o Sample preparation procedures in the field
o Sample preparation procedures at the sample bank
Due to a lack of specifically tested and recommended methods
dealing with the storage, handling, construction and types of
containers, cleaning and decontamination of containers, and
101
-------
suggested materials for container lids for soil samples it is
suggested that the specifications and methods identified in
USEPA, Federal Register Vol. 44 No. 233 (1979) be utilized.
Table B-l provides general information on recommended
containers, preservation requirements, and holding times for
measuring selected contaminants. Even though these procedures
and methods were specifically designed and tested for water
samples, they are applicable for soil sampling studies.
For sampling studies that require a large number of samples
and/or extensive preanalytical sample preparation a sample bank
may be established. The sample bank is the element that operates
between the field sampling effort and the analytical laboratory.
However, for smaller studies the sample banks responsibilities
are often incorporated into the responsibilities of the field
sampling team or the analytical laboratory.
If a sample bank is established, sample bank personnel can
assume responsibility for the following procedures:
o Custodian for all records pertaining to the sampling,
sample preparation as required, and shipment of soil
samples to analytical laboratories.
o Responsibility for record filing and storing, for
storing and preparation of soil samples, and for
dispensing containers, sampling equipment and all
custody documents such as chain-of-custody forms and
sample collection and analytical tags, as required.
o Responsibility for updating and maintaining the
projects' master log book, auditing the records as
required, generating sample bank QC sample blanks,
accepting QA/QC samples for inclusion into the
analytical scheme, and for scheduling the collection of
field sample blanks.
o Responsibility for completing, as required, analysis
data reporting forms and for assuring that all
chain-of-custody requirements pertaining to all
field sampling, shipping and sample bank operations,
are adhered to.
102
-------
table o-l toapllno, Contatnore, Prooer-etion toatuimente, and Holding liaoe for toll taaeloe
CONTWIMMT CONTAJNie. •H£SC«»*T10«. MXD1NC TMI
teiaity P.C Cool, «°C 1* toye
Alkalinity P.C Cool, e«C ]« a»yi
OtoBrue P.C COOl, «°C
Sulfate P.
Sul'ito P.
lulfite P.
Mitrate f,
Nitrate-Nitrite P,
Nitrite F,
Cool, a°C 21 a*re
Cool, »°C M dkye
Cool, *°C *J houre
Cool, *°C *• houre
Cool, *°C 21 toy*
Cool, «° o* hourt
Oil and Crem G Cool, *°C 21 toys
Organic Carbon P,C Cool, *°C tt toy*
to UU
OVOBIUB VI P,C Cool, *°C OJ noura
Mtrtury P.C tt tore
Netele oieept oben »,G i oonthe
OTaonic
Citroctotloe (including C, teflon-lined Cool, a°C 7 tor* (until attraction)
phthalate*, futroao»ina« coo XI toy* (after extraction)
aigenocMennr ao*ticioet,
PCf*
•tic
Holoethire, etilorinoted
hydroeortone and 1CDt»
CitrertoblM (ohenole) C, te'lorv-linta Cool, «<* 7 toyi (until vtrection)
top 10 toy* (oftor 01 traction!
•urgtblaa (holonrton* C, teflon-l»»fl Cool, »°C 1* toy*
ond •TOBotir*) toptu*
•urgoblee (ocrolein and C, teflon-lined Cool, *°C J toy*
•rrylomtrott) ooptu*
P.C Cool, *°C *t hour*
C, tefion-lined Cool, ••C 7 tore (until o»tract 10"-
oop • toy* (after attroctton)
P.C Cool, t»C tl toy*
C Cool. «°C «• houri
>ruo, total P.C Cool, «°C n tov*
C. toflon-lined Cool, »*C 7 toy* (until totroction)
aap » toy* (oftar • traction)
Polyethylene If) or UoaeCC)
Satole froooriation ohould to oerfomed iaoediotaly uoon oBeple oollaetion. for aoipoeiU aaocle*
oat*- aaiouet onould to »r»o»r»od ot tto tio* of collection. Men tao«**lbJe to pre*er«* oocn
aliavtl, tt«n *aople» aay to arooomod k> auntoining ot «*C until eatoooiting ond oooelr ovlitting
to ooBpleted.
looele* onould to onelyiod •• toon o* oaooiblo after eollortion. The Uaee lie ted or* tne aniour
Uaoo that oooplo* oo> to hold before onolyaio and etill eonaitorod valid. Saoelae My to hold for
tanoer aeriad* only if the analytical laboratory n§* tot* on file to •*» ttwjt the tootific tyoe* of
aoaple* vntor otudy arc otoble for the longer tie*.
for ortttionol mfom*tior< ox ford ot ol (1M3).
103
-------
The following sample bank procedures have been used
successfully on a number of soil monitoring studies.
A. Issuing Supplies:
(1) The sample bank issues as required sample
containers, sample collection tags, chain-of-custody
forms and site description forms to the sampling
teams. Sample collection tags and chain-of-custody
forms are normally accountable documents; the sample
bank will log the forms by numerical lot identifying
the team and/or the individual responsible for the
temporary custody of these documents.
(2) The sample bank may be required to store sampling
equipment in a suitable environment. If sampling
equipment is stored at the sample bank, issuing this
equipment to the sampling teams as required will be
necessary.
B. Accepting and Logging Samples:
(1) Transfer of sample custody from the sampler to
sample bank personnel will normally occur at the
sample bank.
(2) Before accepting custody of any samples, sample bank
personnel must check all tags and forms for
legibility and completeness.
(a) All individual samples must have a completely
filled out sample collection tag attached.
(b) Every sample must be identified on the
chain-of-custody form.
(c) Each site sampled must have a completely filled
out site description form.
(d) Any discrepancy will be corrected before sample
bank personnel will assume custody. If a
discrepancy exists that cannot be resolved to
the satisfaction of the sample bank personnel,
resampling, filling out additional tags and
forms, and/or revisiting the site to obtain
necessary documentation may be required.
104
-------
(e) All unused accountable documents as shown in
Table B-2 must be returned to the sample bank
on a daily basis. However, depending upon
circumstances such as a sampling team's
schedule and route, accountable documents may
be retained by the sampling team leader. The
sample bank supervisor, however, Bust be aware
of the situation.
(3) After the sampler relinquishes custody and the
sample bank personnel assumes custody of the
samples, each sample must be logged into the master
log book.
Preparation of soil samples for analysis may require sample
bank personnel to dry, sieve, mix and aliquot samples
appropriately. The preparation procedures selected are
determined by the contaminant to be measured and the analytical
requirements. Various techniques and methods for mixing and
compositing soils have been described by Oregon State University
(1971), USEPA (1984), and Peterson and Calvin (1965).
It is inappropriate to initiate a sampling study without
first consulting with analytical personnel. Collecting samples
that cannot be suitably analyzed will not provide data necessary
for satisfying the sampling objectives.
There is the possibility of errors being introduced in
sample preparation procedures involving the discarding of
non-soil material or of non-sieved material as well as possible
losses during any grinding or drying operation. The definitive
study decisions concerning the non-soil fraction must be made on
the basis of the data obtained from the exploratory study. For
example, available data nay indicate that significant
contamination is in the discarded portion. If so, it is
recommended that the discarded portion from ten percent of the
samples collected from the area having the highest concentrations
be analyzed. An estimate can then be made of the total amount of
contamination being discarded by multiplying the measured
concentration in the discarded material by the total amount of
the discarded material. Assuming that this amount is uniformly
distributed through the soil sample remaining after non-soil
materials and non-sieved materials have been discarded, one can
then calculate an estimated value for the potential soil sample
total concentration if none of the contamination had been
discarded. Comparison of this potential concentration to the
actual measured concentration will enable an estimate of the
possible error to be made.
105
-------
TAMLR •-?. ACCnUNTAILE OOCttNflfT CONTROL BHJUI «**.«•*
O
Ot
Oocao*ntation
SMO|* Collection Tan*
C«atoo> Rvcnrtfa
rieU Loabook*
Sit* OcacriMl»i> Tort*
Analytical Sanolr Ta«a
Laboratory Not*booka
Analytical Data •***»•
laa««d by
Saaol* ••nk
Sao*U Hank
•anylo lank
Saa»l» lank
Jlaa^U lank
Laboratory
Itooyla lank
Interim
Nunbvrinit Meaoonfibi 1 ity
Pr*»*ri«lt«»d Sanollni Team
Prraerialiicd SaoolinR Teao>
tan* lint f*«*
ilaaolint Team
•r*a»r ia!it*
-------
If the error estimated by this process exceeds acceptable
limits specified in the QA/QC plan, it night be necessary to
modify sample preparation procedures for the definitive study.
One might consider a sample preparation procedure in which the
entire collected sample (soil and non-soil materials) is
extracted in the analytical laboratory. The analytical results
could then be reported as amounts of contaminant per gram of
mixed material. At present there is no acceptable method for
proceeding in cases such as these. One problem is the lack of
standard reference materials for determining and measuring errors
in extraction efficiency. One solution may be to try different
methods of extraction and compare the results. The final
interpretation of the data must then take into consideration
these estimated errors.
QUALITY ASSURANCE ASPECTS
The problem is to quantitate overall errors. The
recommended procedure for verifying that the QA/QC plan is being
carried out properly for this chapter's factors is a periodic
ai'dit, combined with a modest amount of extra samples and
analyses related to factors discussed above.
107
-------
APPENDIX C
ANALYSIS AND INTERPRETATION OF QA/QC DATA
INTRODUCTION
One goal in the analysis and interpretation of data is to
show how all aspects of QA/QC for a soil monitoring study combine
to give an overall level of precision and confidence for the data
resulting from the study. Another goal may be to determine
whether all QA/QC procedures which were used were necessary and
adequate and should definitely be incorporated into future
studies of the same type. This entire evaluation must be closely
linked to the objectives of the study. In summary the important
questions to be answered are, "What is the quality of the data
(maximum accuracy attainable)?" and also," Could the same
objective have been achieved through an improved QA/QC design
which may have required fewer resources?"
PRESENTATION OF DATA SUMMARIES
It is desirable to provide summarized tables of validated
QA/QC data in the final report. For example, QA/QC data
validation procedures used in a number of soil sampling studies
reported by Brown and Black (1983) included validation of sample
data sets by checking and assessing the accompanying QA/QC data.
The criteria for QA/QC samples and procedures used to validate
all data included:
108
-------
Samples and Procedures Example Criteria
1. Reagent BlanksConcentrations had to be less
than 0.25 g/ml'1.
2. Calibration Check Recovery must be between 95% and
Standards 105% of the known value for
either the first analysis or the
first re-check analysis.
3. Laboratory Control Recovery must be between 90% and
Standards 110% of the known value for
either the first analysis or the
first re-check analysis.
Data produced by any sampling and analyzing system are
affected by two types of errors; random and systematic. The
accuracy of any one result then, is a function of the bias (due
to systematic error) and precision (due to random error) of the
collection and analysis methodology. Bias has at least two
components, associated with extraction and instrument efficiency/
and is assessed by the mean recovery of Calibration Check
Standards and Laboratory Control Standards (LCS). The LCS check
overall bias for the system; the Calibration Check Standard
determines the instrumental bias.
Total random error can be assessed by analyzing duplicate
samples, but it includes errors due to sample collection, sample
homogeneity, sample extraction, sample composition (matrix
effects) and instrumental reproducibility. These errors can be
evaluated by the use of the other QC procedures stated above and
are assessed by calculating the standard deviations of the
various analyses.
The accuracy of analysis, i.e., bias and precision, are
evaluated separately below for the two types of samples, using
the following equations:
Recovery » Amount Found/Known Amount (1)
Bias (B) • Recovery -1. ' (2)
109
-------
Difference (D) » | xi - *2\ where x^ and X2 are the analytical
results of paired analyses and the average is:
n
and the precision is:
Precision « 0.8862 5
(4 )
where 0.8862 converts the range of two results to the standard
deviation (Natrella, 1963).
then
If component errors are used to assess total random error,
0 « (Di + 02 + ...)/n and
» Precision « [0.8862 (Di2 + D22
) + sj.2 + S2
(5)
. .. J1/2
Equation (3) is suitable for use on results where the
concentration varies over a very narrow range. If the
concentrations found vary by an order of magnitude or more, then
the difference should be normalized by dividing by the average of
the two values and the precision is expressed as the coefficient
of variation (CV) which is s/x
T |X - X
2 i LL,
n ,.
i .
a II
CV - 0.8862 D
(7)
One of the studies discussed by Brown and Black (1983)
involved lead contaminated soils. The use and evaluation of the
QC analyses for this soil monitoring study was presented as
follows:
110
-------
The limit of detection, approximately 0.25 yg ml"1, was tested on
about 10 blank analyses using a more sensitive absorbance
wavelength for lead on an AAS. The result was less than 0.1 yg
ml"1, or 2 yg g"1 for sample analysis. This suggests that most
of the blank analyses were less than 2 yg g"1, but this cannot be
stated with any confidence. The results of the QC analyses were
as follows:
QC Sample
Calibration Check Standard
Laboratory Control Standard
Field Blank (yg ml'1)
Sample Bank Blank (yg ml"1)
Reagent Blank (yg ml'1)
Re-extraction Analysis
Total Recoverable
Split Extract (CV)
Spiked Extract
Spiked Sample
Duplicate Aliquot (CV)
Duplicate Sample (CV)
Triplicate Analysis (CV)
No.
150
147
76
77
148
17
144
147
147
147
134
129
220
Mean
101.5%
101.2%
<0.25
<0.25
<0.25
1.7%
99.8%
0.0089
99.4%
100.4%
0.053
0.189
0.144
S
2.6%
4.1%
1.4%
8.0%
0.0079
5.0%
5.1%
0.047
0.168
0.128
(1) Bias: The percent recoveries indicated above .for the
Calibration Check Standards and LCS's suggest a small positive
bias for the method of soil analysis, due principally to
instrument reproducibility. The result, using Equation (2), is:
Bias * Recovery - 1 - 1.012 - 1 - 0.012.
(2) Precision: The recovery of the analyte by the analytical
method compared to the "total" recoverable method was essentially
equal and re-extraction of the residue left from the initial
extraction indicated an additional 1.7 ±1.4 percent recovery,
also essentially equivalent. Furthermore, the results of the
three types of blank analyses indicate no measurable
contamination from reagents, sample collection, or sample
preparation. The remaining random errors are evaluated below.
Because of the wide range of concentration of lead in the
samples, the coefficient of variation is used, Equation (7).
-------
Precision (total random error) from Duplicate Sample
Analysis:
CV « 0.168 or 16.8% of sample concentration.
The component random errors, summed as per Equation (5), are:
»x - (0.00792 + 0.052 + 0.0512 + 0.0472)1/2 . 0.085.
These random errors suggest that reproducibility
errors(0.0079) are small and that extract matrix, sample matrix,
and sample homogeneity errors are equivalent. The sum of these
errors is about half the total random error so the sampling error
is essentially equal to all other errors combined.
Inter laboratory precision as calculated from the results of
triplicate analyses, using Equation (7) is:
Precision - CV - 0.128 or 12.8% of sample concentration,
(3) Uncertainty: The data for bias and precision can be
combined to yield the uncertainty for any reported concentration
by use of the following equation:
0 - (1 + B + 2 C) (8)
where B is the bias, C is the standard deviation or coefficient
of variation as appropriate, and 2 converts these to the 95
percent confidence limits. For soil analyses, using Equation (8)
and the bias and CV derived above, the 95% confidence bounds on a
reported value, x, are:
Soil result will lie between 0.676x and 1.348x Vg g'1.
It is required that the QA/QC plan ensure and document that
all data collected, whether used for research or for monitoring
purposes, is scientifically valid, defensible and of known
precision and accuracy. The described presentation of QC data,
though designed for analysis of lead in soil, can be used as a
112
-------
guide for other sampling and data analysis protocols and/or QA/QC
plans.
Presentation of QA/QC data allows readers to verify
conclusions drawn as to the reliability of the data. Such an
approach also contributes to the building of a body of QA/QC and
monitoring experimental data in the literature which allow
comparisons to be made between and among studies. Procedures
used to validate the individual data points should be presented
and where some points are discarded arguments should be presented
to support these decisions.
PRESENTATION OF RESULTS AND CONCLUSIONS
Special emphasis should be placed on how overall levels of
precision and confidence were derived from the data. Great care
must be exercised to insure that, in determining results and
conclusions, assumptions are not made which were not part of the
study design and which cannot be tested by data derived from the
study. If portions of the study results are ambiguous and
supportable conclusions cannot be drawn with regard to the total
reliability of the data, that situation must be clearly stated.
" that event it is desirable to include recommendations for
conducting an improved study in such a way as to clarify the
observed ambiguities.
QUALITY ASSURANCE ASPECTS
The adequacy of all aspects of the QA/QC plan should be
examined in detail with emphasis on defining for future studies
an appropriate minimum adequate plan. Some aspects of the plan
actually used may have been too restrictive, some may not have
been restrictive enough. Appropriate analyses and interpretation
of the data should identify the actual situation.
Future soil monitoring studies should have checks and
balances built into the QA/QC plan which will identify early in
the study whether the plan is adequate and if necessary, allow
for corrective action to be taken before the study continues.
This is one of the major advantages of conducting an exploratory
study along the lines outlined in this report. If there are
problems with the QA/QC plan, they will often be identified in
the exploratory study and be corrected before major resources are
expended.
-------
There is insufficient knowledge dealing with soil monitoring
studies to state with confidence which portions of the QA/QC plan
will be generally applicable to all soil monitoring studies and
which portions must be varied depending on site-specific factors.
As experience is gained, it may be possible to provide more
adequate guidance on this subject. In the meantime it is
recommended that the best approach is to assume that important
factors of QA/QC plans are site-specific and to conduct an
appropriate exploratory study at each new study site to verify
that various aspects of the QA/QC plan are adequate to meet
program objectives prior to proceeding with the final definitive
study.
114
-------
APPENDIX D
SYSTEM AUDITS AND TRAINING
INTRODUCTION
The material for this chapter has been obtained primarily
from USEPA Kellogg Idaho Study (1984). The first phase of an
auditing program for soil monitoring projects should be the
preparation of standard operating procedures (SOP) that identify
the methods and techniques necessary to perform all aspects of
the required audit. The SOP must be adequate to perform onsite
sampling and sample bank (where applicable) audits. The second
phase should then be the actual conduct of the required field
audit. Audits are conducted by appropriate elements of agencies
or organizations having cognizance over the monitoring project.
The frequency of auditing should be determined by the project
officer. Juran et al. (1979) state that, "the activities subject
to audit should include any that affect quality regardless of the
internal organizational location."
A system audit is an overall evaluation of a project to:
o Verify that sampling methodology is being performed in
accordance with program requirements
o Check on the use of appropriate QA/QC measures
o Check methods of sample handling, i.e., packaging,
labeling, preserving, transporting, and archiving in
accordance with progam requirements
o . Identify any existing quality problems
o Check program documentation, i.e., records (site
description, chain-of-custody collection and analytical
tags, field and sample bank log books and field work
sheets)
o Initiate corrective action if a problem is identified
115
-------
o Assess personnel experience and qualifications if
required
o Follow-up on any corrective action previously
implemented
o Provide onsite debriefings for sampling team and sample
bank personnel.
o Provide a written evaluation of the sampling and sample
bank program
The purpose of the system audit is to ensure that the QA/QC
system planned for the project is in place and functioning well.
The auditor first must review Work Plans, Protocols, Test
Plans, QA/QC Project Plan, and all Program Reports. A discussion
of the current status of the project, and the identity of any
problems encountered, with the project officer is suggested
before conducting the onsite sampling audit. Sample
chain-of-custody procedures and raw data are checked as
appropriate and results of blind QC samples routinely inserted in
the sample load by sample bank personnel are reviewed.
Spot-checks of sampling methods and techniques, sampling and
analysis calculations, and data transcription are performed.
SAMPLE BANK AUDIT
The primary objective is to determine the status of all
Sample Bank documentation and archived samples. Emphasis is
placed on:
o Verifying that the documentation is in order and
sufficient to establish the disposition of any sample
collected
o Determining any discrepancies that currently exist and
initiating corrective action as appropriate
o Verifying that the recording of QA/QC measures (blanks,
duplicate spikes, blinds) is in accordance with the
QA/QC Plan
o Establishing procedures for final disposition and
mechanics of transfer of all Sample Bank holdings upon
termination of the operation.
116
-------
The first step of the audit is to inventory the Sample Bank
records and archived samples. The records that must be inspected
are:
o Chain-of-custody forms
Field forms
Analysis forms
o Sample tags
Field tags
Analysis tags
o Analysis forms
Individual samples
Batch sheets
o Shipment forms
o Logbooks
Soils
Daily log
The operational procedures inspected should include:
o Preparation Procedures (sample bank or analytical
laboratory)
Drying (if used)
Sieving
Mixing
Packaging
Shipping
o Housekeeping
Safety
Decontamination
Evaluation of Swipe Samples
o Security
- Forms (documents)
Samples
o Storage
Sampling equipment
Archived samples
Check that required documentation has been maintained
in an orderly fashion, that each of the recorded items is
properly categorized, and cross-checking can be easily performed.
In addition, ensure that data recording conforms to strict
document control protocols and the program's QA/QC Plan.
117
-------
The archived samples inspected can be categorized as
follows:
o Soil
o Blanks
o Splits
o Standard Reference Materials (SRM)
o Non-Soil Materials Collected with the Soil Sample
Conduct an audit of the archived samples. Verify that
appropriate samples exist for each entry in the logbook. Field
sample tags should be replaced by the appropriate analytical
tags/ and chain-of-custody forms are prepared in order to
transfer the samples. Detailed sample bank procedures are
presented by USEPA Dallas Lead Study (1984).
DAILY LOG
Check for clear, concise entries detailing events of the day
(such as numbers of samples processed), problems encountered, and
actions taken to solve them. This log can provide excellent
documentation of the operation of the Sample Bank.
SAMPLE BANK LOGS
Review these logs for complete sample information entered.
Changes made should be by crossing out so the original entry is
still visible, and initialing. In addition checks for the
identification and documentation of split and duplicate samples,
and field and Sample Bank blanks must be performed.
SAMPLE COLLECTION AUDITS
It is recommended that an audit of the overall QA/QC plan
for sample documentation, collection, preparation, storage, and
transfer procedures be performed just before sampling starts.
The intent of this audit is to critically review the entire
sampling operation to determine the need for any corrective
action early in the program. Additional total program or partial
audits can be conducted at various times throughout the sampling
program.
It is recommended that the Project Officer maintain a QA/QC
Coordinator onsite during cample collection to monitor the
sampling team's activities, provide technical and corrective
118
-------
action suggestions to the sampling teams, and supplement
performance audits on sampling as needed.
FIELD AUDITS
The primary objective is to determine the status of sampling
operations. Emphasis is placed on:
o Verifying that operational aspects and procedures are in
accordance with the protocols and QA/QC plan.
o Verifying the collection of all samples including
duplicates and field blanks.
o Verifying that documentation is in order and sufficient
to establish the collection location of any sample
collected.
o Determining discrepancies that exist and initiating
corrective action as appropriate.
>..
o Collecting independent samples.
*
The on-site field audit is to inspect sample records and
equipment. Records inspected include:
'a. Chain-of-Custody Forms
b. Sample Tags
c. Site Description Forms
d. Log Books
The operational procedures inspected should include:
o Sampling Procedures
Equipment
. - Techniques
- Decontamination
Collection of duplicate and field blank samples
Security
Sample storage and transportation
- Containers
Contaminated waste storage and disposal
Site Description Form entries
119
-------
DATA MANAGEMENT AUDITS
An audit of the data management system by tracing the flow
of specific samples through the system should be performed. In
particular, the ability of the system to correctly identify a
sample from any one of its identification numbers should be
checked.
Entries in the sample bank's logbook will be the basis for
these performance checks. From time to time, erroneous input
information may be used to audit the system.
TRAINING
The project leader of a soil monitoring project is
responsible for ascertaining that all members of his project team
have adequate training and experience to carry out satisfactorily
their assigned missions and functions. Until a field sampling
team has worked together long enough for the project leader to
have verified this from first hand knowledge it is good practice,
in addition to any classroom training or experience, to conduct
comprehensive briefing sessions for all involved parties during
which all aspects of the sampling protocol, including the QA/QC
plan, are presented and discussed in some detail. This approach
will help the project personnel to develop into a team where each
team member knows his own job well and knows how it fits into the
overall team effort. Sufficient field training exercises should
follow the briefing sessions until each team member can
demonstrate successfully that he can perform his job routinely
well and without delay. Of course, on subsequent projects of the
same general type with the same team, the training exercises may
be reduced in number or dispensed with as deemed appropriate by
the project leader.
In summary, the sampling effort must include classroom and
field training programs that have provided detailed instruction
and practical experience to personnel in sample collection
techniques and procedures, labeling, preservation, documentation,
transport, and sample bank operational procedures. Also, special
training programs concerning procedures and program documentation
should be completed by all personnel prior to their involvement
in the conduction of any audits.
120
------- |