PB83-206979
Preparation of Soil Sampling Protocol: Techniques and Strategies
Environmental Research Center
Las Vegas, NV
May 83
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
NTIS
-------
EPA-600/4-83-020
May 1983
FB33-2CC979
PREPARATION OF SOIL SAMPLING PROTOCOL:
TECHNIQUES AND STRATEGIES
by
Benjamin J. Mason
ETHURA
McLean, Virginia 22101
under subcontract
to
Environmental Research Center
University of Nevada, Las Vegas
Las Vegas, Nevada 89154
Cooperative Agreement Number: CR808529-Q1-2
Project Officer
Robert D. Schonbrod
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
Environmental Monitoring Systems Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Las Vegas, Nevada 89114
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
LPA-bOO/4-83-020
2.
3 RECIPIENT'S ACCESSION NO.
' * •» * A C 7
4 TITLE AND SUBTITLE
PREPARATION OF SOIL SAMPLING PROTOCOL: TECHNIQUES
AND STRATEGIES
5 REPORT DATE
May 1983
6. PERFORMING ORGANIZATION CODE
7. AUTHORis) Dr. Benjamin J. Mason
Ethura
McLean. VA 22101
8 PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Research Center
University of Nevada, Las Vegas
4505 S. Maryland Parkway
Las Vegas, NV 89154
10. PROGRAM ELEMENT NO.
CCRL1A
11 CONTRACT/GRANT NO.
CR808529-01-2
12 SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
P.O. Box 15027
Las Vegas, NV 89114
13 TYPE OF REPORT AND PERIOD COVERED
Final
14 SPONSORING AGENCY CODE
EPA/600/07
is SUPPLEMENTARY NOTES
as a product of the Exposure Assessment Research Project
16 ABSTRACT
This report sets out a system for developing soil sampling protocols that
can be used to meet the needs of the environmental scientist working under a
number of situations. The body of the report discusses the factors that
influence the selection of a particular sampling design and the use of a
particular sampling method. Statistical designs are discussed along with the
appropriate analysis of the data. Three appendices are included. One is a
djscussion of the steps that must be taken to arrive at the desirod protocol.
The remaining appendices present two examples of protocols; one fox a shallow
spill situation and the second for a deep contamination plume. A technique
called kriging is presented as an approach for handling the analysis of data
collected during a soil sampling program. This technique allows the researcher
to develop maps of the pollution levels and to assign a statistical precision
to the data at each point on the map. The data presented on these two maps can
be evaluated to identify areas where additional samples are needed to reach a
desired level of precision for the area covered by the maps.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATi Field/Group
Sampling, Investigations, Soil Surveys,
Pollutants, Design Criteria, Statistical
Analysis.
monitoring network
design, geostatistical
methods
1406
1407
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (Tins Report)
N/A
20 SECURITY CLASS (Tim page)
U/A
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
NOTICS
Tne information in this document has been funded wholly or
in part by the United States Environmental Protection Agency
under assistance agreement number CR 808529-01-2 to the
Environmental Research Center of the University of Nevada, Las
Vegas, it has been subject to the Agency's peer and
administrative review, and it has been approved Eor publication.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
11
-------
ABSTRACT
This report sets out a system for developing soil sampling
protocols that can be used to meet the needs of the environmental
scientist working under a number of situations. The body of the
report discusses the factors that influence the selection of a
particular sampling design and the use of a particular sampling
method. Statistical designs are discussed along with the
appropriate analysis of the data. Three appendices are included.
One is a discussion of the steps that must be taken to arrive at
the desired protocol. The remaining appendices present two
examples of protocols; one for a shallow spill situation and the
second for a deep contamination plume. A technique called
kriging is presented as an approach for handling the analysis of
data collected during a soil sampling program. This technique
allows the researcher to develop maps of the pollution levels and
to assign a statistical precision to the data at each point on
the map. The data presented on these two maps can be evaluated
to identify areas where additional samples are needed to reach a
desired level of precision for the area covered by the maps.
This report was submitted in partial fulfillment of the
requirements under Cooperative Agreement No. CR808529-01-2 by
ETHURA under subcontract to the Environmental Research Center of
the University of Nevada, Las Vegas and under the sponsorship of
the U. S. Environmental Protection Agency.
111
-------
CONTENTS
Abstract iii
Figures viii
Tables viii
1. Introduction 1
2. The Soil System 3
3. Initiating the Soil Sampling Study 6
3.1 The Objective 6
3.2 Data Reliability 7
3.3 Resources 8
4. Types of Soil Pollution Situations 9
4.1 Large Areas Characterized by Shallow
Pollution Deposition 9
4.2 Large Areas Characterized by Deep
Pollution Deposition 10
4.3 Localized Areas of Surface Contamination. 11
4.4 Localized Areas Characterized by a
Deeply Penetrating Plume 12
5. Review of Background Data 14
5.1 Historical Data 15
5.2 Geological Data 16
5.3 Soils Information 17
5.4 Environmental Studies Data 17
5.5 Legal Cases 18
5.6 Remote Sensing 18
6. Statistical Designs 20
6.1 Background for Statistical Sampling Plans 20
6.2 Simple Random Sampling 23
6.2.1 Determination of the Number of
Samples Required 24
6.2.2 Location of Sampling Points ... 27
6.3 Stratified Random Sample 28
6.4 Systematic Sampling 29
6.5 Judgement Sampling 31
6.6 Phasing the Study 31
6.7 Control Areas 32
IV
-------
7. Sample Collection 33
7.1 Surface Sampling 33
7.1.1 Sampling with a Soil Punch .... 34
7.1.2 Ring Sampler 34
7.1.3 Scoop or Shovel Sampling 35
7.2 Shallow Subsurface Sampling 35
7.2.1 Soil Probes and Hand Augers ... 36
7.2.2 Power Augers and Core Samplers . . 37
7.2.3 Trenching 37
7.3 Sampling for Underground Plumes 38
7.3.1 Usual Procedure for Underground
Plume Sampling 38
7.3.2 Variations in the Procedures ... 41
7.4 Compositing 41
7.4.1 Estimating Sample Variance .... 42
7.4.2 Compositing with a Mixing Cloth . 42
7.4.3 Compositing with a Mixing Bowl . . 43
7.4.4 Laboratory Compositing 43
7.5 Replicate Samples 44
7.6 Miscellaneous Tools 44
7.7 Record Keeping 44
7.7.1 Log Book 45
7.7.2 Site Discription forms 45
7.7.3 Sample Tags 45
7.7.4 Chain-of-Custody Forms 45
7.8 Decontamination 46
7.8.1 Laboratory Cleanup of Sample
Containers 46
7.8.2 Field Decontamination 47
7.9 Quality Assurance 48
7.10 Safety 48
8. Data Analysis 49
8.1 Analysis of Data for Simple Random Design 50
8.1.1 Basic Parameters 50
8.1.2 Use of t-Test 51
8.1.3 Analysis of Variance of Simple
Random Design 52
8.2 Data Analysis for Stratified Random
Design 53
8.2.1 Basic Parameters for the
Stratified Random Design .... 53
8.3 Data Analysis for Systematic Sampling
Designs 55
8.3.1 Kriging Analysis of Systematic
Data 56
References 62
-------
Appendices
A Soil Protocol Development Process A-l
Section A-l Purpose A-l
Section A-2 Preliminary Background Information A-2
Section A-3 Development of Protocol A-3
3.1 Background Information A-4
3.1.1 Questions to answer A-5
3.2 Develop an Objective Statement for
the Study A-5
3.2.1 What is the study attempting
to accomplish A-6
3.2.2 How close must the estimates
be to the real mean value . . A-6
3.2.3 How reliable must the answers
be A-7
3.2.4 What resources are available
for the study A-7
3.2.5 When must the study be finished A-8
3.3 Determing the Magnitude of the
Problem A-8
3.3.1 What is the known areal extent
of the contamination .... A-8
3.3.2 What is the vertical extent
of the contamination .... A-9
3.3.3 What chemicals have been
identified in the study area A-9
3.3.4 What are the concentrations . A-9
3.3.5 Are the chemicals toxic . . . A-9
3.3.6 What is the attitude of the
community toward the problem A-10
3.3.7 Place the study into one of
the four classes of studies
listed in 3.3B and go to 3.4 A-10
3.4 Selection of a Statistical Design . A-10
3.4.1 How many samples do you need . A-10
3.4.2 What is the distribution of
the sample sites- A-10
3.4.3 What is the frequency of
sampling A-ll
3.5 Select a Sampling Method A-ll
3.6 Data Analysis A-12
B Sampling Protocols for Surface Soils .... B-l
Section B-l Overview B-l
VI
-------
Section B-2 Sampling Design B-2
2.1 Minimum number of samples B-2
2.1.1 The equation B-3
2. 2 Grid Layout B-3
2.2.1 Selection of a starting point B-5
2.2.2 Grid size B-5
2.2.3 Physical location of the grid
in the field B-6
2.3 Compositing of the Samples B-6
2.4 Replicates B-7
2.5 Controls and Background Samples . . B-7
2.5.1 Location of Controls B-7
Section B-3 Sample Collection on the
Triangular Grid B-9
3.1 Sampling for Non-volatile Chemicals B-9
3.2 Sampling for Volatile Chemicals . . B-9
3.2.1 Soil Sampler B-10
3.2.2 Collection of sample B-10
3.2.3 Transport and analysis .... B-10
Section B-4 Records, Security and Safety . . B-ll
4.1 Records B-ll
4.2 Security B-ll
4.3 Safety B-ll
4.4 Site Restoration B-12
Section B-5 Data Analysis B-13
Section B-6 Staffing, Equipment and Supplies B-14
6.1 Equipment and Supplies B-14
C Sampling Protocol for Contaminant Plume . . . C-l
Section C-l Overview C-l
Section C-2 Sampling Design C-2
2.1 Minimum Number of Samples C-2
2.2 Grid Layout C-3
2.3 Control Area C-4
2.4 Preliminary Study C-4
Section C-3 Sample Collection C-5
3.1 Sampling Equipment C-5
3.2 Non-Volatile Pollutant Sampling . . C-5
3.2.1 Sample extraction C-5
3.2.2 Sample preparation C-6
3.3 Volatile Chemical Sampling C-7
3.4 Security C-7
3.5 Safety C-7
3.6 Decontamination C-8
Section C-4 Data Analysis C-9
Section C-5 Staffing, Equipment and Supplies C-10
5.1 Staffing C-10
5.2 Equipment and Supplies C-10
vn
-------
FIGURES
Number Page
7.1 Trench sampling techniques 39
8.3.1 Flow sheet for kriging 59
8.3.2 Examples of semi-variogram developed during
kriging 60
B-2.1 Triangular grid design B-4
TABLES
Number Page
8.1.1 Analysis of Variance for Simple Random Design 52
8.2.1 Analysis of Variance for Stratified Random
Design 54
B-2.1 Number of Sampling Points in a Triangular Grid B-5
Vlll
-------
SECTION 1
INTRODUCTION
Environmental assessments are designed to evaluate the
impacts of chemical pollutants upon man and his environment.
Considerable effort has been expended in developing protocols for
use in monitoring for air and water borne pollutants. The
complexities encountered in sampling the soil system have been a
major handicap in the development of field procedures. The
Office of Pesticide Programs has promulgated a document outlining
procedures for sampling for the National Soil Pesticide Survey
Program (Office of Pesticide Programs, 1976). This document has
proven to be useful in the sampling of soils on a nationwide
basis but cannot be used effectively in many situations where
hazardous wastes are encountered.
The system presented below is designed to provide the
environmental scientist with a means for developing a protocol
that will satisfy the soil sampling needs of agencies such as the
U. S. Environmental Protection Agency (EPA). The methods
presented will provide the scientist with adequate tools for
producing reliable estimates of the spatial distribution of soil
borne pollutants. The techniques described in this report have
been borrowed from a number of areas of soil science. Each
section provides information that will enable a person that is
knowledgeable in the behavior of pollutants with the necessary
tools for acquiring soils data that will meet most of the needs
encountered in environmental monitoring work.
In order to use this document, the environmental scientist
should have a familiarity with the general properties of soils
and have some idea of the behavior of pollutants in the
environment. A chemist and a statistician must be available for
consultation from the outset of any soil sampling study in order
to provide the technical input into the selection processes that
are needed in developing the protocol. The field scientist must
be given latitude in modifying the protocol to meet unusual
conditions not covered during planning; but, these judgement
calls must be made in concert with the chemist and the
statistician.
Scientists working for the EPA will need to follow
procedures in the areas of chain-of-custody and quality assurance
-------
that are more riqid than those likely to be required by a
non-regulatory agency. Sections dealing with these areas are
included in this report; but they have been made brief because
documents are available from the EPA National Enforcement
Investigations Center (NEIC) that cover these areas in
considerable detail. Chemical analytical methods are not
discussed in this report even though these methods are factors
that must be considered in selecting the appropriate sample
collection techniques.
The basic format of this report is designed to allow the
environmental scientist to arrive at the selection of a sampling
protocol by answering a series of questions found in Appendix A.
The answers to the questions will then lead to one of several
approaches that can be used in a particular study setting. The
sections of Appendix A are arranged in a chronological order
leading from initiation of the study through to the statistical
analysis of the data and point back to one of the sections in the
main body of the report. Sections 2, 3, and 4 outline some of
the basic concepts needed to plan and initiate a soil sampling
study. Section 5 sets out the types of background data that are
needed in order to properly evaluate the situation in a
particular soil sampling study area. Section 6 presents
statistical designs, Section 7 the methods for collecting samples
and Section 8 data analyses.
The sections on sample design and data analyses present an
approach that allows the scientist to handle and present the data
that has been acquired through a sampling program designed to
estimate the levels and distribution of pollutants found in the
soil environment. This approach, called kriginq, has been used
in evaluating the distribution of radioactivity at a number of
sites both in the United States and on the Enewetak Atoll. The
method was developed in France for use in ore evaluation and has
recently been used in the United States by a number of scientific
disciplines. The use of the technique is still in its infancy in
the environmental fields but appears to hold considerable promise
as a tool for evaluating pollution patterns. The major advantage
of kriging is in the abil.ity to develop estimates of
concentrations over a geographic area and also provide a measure
of the confidence limits to be placed on the data at any point.
The statistical errors at each point can be plotted on a map
which shows the isopleths for the error terms. The resulting map
allows one to identify where additional sampling sites would be
most beneficial in increasing the reliability of the data set.
-------
SECTION 2
THE SOIL SYSTEM
Soil sampling as presented in this report encompasses the
entire mass of unconsolidated mantle of weathered rock and loose
earth material lying above solid rock. This definition is that
used by the engineer rather than the agriculturalist (Soil
Science Society of America, 1965).
The physical and chemical characteristics of the soil system
influence the transformation, retention and movement of
pollutants through the soil. Clay content, organic matter
content, texture, permeability, pH, and cation exchange capacity
will influence the rate and route of migration. These factors
must be considered in the process of designing a sampling plan.
The agricultural worker considers these factors but does not
focus the sampling design on them because a farmer is interested
in how much lime or fertilizer to apply to a field and not in the
avenues of movement of that fertilizer through the soil system.
Little consideration is given to the spatial variability of a
field. Occasionally a farmer may fertilize two different soil
types at different rates if the yield gains and fertilizer cost
savings can justify the time and effort required. In such a case
the soil scientist sampling the farmer's fields may take separate
samples of each soil type. The environmental scientist on the
other hand is interested in a number of possible types of
pollution and routes of migration. This cannot be addressed with
a single sample or a single composite sample. Therefore, some
form of statistical sampling design must be used to evaluate the
pollution in a soil system. This report will present several
options for sampling soils that are available for evaluating
pollution migration.
Environmental sampling must take into consideration one of
the key characteristics of the soil system -- extreme
variability. Cline (1944) noted that even though it was common
knowledge that analytical errors are much less than sampling
errors, little attention has been given to studies that would
provide data for developing a sound sampling procedure. Cline
further noted that, "the limit of accuracy is determined by the
sample not by the analysis". The limited knowledge about soil
variability changed very little until about two decades ago.
-------
Nielsen and his associates (Nielsen et al., 1973; Warrick et
al.f 1977; and Vieira et al., 1981) have begun to study the
spatial variability of the soil system in an attempt to develop
reliable predictions of water movement through the soil. Campbell
(1979) makes note of the fact that, in 1915, J. A. Harris
discussed the effects of soil variability on the results of
experiments. Campbell has used this knowledge in evaluating
approaches to delineating soil mapping units (Campbell 1977,
1978, 1979, 1981). Rao et al. (1979) reviews other work where
spatial variability was considered.
Petersen and Calvin (1965)"also note: "Soil properties vary
not only from one location to another but also among the horizons
of a given profile. The horizon boundaries may be more distinct
than are the surface boundaries of a soil classification unit.
Here, also, however, zones of transition are found between
adjacent horizons". The magnitude of sampling errors between
layers of soil tends to be less than the magnitude of sampling
errors in a horizontal direction. Disturbed or plowed soils are
reported to be more variable than virgin soils in most cases
(Chapman and Pratt, 1961).
One measure of variation is the coefficient of variation
(CV)*. Coefficients of variation for soil parameters have been
reported ranging from as low as 1 to 2% to as high as 850%.
White and Hakonsbn (1979), for example, noted that the CV for
plutonium in the soils of a number of test sites ranged from 62%
to 840%. Mathur and Sanderson (1978) reported coefficients for
natural soil constituents (i.e., part of the soil itself) varying
from 5.6% to 75.2%. Harrison (1979) evaluated four phosphorus
properties of soil and reported CV values ranging from ll% to
144% with the highest values being for available P. Hindin et
al. (1966) reported a CV of 156% for insecticide residue
concentrations in a square block of soil that was 30 inches on a
side.
Mausbach et al. (1980) reported on a study conducted by the
Soil Conservation Service (SCS) laboratory in Lincoln, Nebraska.
Matched pairs of samples were collected from areas within a soil
series. The samples were stratified by a number of factors in
order to reduce the variability. The samples were selected from
the modal phase of the series and were collected at distances
that ranged from 2 to 32 km from the other member of the pair.
The authors note that the literature indicates that up to half of
the variability between similar soils may occur within a distance
CV = +s/y X 100
where CV = coefficient of variation in %
_s = standard deviation of sample
y = mean of sample
-------
of one meter. (Studies are underway at Lincoln to determine
variability within this one meter distance.) Mausbach et al.
(1980) reported that in their study of the variability within a
soil type the CV s for physical properties ranged from 9 to 40%
for loess, 23 to 35% for glacial drift, 33 to 47% for alluvium
and residuum, 18 to 32% for the A and B horizons, and 33 to 51%
for the C horizons. The CV's for the chemical properties tended
to be higher ranging from 12 to 50% for Alfisols, 4 to 71% for
Aridisols, 6 to 61% for Entisols, 10 to 63% for Inceptisols, 9 to
46% for Mollisols, 16 to 132% for Spodosols, 10 to 100% for
Ultisols, and 8 to 46% for Vertisols.
The variation that seems to be inherent in the data collected
from any soil sampling study must be taken into consideration
during the design of a sampling plan for whatever the purpose of
the study. Technologies designed to take the variation into
account must be employed in anv soil sampling plan. This
includes the sampling design, the collection procedures, the
analytical procedures and the data analyses.
-------
SECTION 3
INITIATING THE SOIL SAMPLING STUDY
The identification of the components needed in a specific
soil sampling protocol begins with a clear statement of the
objectives of the study and follows through a series of steps
that are required to select the sampling design, the tools to
use, the size of crew, etc. Each step is characterized by a
number of decisions that must be made before the protocol can be
finalized. A number of the questions that must be answered in
making the decisions are given in Appendix A in order to assist
the reader in making the appropriate selections. The key
components of any soils study are discussed in the sections that
follow this brief introduction to the protocol development
process.
A further assumption has been made that the scientist, his
supervisors and the administrators calling for the study are
committed to producing a quality study. This requires that a
clear statement of the objectives be made and that adequate time
is given for planning and reviewing the study. This report can
help shorten the planning time by focusing the reader into the
areas where decisions must be made; but, the commitment to
succeed must come from those responsible for the study.
3.1 The Objective
The objective statement sets forth the specific goals that
are to be met by the sampling program. This statment should be a
clear, concise definition of why the study is needed and what
questions the studv is to answer. It could be as brief as a
question such as "How much of the surface area around the
accident at the XYZ railroad siding has been contaminated with
trichloroethylene (TCE)?"; or, it could be the request for a
detailed study in support of an enforcement action that must
determine the likely avenues of contamination leading from a
hazardous waste site into a community. The important point that
is being made is that the goals of the study must be spelled out
and must be agreed upon by all parties involved with the study.
-------
As part of the objective statement, the environmental
scientist responsible for the study should attempt to obtain a
statement of the data reliability desired and the resources
committed. If this cannot be clearly given at the outset, some
initial goal should be indicated with the final goal selected
after a number of iterations during the planning process.
3.2 Data Reliability
The scientist needs to know two things before selecting the
components of the study plan — the confidence level desired and
the allowable margin of error to be met by the results.
Too often, soil sampling is done without a clear knowledge
of the level of precision that can be met by the study.
Laboratory systems have been developed to the point where
considerable confidence can be placed in the results produced by
a quality laboratory. Quality assurance is maintained throughout
the studies. The sources of uncontrolled variation are too great
for a field study to meet the precision found in a laboratory.
The only alternative is to select a level of confidence that is
acceptable and attainable within the limits of the resources
available for the study.
The reliability expressed by the confidence level states the
level of precision of the results generated by the study. Three
confidence levels are normally used by the scientific community.
These are usually expressed as + 1 standard deviation, + 1.96
standard deviations and + 2.58 standard deviations which covers
68%, 95% and 99% of the total population, respectively. Another
way to state this is to say the probability is 0.32 (or 1 in 3)
that the value is outside of one standard deviation on either
side of the mean; 0.05(or 1 in 20) that the value is outside of
1.96 standard deviations; or 0.01 (or 1 in 100) that the value is
outside of 2.58 standard deviations. Where results must be
absolute, a 99% confidence level should be used. Where resources
are limiting or reliability is not of paramount importance, the
68% confidence level may be acceptable. Environmental sampling
often attempts to attain a level of 95% confidence. The actual
level is not as important as the fact that the level is known and
agreed upon before the study is started.
Environmental studies can often be conducted in phases with
an increase in reliability attained as each phase is completed.
The design of the study should allow for this by designating the
-------
confidence level for each phase and planning the study so that
each phase produces an answer. For example, the first phase of
many soil sampling programs is a pilot or exploratory study. If
properly designed, this can produce results with a known
precision and thus avoid the problems of the "quick and dirty"
look at a soil study area. The confidence level for this phase
might be the 68% level. The second phase might be more
definitive whose study design makes use of the data generated by
the pilot study. The results are expected to reflect this
increased knowledge; therefore, the study should be designed to
acquire the data to meet this precision. The confidence level
might be increased to 95% for this phase. There may be
particular situations where it is necessary to reach a 99%
confidence level. An example where this might be required would
be a litigation case where there was a possibility that a home
had been contaminated by chemicals from an abandoned hazardous
waste landfill. The emotions and the liabiity force the
scientist to attempt to meet a higher level of reliability.
The second item listed, the margin of error, is needed in
determining the number of samples required to meet the precision
specified above. This is often expressed as a percentage error
that the scientist is willing to accept or it may be the
difference that he hopes to detect from the study. The value
designated should reflect the importance and use to be made of
the numbers. Clean-up of a chemical spill might require a less
precise estimate of pollutant levels than would a study
attempting to identify subtle variations in background levels.
In the spill case, the scientist wants to know if a soil sample
is contaminated or not. This situation could change, however, if
the spill was of a known carcinogen located next to a city's
water treatment plant. The margin of error chosen is combined
with the confidence level to derive an estimate of the number of
samples required. The smaller the margin of error; the larger
the number of samples required.
3.3 Resources
The data reliability is often modified by the time and money
available for the study. An attempt should be made to estimate
as precisely as possible the amount of resources that can be
committed to a particular study. This commitment should be
spelled out and monitored on a regular basis. The resources
needed include not only money but personnel, laboratory capacity,
equipment and time. Where time and staff are both in short
supply the reliability of the study will have to be lower in
order to meet the schedule. Planning for the study must also
include resources for the handling of paper work and
chain-of-custody along with shipping costs of samples where these
are not handled directly by the sampling crew.
8
-------
SECTION 4
TTPBS OF SOIL POLLUTION SITUATIONS
There are essentially four major types of sampling situations
that the environmental scientist is likely to encounter. These
are:
o Large area studies where pollution is in the
surface layers, e.g., in support of an ambient
monitoring effort.
o Large area studies where pollution has moved
down into the soil profile, e.g., assessing
the impact from a major industrial complex.
o Localized area studies where the pollution is
in the surface layers, e.g., sampling around a
recent hazardous chemical spill.
o Localized plume studies where the major source
of contamination is below the surface at some
depth, e.g., sampling near a leaking hazardous
waste disposal site.
The environmental scientist should attempt to determine which
of these categories exist in the area to be sampled. Factors
such as length of time that the site has been contaminated, the
type of pollutant, the type of soil and the past use of the area
must all be 'considered in determining which study category to
use. The sampling methods and the statistical designs used in
each case are slightly different. These are briefly covered in
the paragraphs presented below.
4.1 Large Areas Characterized by Shallow Pollutant
Deposition
The basic characteristic for this category of study is that
the pollutant covers a wide area and is expected to be primarily
-------
on the surface. Studies in this class are usually applicable to
surveys of ambient monitoring or monitoring downwind of a major
pollutant source that has been present for a short time span.
The pollutants may have migrated into the soil up to one to two
feet but they tend to be located only in the surface layers of
the soil. Where organic matter is present in these layers,
penetration may only be a few inches. Hand tools are usually
used to collect samples. The designs are usually either the
stratified random or the systematic grid designs. In cases where
there has been penetration and where a large number of samples
are to be collected, the scientist may desire to use a power
driven sampler. In this latter case, the samples may be
sectioned into several layers located at depth below the surface.
This will enable the scientist to evaluate the movement of the
pollutant through the soil mass if this is desirable.
The fact that the pollution covers a large area, encompassing
a number of soil types and topographic features, makes the use of
stratified sampling desirable. The strata can be identified by
soil type, by position on a slope, by parent material and in some
cases by aspect and vegetation type. Compositing is often used
to further reduce the variability and to reduce the cost. The
large area covered by the study creates a large number of samples
thus cost reduction techniques are important. The National Soils
Pesticide Monitoring Program of the Office of Pesticide Programs
in EPA is an example of a study conducted at this scale (OPP,
1976).
4.2 Large Areas Characterized by Deep Pollutant Deposition
These study situations are similar in character to those
presented in section 4.1 with the exception of the depth of
pollutant migration. The sources are such that one would expect
the chemical pollutants to move down into the soil profile to a
considerable depth. Sources might be major industrial complexes
that existed prior to implementation of the Clean Air Act, major
agricultural areas where pesticide useage has been in effect for
a number of years, world wide nuclear fallout, lead contamination
from automobile exhaust or benzo-a-pyrene in an urban area. The
pollutants would be those that are not considered to be reactive
with the surface soil constituents and would likely be soluble in
water to some extent. Metals that are known to chelate with soil
organic matter may fall into this class in some cases.
Sampling must be done with power equipment but the depths of
sampling are usually not as deep as those presented in Section
4.4. Composite samples are not recommended if they can be
avoided because of the information that is lost during the
mixing. The migration of the chemical through the soil is
usually one of the factors of primary concern in these
s ituations.
10
-------
The statistical designs are somewhat similar to those used
in the cases discussed in Section 4.1. Stratification is
desirable in most cases. Systematic grid sampling can prove to
be effective if local variation is not of interest. Kriging of
the results usually does not work too well because the range
identified by the variogram is usually around one to two
kilometers thus the intensity of sampling required to use this
technique requires too many samples to be cost effective. If
local variation is of interest, the environmental scientist may
desire to identify specific local areas where information is
needed and conduct a study on a smaller scale in those areas.
This allows the intensity of sampling to be increased and still
not run up the costs of the study.
An example of a study that is similar to the last situation
would be the soil monitoring done in support of the clean-up of
Plutonium contaminated soils found on Enewetak Atoll (Barnes
1978).
4.3 Localized Areas of Surface Contamination
This category of sampling is probably one of the most likely
to occur in the future. Spills that result from industrial or
transportation accidents, fires, or unexpected leaks from storage
containers usually pollute the soils in local areas near the
source. The primary purpose of the sampling efforts are to map
the extent of pollution and to determine the effectiveness of the
clean-up operations. In some cases the polluted areas can be
identified by color or other observable indicators. If the
Emergency Response Team is able to be on-site within a short time
after the accident or spill, sampling may be very rapid and can
be done by either grid, simple random sample or, in some
specialized cases, stratified sampling. Judgement sampling may
be permitted in some cases but this method should not be used if
follow-up monitoring is to be done. A simple random or
stratified sampling plan can be done with a few samples and
provide a basis for measuring the degree of clean-up that has
been accomplished; thus, nothing can really be gained by using
judgement sampling.
The samples are usually collected with a King tube sampler
and can be composited in some cases. The fact that the
pollutants found at spill sites are limited to a few identified
chemicals greatly reduces the costs of analysis. Therefore,
individual samples can be collected and analyzed without
prohibitive costs. The scientist may desire to collect and
composite a number of individual samples at each sampling
location, thus reducing the variation in the results of the study.
The use of individual samples allows the investigation
11
-------
teams to make more definitive conclusions about the pollution
deposition but this is often not justified in these limited study
areas.
4.4 Localized Areas Characterized by a Deeply Penetrating
Plume
The EPA is currently involved in a number of studies where
groundwater plumes have migrated from landfill sites or from
leaking storage tanks. These plumes tend to follow the
groundwater flow and tend to occur at considerable depth below
the surface. Sampling is often limited to the collection of
groundwater samples rather than soil samples because of the
relative ease in sample collection and analysis compared to the
sampling and analysis of soils. The results obtained from water
samples tend to be more uniform than data from soil samples.
The location of the pollutant plume will probably require
some form of phased approach that will be done in conjunction
with the geologists. Sampling requires heavy drilling equipment
and frequently is limited in scope due to the costs involved with
obtaining the samples. Split spoons are usually used to obtain a
core of the soil from depth and provide a means of obtaining a
relatively undisturbed sample of the soil for use in both
physical and chemical analyses. The cores obtained by this
method are easily viewed at the time the sample is removed from
the split spoon. Records or logs of the strata observed can be
used for selecting samples for analysis and for interpretation of
the results obtained.
The statistics for this type of sampling is complicated by
the fact that depth plays a prominent role in the analysis of the
results of the sampling. Regression may be required along with
stratification with depth if any meaningful interpretation is to
be made from the results. Frequently in deep sample collection
situations a layer with high permeability will be encountered.
Some flexibility must be included in the design to allow these
samples to be collected if they are encountered. A phased study
approach will allow these areas to be identified and
incorporated into the final sampling phases.
Mapping of the contaminant plume becomes quite important in
many enforcement cases. This fact should encourage the use of
some form of systematic sampling array. Kriging can be used
effectively in this kind of situation. The mathematics required
for three dimensional kriging analysis may preclude the use of
this technique for other than selected layers through the plume.
12
-------
Sample replication, quality assurance, and decontamination
all become more difficult in these cases because the study is
actually being conducted below ground beyond the sight of the
researcher. Cost may force the researcher to compromise on the
precision of the numbers obtained. The fact that a number of
samples may be collected from each hole drilled also compounds
the cost and complexity of the operation. Careful planning,
however, can overcome many of these problems.
It should be apparent that these four categories of studies
do not cover all possible situations. There are many instances
where a hybrid between two or more of them may occur. It is
believed that the use of these four categories will allow the
environmental scientist to adapt to meet the situations
encountered in the field. The field researcher should be aware
that modifications may have to be made in the plan to meet the
unusual situations that often occur in field sampling. He should,
however, not make changes without carefully planning the changes
and consulting with the chemist and the statistician before
altering the designs and techniques used.
13
-------
SECTION 5
REVIEW OF BACKGROUND DATA
Any monitoring effort requires a familiarity with the study
area. Too little time is usually spent in preliminary data
collection, evaluation and planning. It is difficult, if not
impossible, to undertake a reliable soils study without review of
existing data. The sources presented below should be evaluated
and studied prior to developing a plan. The areas below are
presented in order to draw the scientist's focus onto those types
of data that will reveal the potential location of pollutants and
help evaluate their migration through the environment. Combined
with site visits and interviews with local citizens a good grasp
of the situation can be gained.
Libraries, museums, governmental agencies, public agencies,
data bases and researchers are all sources of information that
can be accumulated prior to finalizing the study plan. Often the
local citizens can provide information that is not available in
any of the normal research channels. The environmental scientist
working on abandoned hazardous waste sites will find that often
the public citizen is one of the most useful sources of
unpublished data. They have often lived in the area and are
familiar with the operation of the site and may even provide
insight into the types of chemicals and the methods of disposal
at the site. The scientist working in these cases must become a
detective. Any piece of information that will help determine how
and where the pollutants may migrate is useful in planning the
study. Each piece of information must be sifted and evaluated in
an attempt to determine how the soil system responds to such
factors as flooding, movement and use.
The following listing of information is only partial and
reflects the author's own experience. Each researcher should be
able to use this listing as a starting point from which to
develop the needed data for other studies.
14
-------
5.1 Historical Data
The scientist should attempt to collect all available
documents dealing with the study area including newspaper
accounts, if time permits. The more informed the investigator
is, the better his grasp of the situation. The result should be
a knowledgeable study that addresses the pollutant problem in the
context of the soil system in the study area. Historical data
can help answer questions about the sources of pollution, routes
of migration, uses of the area, or any data that will aid in
designing a study that will acquire the necessary data. The kinds
of information will vary with the site; but, in general, they
deal with the history of use of the area, historical drainage
patterns, groundwater flow and use, and environmental and health
problems associated with the study area.
Wildlife biologists and other conservation workers familiar
with the natural environment in the study area along with
hunters, conservation groups and scout groups can prove to be
valuable sources of information about the wildlife and vegetation
changes that can reflect the impacts of pollution in the area.
Stream gauging station operators, boating clubs and
sportsmen are valuable sources of information about the possible
routes of migration for groundwater and pollution. Often they
have noted changes in color, sediment loading, algal blooms,
etc., that indicate chemicals are entering the streams in the
area. This becomes especially important when abandoned hazardous
waste sites are the source of pollution.
Local authorities, such as fire, police, health,
engineering, highway and maintenance departments, tax
departments, forestry and conservation workers can all provide
valuable information on prior land use. Where spills have
occurred the local fire department are often able to provide
information on the movement of the spilled materials. This is
especialy important if they have used any particular
countermeasures on unusually toxic chemicals.
The U. S. Soil Conservation Service (SCS) along with
the Cooperative Extension Service and the Agricultural
Stabilization and Conservation Service (ASCS) have frequent
contact with the local community and are often in the rural areas.
They are interested in the soil system and are usually qualified
to assist in obtaining the kinds of information that are needed
about not only the history of the area but also the presence and
effects of pollution. The staff of all three of these groups can
usually identify the local historians in the community. SCS and
ASCS both maintain files of aerial photographs of the area.
These files often go back for a number of years and can give
information on the uses of the area along with changes in soil
character with time.
15
-------
Basically the environmental scientist is attempting to
reconstruct the situation with time. An attempt must be made to
determine where the pollution came from, how long it has been
present, where it has gone in the past, and'what effects it
had. Any information that will aid in answering these types of
questions will assist in developing a meaningful study plan.
5.2 Geological Data
The geological character of the area is important not only
for determining the routes of migration of soil pollutants but
also as a factor in any attempt to stratify the area into
homogenous soil types. Parent materials and bedrock can often
play an important part in determining how the pollutants will
react in the soil.
The U.S. Geological Survey (USGS), the Corps of Engineers and
the Bureau of Reclamation all maintain information on stream
conditions and stream flow. These agencies are valuable sources
of data about the history of the stream channels, about dredging
of channels in the streams and about flooding. These factors may
play an important part in determining the rate and route of
pollution migration. Groups such as the Tennessee Valley
Authority, the Colorado River Commission and the Great Lakes
Basin Commission have environmental scientists on their staffs
that are often able to provide insight into the environmental
setting of the streams and lakes in the area.
The USGS has produced many reports on the geology of parts of
the U. S. Their staffs are knowledgeable on rock formations,
drainage, groundwater flow and quality, and can provide maps and
remote sensing data in many cases. The USGS field geologists
often work closely with the various state agencies that cover
areas such as mining, groundwater, construction and environmental
geology. These scientists are usually familiar with the settings
where studies are to be conducted; in fact, they have often been
the first persons contacted when a problem with groundwater has
occurred.
Any information that will tell the scientist about the nature
of the bedrock, the groundwater elevations, the direction of
groundwater flow and the sources of recharge to the aquifer
should be acquired prior to developing the final study plan.
16
-------
5.3 Soils Information
As was mentioned above, the SCS, the ASCS and the Cooperative
Extension Service (County Agent) are three of the best sources of
information on soils in an area and should be the first point of
contact before any other soils data searches are undertaken. The
state offices and the local offices of the SCS maintain
information on the status of the agricultural system in the areas
under their responsibility. The SCS soils reports are a good
place to develop a familiarity with the soil types in the study
area.
Most states maintain an agricultural school that is closely
aligned with the U. S. Department of Agriculture's various
offices. The Soils Departments of the Land Grant Universities
are in close contact with SCS and are often closely involved in
agricultural soils analysis work. Their files often contain
valuable information on the nature of the soils in an area and
they often know of problems that have surfaced in the past. Some
of the universities have maintained samples of soils from past
studies. These can, on occasion, provide a valuable insight into
past pollution levels if the samples have been property
maintained.
Any data that will assist in determining the soil properties,
chemical composition, amount of organic matter, rates of
percolation into the soil, crop history, type and amounts of
clay, drainage patterns within the soil and spatial variability
in the study area can be a valuable asset when time comes to
interpret the results of the study as well as during the planning
phases of the study.
5.4 Environmental Studies
Other scientists often are interested in the same areas where
the environmental scientist is attempting to determine the levels
of soil pollution. These studies often provide valuable insight
into the problem, the system and are possible sources of
information. Frequently the geologist working on a groundwater
problem will have information on pollutant migration and soil
properties that can prove to be valuable. The well driller's log
books kept when exploratory borings are made for construction of
highways can be used to augment the data collected by the soil
investigator.
Universities in the area frequently have accumulated data as
part of thesis projects and other research studies that can be
used to increase the understanding of the soil system.
17
-------
Where the EPA or one of its state counterparts has been
investigating a particular pollution incident, the data
accumulated by them along with any analyses should be consulted
prior to undertaking the study. This search for data at the
state level should include both the environmental agencies and
the health agencies. Where the data is archived will depend upon
the state involved. Each has a slightly different organizational
structure.
Environmental impact statements are a gold mine of
information that can save considerable time for the field
scientist. Studies where highways and canals, etc., have been
the subject of the EIS can greatly increase the information
available for planning with little cost involved on the part of
the researcher.
The investigator is attempting to find information on the
pollutants, routes of migration, and effects of that migration.
Therefore, any environmental study that has been undertaken in
the past can provide the keys to preparing a viable study plan.
5.5 Legal Cases
Where legal action is pending at a particular location, data
often is available through the various enforcement channels.
This type of information is sensitive and often difficult to use
due to chain-of-custody and confidentiality. Frequently
government agencies will share data with each other under normal
conditions but when court action is involved or possible, data is
difficult to obtain and even more difficult to use in an open
forum.
Where a case is closed, considerable data may be available
in the various enforcement offices and in the court proceedings.
This is available and can usually be obtained if the need exists.
The time involved can be extensive but the data may well be worth
the effort if the soils study being planned has the potential for
creating controversy or of being used in litigation.
5.6 Remote Sensing
Imagery obtained from either aircraft or satellite can prove
to be valuable in determining the impacts of pollutants and in
identifying routes and effects of migration. Old landfill sites
can often be identified from archived aerial photography which
are perhaps one of the best historical records available. The
18
-------
USEPA's Environmental Monitoring Systems Laboratory in Las Vegas,
Nevada (EMSL-LV) is the best resource available for pollution-
oriented imagery. They are knowledgeable about sources of
existing imagery and also can assist in obtaining new imagery.
Photographs taken in conjunction with accidents or chemical
spills are a valuable resource for determining the areas where
samples should be taken.
The following sources can often provide information on
available imagery.
Agricultural Stabilization and Conservation Service
Bureau of Reclamation
Colorado River Commission
EROS Data Center in Sioux Falls, SD
National Aeronautics and Space Administration
National Archives and Record Service
National Oceanic and Atmospheric Administration
National Park Service
National Weather Service
Tennessee Valley Authority
U.S. Air Force
U.S. Army Corps of Engineers
U.S. Army Map Service
U.S. Coast and Geodetic Survey
U.S. Commodity Stabilization Service
U.S. Forest Service
U.S. Geological Survey
U.S. Soil Conservation Service
19
-------
SECTION 6
STATISTICAL DESIGNS
6.1 Background for Statistical Sampling Plans
This section outlines the basic statistical designs that are
available for use in soil sampling. The procedure for selecting
the appropriate design is covered in Appendix A.
The purpose of any soil sampling program is to obtain
information about, or a constituent of the soil. The information
obtained from the study should be representative of the soil
system in the study area if it is to be useful to the scientific
community. Much of the data collected for soil systems in the
past have been based upon samples collected according to sampling
plans designed for agricultural systems. These were patterned
after reports such as Cline (1944), the Soil Conservation
Service's Soil Survey Manual (1951) or monographs similar to the
USDA's Handbook #60 - Diagnosis and Improvement of Saline and
Alkali Soils (Richards, 1969). These resources provide guidance
on soil sampling but the approaches often provided are not
adequate for studies dealing with soil pollution.
Soil pollution studies require that sampling results provide
input into decisions that often have profound health and economic
consequences. The environmental scientist desires to determine
an average concentration, a measure of data reliability, the
direction of movement and the location of any "hot spots" that
are likely to create an undue hazard to the public or the
environment. The sampling designs used must provide this type of
information with maximum reliability and minimum cost. Limited
laboratory capacity for conducting sophisticated analyses such as
those for pesticides, priority pollutants and TCDD, the potential
for litigation, and public awareness further force the
environmental scientist into a detailed planning mode. When the
time and expense invested in analysis, data handling, and
reporting are considered, it makes little sense to invest
resources in a study that lacks the planning needed to produce
reliable results. Statistical designs must be incorporated into
20
-------
the plan at the outset; thus, the statistician and the
environmental scientist must work together throughout the study.
This facilitates the design of data forms, analysis of the data
and interpretation of results when the study is completed.
The need for a valid statistical plan cannot be over
emphasized. It is essential to know the expected variability and
confidence limits of both the analytical methods used and the
sampling designs employed. The sampling designs must take the
natural variability of the soil system into consideration. Too
often there has been a tendency to do a "quick and dirty" study
with no design. A few grab samples taken at some point of
interest may provide some information, but, more often than not,
the results of these studies eventually come home to haunt the
scientist.
The following section outlines the types of designs that can
be used in soils work. Four basic sampling approaches are
presented below -- simple random, stratified random, systematic
and judgement sampling. Judgement sampling is included in this
discussion in order to complete the list of options; however, it
is not recommended as a viable approach in most pollution control
work.
In this report the term sample is used to describe the
individual sample of soil collected at a specific sampling site
or location. The sampling site is that location within the study
area that is chosen by some random procedure to be the location
from which to collect a particular sample of soil. One can look
upon the soils in the study area as an assemblage of all possible
samples that might be collected from the area. Sampling theory
is based upon the selection of some subset of the total number
(N) of samples by a random selection process. The object of the
sampling effort is to collect a prescribed number (n) of the
individual samples at randomly selected locations. The number of
samples needed to estimate the pollution level with a prescribed
precision will depend upon the magnitude of the variation within
the soil system. In a relatively homogeneous soil, a small
number of samples may be adequate to satisfy the information
needs of the scientist. A greater variation will require more
samples to reach the same level of precision.
One technique that can help reduce the effects of the
variation upon the statistical analysis of the data is to divide
the sampling area into smaller, more homogeneous sub-areas called
strata. These strata are defined by some identifiable boundry
that is based upon topography, soil chemical or physical
properties, or some stratigraphic feature. The identification of
the strata in an area required a pilot study or prior knowledge
of the area if it is to be effective. The soil population to be
sampled should be subdivided into sampling units that are as
homogeneous as possible. The different sources of variation
within the population should be sampled if valid inferences are
21
-------
to be made about pollution patterns found in the study area
(Petersen and Calvin, 1965). This division into relatively
homogeneous sub-areas allows the statistician to remove a portion
of the variation and thus reduce the statistical error term in
the statistical tests. Stratification often allows a study to be
conducted with fewer samples or allows the study to reach
conclusions with a higher statistical precision. Generally
speaking, the more stratification of the area, the greater the
increase in precision. Petersen and Calvin (1965) noted,
however, that "The precision increases at a decreasing rate as
the strata are divided more and more until a point is reached
where no further gain in precision is obtained."
Environmental pollution behavior often is difficult to
understand without some means of graphic display of the spatial
relationship of the data. Maps have provided a useful means of
viewing and grasping the data collected in soil sampling studies.
Most mapping techniques use some form of data grid to plot the
analytical results. The use of the grid, which is a systematic
sampling design, has evolved as a result of a desire to provide
sampling coverage for portions of the entire study area rather
than at only certain randomly selected points. Stratification by
soil types offers some improvement over a simple random sample
(see discussions of these designs presented below) but the grid
provides the most uniform coverage of the study area. (Actually
the grid pattern is nothing more than a systematic or uniform
stratification of the area into blocks or sub-areas.) The
variance obtained by systematic sampling is often less than that
derived from simple random sampling. Where plumes of pollution
are expected, this approach appears to be the only reliable
method for locating the plume and measuring concentrations in the
plume. Another advantage of the systematic sample grid is that
the data can be easily mapped by most computer plotting routines.
The technique called kriging is most effective when used in
conjunction with a systematic sampling plan (see below for
discussion on kriging) .
The stratified random sample plan and the systematic sample
plan can be considered to be refinements in statistical designs
whose purpose is to make the survey more efficient. This
efficiency may result from either obtaining a smaller sampling
error with the same number of samples or from reducing the number
of sample units required to produce a specified sampling error.
Where the scientist knows little about the area to be sampled, a
preliminary study may be required. This preliminary study should
be either a simple random design or a systematic design with a
coarse grid, if kriging is to be used, a transect of sufficient
length is necessary in order to conduct the calculations.
22
-------
Compositing of a number of subsamples is another technique
that is often used to reduce the effects of variation and thus
increase the precision of the numbers obtained during the
sampling study. One of the major advantages of compositing is
the gain in precision obtained at no increase in analytical cost.
Frequently soil scientists will collect a larqe number of samples
from a farmer's field. These are then mixed in the laboratory and
an analysis performed on the composited sample. If a number of
subsamples are analyzed from the composite sample, the range of
the value obtained decreases in proportion to the square root of
the number of sampling units contributing to the composite sample
(Cline, 1944).
The problem with compositing of samples is the fact that only
an unbiased estimate of the mean is obtained. Additional data on
individual samples must be collected to augment the composite
sample data. If any statistic other than the mean is required, a
single composite sample is completely inadequate (Cline, 1944).
Compositing to reduce costs assumes that the soil is
homogeneous and therefore the number of analysis required can be
reduced. Compositing done to reduce the variability in the data
acknowledges that the variability is present but chooses to
overcome this by smoothing the effects of the variation.
Pollution studies often use composite samples in order to reduce
costs but in the process the very data desired is lost. The
environmental scientist is looking for the presence of chemicals.
If a small area of contaminated soil is composited with a large
volume of uncontaminated soil the resulting analysis often is
below the minimum detection limit of the analytical methods;
thus, valuable data is lost and an erroneous conclusion may be
reached.
With this basic background on the statistical concepts used
in soil sampling in mind, each of the statistical designs are
discussed in the following subsections. The first two
subsections are based primarily on Peter sen and Calvin (1965) and
Cochran (1965) while the remaining sections are a compendium of
information from a number of the sources listed in the Reference
Section.
6.2 Simple Random Sampling
A random sample is any sample in which the probabilities of
selection are known. Random samples are selected by some method
that uses chance as the determining factor for selection. The
chance mechanism used may range from a simple "toss of the coin"
to the use of a random number table. The choice can be one of
convenience as long as chance is the means used to make the final
sample selection. The chance of selection of any individual in
23
-------
the population can be calculated using the laws of probability.
The random sample by definition is free of selection bias.
Simple random sampling is a limiting case of random sampling.
In simple random sampling of soils, the chances of selection of
any particular segment of the soil system must be the same, in
other words each member of the soil population must have an egual
probability for selection. If a two inch core sampler is used to
sample the soil, the total number of possible samples is
determined by dividing the total area of the study boundaries by
the cross sectional area of the soil core. For example a one
square mile area would contain approximately 1,278,000,000
individual soil samples (640 X 43,560 X 144) / (1 X 1 X n ).
Simple random sampling is the basis for all probability
sampling techniques used in soil sampling and serves as a
reference point from which modifications to increase the
efficiency of sampling are evaluated. Simple random sampling in
itself may not give the desired precision because of the large
statistical variations encountered in soil sampling; therefore,
one of the other designs may be more useful. Where there is a
lack of information about the area to be studied or about the
pollution distribution, the simple random sampling design is the
only design other than the systematic grid that can be used
effectively.
In soil sampling, the unit of soil taken from the area is
usually a volume of soil, i.e., a core, a cube of soil or a
shovel full of soil. Occasionally there is a need to determine
the deposition of a particular pollutant on a per unit area
basis. In this case a known area of soil is collected for the
sample. This has been done with radioactive fallout in the past.
Results of soil sampling programs are usually expressed in either
a per unit volume, per unit area, or per unit weight. The bulk
density of the soil is the common denominator for all three of
these units. If conversions between these units are planned,
several measurements of bulk density should be made.
6.2.1 Determination of the Number of Samples Required
The procedures used in this section for determining the
number of samples required to meet a predetermined precision is
the basis for the allocation of samples to a strata in the
stratified random design and can be used to determine the number
of sample points required in the systematic sample design;
therefore, the information in this section will only be presented
once in this report.
24
-------
The number of samples required to obtain a given precision
with a specific confidence level can be obtained from equation
6.2.1 if some measure of the variance can be obtained from either
a preliminary experiment, a pilot study, or from the literature.
References such as Beckett and Webster (1971) and Mausbach et al.
(1980) can be used as a first approximation of the variance of
soil samples. This can be used to develop the initial estimate
of the number of samples needed. A preliminary study will then
further refine this number once an estimate for the variance of
the soils in the specific area is known.
n ' £*2/*2 (6.2.1)
Where D is the precision given in the specifications of the
study; s2 is the sample variance and t is the two-tailed t-value
obtained from the standard statistical tables at the a level of
significance and (n-1) degrees of freedom. D is usually
expressed as + a specified number of concentration units (i.e. +
5.00 ppm) . Equation 6.2.1 can also be written in terms of the
coefficient of variation (CV). This conversion yields Equation
6.2.2.
n = (CV)2ta2 /p2 (6<2>2)
where n = number of samples
CV = coefficient of variation
v = mean of the samples
p = allowable margin_of error expressed
as a percent (D/y).
t = the two-tailed t value obtained
from standard statistical tables at
the a level of significance and
at (n-1) degrees of freedom.
Since the t-value is dependent upon the number of degrees of
freedom, it is necessary to use an iterative approach to arrive
at the number of samples to use. Curves can be prepared that
plot the number of samples against the coefficient of variation
and thus avoid the bother of the iterations.
Use of this equation assumes that the population is
normally distributed and that less than 10% of all possible
samples in the study area are being collected. The latter
criteria is seldom exceeded in soils sampling. (In those very
limited situations where this may be the case, the finite
population correction must be applied. This correction, (N-n)/N,
is multiplied by the variance obtained from the sampling
experiment.)
25
-------
The environmental scientist can gain more information on the
sampling error if more than one sample is taken at each location.
These replicates are used to provide a measure of the sample to
sample variation.
6.2.1.1 Cost of Collection
Determination of the number of samples in the above section
is based upon the coefficient of variation of the sample
population. There are many cases where the number of samples
required by this method is not acceptable because of the cost of
sample collection, the cost of analysis or limitations imposed by
the lack of available laboratory capacity to handle the analyses.
The following paragraph outlines a means for integrating the
costs with the precision of the estimates obtained by the
sampling program.
The total cost of soil studies often follows a linear form of
equation similar to equation 6.2.3. (After Petersen and Calvin,
1965)
C = C + nC + nC /£ « ,.
o s a (6.2.3)
where n = number of samples
C = total costs
Co = overhead or fixed costs
Cs = cost of sampling
Ca = cost of analysis
The equation is used with equation 6.2.2 to arrive at the number
of samples that will satisfy the budget and still have an
identified precision.
EXAMPLE: Samples costs (Cs) are $1800 for collection,
preparation and shipping. Fixed costs (Co) are $15,000.
Analytical costs (Ca) are $2,000. The budget for the
study is $75,000. The estimated coefficient of
variation obtained from another similar study conducted
nearby was 25%. The precision desired on the results
is +10% at a 95% confidence level.
26
-------
Equation 6.2.2 indicates that approximately 26
samples should be taken. Equation 6.2.3 however
indicates the following:
C = C + nC + nC
o s a
C = CQ + n(Cg + Ca)
n = (C - CQ)/(Cs + Ca)
n = (75,000 - 15,000)/(1,800 + 2,000)
n = 15.8 which is rounded down to 15
The effects of the budgetary constraints must be resolved
either by reducing the precision or else by increasing the
budget. Assuming that the budget cannot be changed, equation
6.2.2 will again be used to arrive at the t value that would
result from the use of the smaller number of samples.
n = (CV)2t2/p2
15 = (25)2t2/102
t2 = 2.4000
t = ti = 1.5491
This value for t is obtained from the statistical t—tables for a
two-tailed t-test for 14 degrees of freedom (n-1). This value
indicates that the significance level for the test would have to
drop to 85% with the smaller number of samples or the allowable
margin of error (p) would have to be increased.
6.2.2 Location of Sampling Points
Once the number of samples is determined their location can
be planned. A map of the study area is overlain with a grid of
an appropriate scale. The starting point of the grid should be
randomly selected rather than located for convienence. This can
be accomplished by selecting four random numbers from a random
number table. The first two numbers locate a specific grid
square on the overlay. The second two identify a point within
that grid square. This point is fixed on the map and the entire
grid shifted so that the lower right corner of the original grid
square lies on the point chosen. This procedure is simple and
fast. Using this technique avoids the questions that are often
raised about bias. A second alternative is to select two map
coordinates at random. This becomes the starting point for the
grid used in sampling. All lines are then laid out on the map
overlay starting at that point.
The grid prepared in this fashion becomes the basis for the
selection of the sample locations. Using the number of samples
27
-------
(n) determined in Section 6.2.1, n pairs of random numbers are
selected from a random number table. These pairs of numbers
become the X and the Y coordinates of the sample location. This
procedure is the basis for locating sampling points in all of the
methods where random samples are collected. In situations similar
to Love Canal, a house lot is the area to be represented by each
sample. The grid intersections can be used to locate the houses
to sample or a listing of the houses can be prepared and the
individual samples identified by a random number procedure.
6.3 Stratified Random Sample
Prior knowledge of the sampling area and information obtained
from the background data can be combined with information on
pollutant behavior to reduce the number of samples necessary to
attain a specified precision. The statistical technique used to
produce this savings is called stratification. Basically it
operates on the fact that environmental factors play a major role
in leaching and concentrating pollutants in certain locations.
For example, a pesticide that is attached to clay particles may
accumulate in stream valleys because of soil erosion from
surrounding agricultural lands. The agricultural land may have
lost most of the pesticide because of the same erosion.
Stratification in this case might be along soil types or along
elevational changes. Soil types are frequently used as a means
of stratification, especially if they are quite different in
physical and chemical properties.
Examples of factors used for stratification are:
Soil type - Comus silt loam and Baile silt loam (The
Comus contains mica that is known to bind a
number of pollutants.)
Texture - Sandy loam and clay loam
Drainage - Stream bottom, valley slope and ridgetop
Uses - Cropland and fence rows
Practices - No till cropland and plowed land
Horizons - A horizon and C horizon (Surface (A)
usually has more organic matter)
The whole purpose of stratification is to increase the
precision of the estimates made by sampling. The stratified
random sampling plan should lead to this increased precision if
28
-------
the strata are selected in such a manner that the units within
each stratum are more homogenous than the total population.
Stratification must remove some of the variation from the
sampling error or else there is no benefit from the effort.
In general, the more stratification, the greater the
increase in precision. As was mentioned earlier, Petersen and
Calvin (1965) have pointed out that the benefit of stratification
has a limit where the law of diminishing returns takes over and
no further gain in precision is encountered.
At least two samples must be taken from each stratum in order
to be able to obtain an estimate of the sampling error. The
number of sampling units is usually allocated according to a
proportion based on the land area covered by each stratum. (i.e.
if the area of soil in one stratum is 25% of the total study
area, then 25% of the samples would be taken from that stratum.)
Proportional allocation is used in soil sampling work primarily
because the variance within a general area tends to be constant
over a number of soil types. A pilot study would allow the
scientist to determine if this is in fact the case. If the
variances are materially different, the allocation must be on the
basis of optimum allocation.
The procedure used once the number of samples is determined
is the same as that outlined in Section 6.2 for the simple random
design. Each stratum is handled as a separate simple random
sampling effort.
6.4 Systematic Sampling
The systematic sampling plan is an attempt to provide better
coverage of the soil study area than could be provided with the
simple random sample. The method is easy to use therefore it has
been popular in many cases. Samples are collected in a regular
pattern (usually a grid or a line transect) over the areas under
investigation. The starting point is located by some random
process; then all other samples are collected at regular
intervals in one or more directions. The orientation of the grid
lines should also be randomly selected. This creates problems
however when a pollution plume is the subject of the
investigation. The orientation of the grid should be such that
the lines in one direction are parallel to the general trace of
the plume. This is especially important if kriging is going to
be used.
The spacing on the grid also becomes important if
regionalized variable theory (this is the basis of kriging) is
used to design the study. The theory is based upon the spacing
of data points along the grid lines. The samples must be close
29
-------
enough to provide a measure of the continuity of the location to
location variation within a soil sampling unit. If on the other
hand a measure of the mean and variance of the population is the
focus of the grid sampling array, the samples must be placed
outside of the "range" of the variance of each point. This
allows the environmental scientist to collect samples that are
not influenced by the regionalized variables. Beckett and
Webster (1971) indicate that about 50% of the reported variation
occurs within the first few meters of a point. This would
indicate that the range beyond which kriging is not effective
probably lies at a distance of approximately ten meters or
greater. Beyond this distance the mean and variance of the
population are the only parameters that can be determined.
A number of studies are reviewed by Petersen and Calvin
(1965) that have compared systematic sampling with a simple
random or a stratified random sampling plan. The results favored
the systematic sampling in nearly all cases. The optimum
sampling is obtained with a triangular grid design located over
the area, but the square grid is almost as efficient. The fact
that the square grid is probably easier to set out in the field
would suggest that a square or rectangular grid should be used
unless there is some reason for desiring to optimize the
placement of sampling points
The systematic sampling plan is ideal when a map is the final
product. This provides a uniform coverage of the area and also
allows the scientist to have points to use in developing the map.
(Most mapping techniques use a grid to generate the points for
plotting isopleths of concentration, etc.)
The location of the grid on the area would be according to
the procedures outlined in Section 6.2.2. At each grid
intersection samples would be collected according to one of the
methods outlined in Section 7. It is desirable to collect
duplicate samples at some of the locations in order to provide a
measure of the sampling error. This will increase the precision
of the estimate of concentration and also allow the researcher to
check the reliability of the sampling at the same time.
There are two problems that may limit the use of this design.
First, the estimation of the sampling error is difficult to
obtain from the sample itself unless double sampling is used at a
number of sites. The variance cannot be calculated unless some
method of mean successive difference test is used to evaluate the
data. The second problem area concerns the presence of trends
and periodicity in the data. Both of these create problems when
the direction of the grid aligns with the pattern in the data.
Soil sampling seldom encounters the cyclic pattern to a degree
that a problem is created. Trends however are common in soil
pollution work. That is the whole purpose for the sampling in
many cases. There are a whole array of methods available for
handling the analysis of data from sequential sampling. An
30
-------
excellent reference for soil scientists working in this area is a
book by John C. Davis (1973) entitled Statistics and Data
Analysis in Geology. Davis spends considerable time discussing
the analysis of sequences of data. Techniques such as least
squares analysis, regression, filtering or time-trend analysis,
autocorrelation, cross correlation, Fourier transformations, map
analysis , nearest neighbor analysis, cluster analysis,
contouring, trend surface analysis, double Fourier series and
moving averages are presented. Kriging and multivariate analysis
are also discussed. A valuable addition to this text is a series
of Fortran computer subroutines for conducting most of these
analyses.
Yates (1948) and Quenouille (1949) present excellent
reviews of the use of systematic sampling from a statistical
point of view.
6.5 Judgemental Sampling
This technique is often used with one of the other methods in
order to cover areas of unusual pollution levels or where effects
have been seen in the past. The problem with the approach is
that it tends to lead to sloppy science and to wrong conclusions.
The scientist's own bias is built into the sampling effort and
the data therefore often suspect. Where the data has a potential
for litigation or where there is a likelihood of emotional
reactions to the results, this system should be absolutely
avoided. A simple random design with a known precision can be
developed that will allow the scientist to determine the presence
of pollutants without the risk of creating problems that cannot
be handled. If it is essential that judgemental sampling be
used, duplicate or triplicate samples should be taken in order to
have some measure of precision.
6.6 Phasing the Study
Often data is not available for use in planning a study in a
particular area. This type of situation leads to a phased
approach. The first phase of the study might be a simple random
study design with a 68% confidence level. The results of this
study would then be used to design a more definitive study with a
95% confidence level. This latter study could use a stratified
random design or a systematic sampling grid. The grid design
would allow the researcher to analyze the data using kriging and
thus find where additional samples are needed to further refine
the sampling design so that the entire area is covered at the 95%
confidence level.
31
-------
Careful planning can provide data at each phase of the study
that can stand the scrutiny of the scientific and legal
communities and at the same time not place all of the resources
into a one shot study that does not meet the situation at a
particular site. Planning takes time, but it will pay off in the
long run by providing the data needed at a precision that is
acceptable to most scientists. The use of phases can greatly
help in this process by allowing the data to grow as the
awareness of the study situation evolves.
6.7 Control Areas
Control sites are used quite often in major soils studies
especially if the study is attempting to determine the extent and
presence of local pollution. Sites for controls must be as
representative as possible of the study area. A careful survey
of the area should be made prior to the final selection. In most
cases it is desirable to spend as much time searching out data
on the control as on the study area. The purpose of the control
area is to serve as a base line against which the results of the
soil sampling study can be compared.
Soil type should be the main factor chosen in selecting the
control but factors such as depth to groundwater, location in
relation to pollution sources and vegetation type all should be
taken into consideration in making the selection. Where
pollution sources are being studied the ideal selection would be
a control site that only differs from the study area by the lack
of the pollution source under investigation. This is seldom
possible but every attempt should be made to reduce the factors
that are different.
32
-------
SECTION 7
SAMPLE COLLECTION
There are two portions of the soil that are important to the
environmental scientist. The surface layer (0-15 cm) reflects
the deposition of airborne pollutants; especially those recently
deposited pollutants. Pollutants that have been deposited by
liquid spills, or by long term deposition of water soluble
materials may be found at depths ranging up to several meters.
Plumes emanating from hazardous waste dumps or leaking storage
tanks may be found at considerable depths. The methods of
sampling each of these are slightly different; but, all make use
of one of two basic techniques. Samples can either be collected
with some form of core sampling or auger device; or, they may be
collected by use of excavations or trenches. In the latter case
the samples are cut from the soil mass with spades or short
punches. The American Society for Testing and Materials (ASTM)
has developed a number of methods that have direct application to
soil sampling. These often need to be modified slightly to meet
the needs of the environmental scientist that requires samples
for chemical analyses since the ASTM methods are designed
primarily for engineering tests. The technioues that are
utilized should be closely coordinated with the analytical
laboratory in order to meet the specific requirements of the
analytical methods used.
The methods outlined below are for the collection of soil
samples alone. At times it is desirable to collect samples of
soil water. In these cases use can be made of some form of
suction collector. The statistical designs would be the same no
matter which of the soil water collectors was used. In those
cases where suction devices are used, the sampling media is water
and not soil even though the samples are a good reflection of
soluble chemicals that may be moving through the soil matrix.
These methods are not discussed in this report.
33
-------
7.1 Surface Sampling
Surface soil sampling can be divided into two categories —
the upper 15 cm and the upper meter. The very shallow pollution
such as that found downwind from a new source or at sites of
recent spills of relatively insoluble chemicals can be sampled by
use of one of the methods listed in Section 7.1.1. The deeper
pollutants found in the top meter are the more soluble, recent
pollutants or those that were deposited on the surface a number
of years ago. These have begun to move downward into the deeper
soil layers. One of the methods in 7.2 should be used in those
cases.
7.1.1 Sampling with a Soil Punch
A number of studies of surface soils have made use of a punch
or thin walled steel tube that is 15 to 20 cm long to extract
short cores from the soil. The tube is driven into the soil with
a wooden mallet; the core and the tube are extracted; the soil
is pushed out of the tube into a stainless steel mixing bowl and
composited with other cores. Two alternates are the short
King-tube samplers or the tube type density samplers used by the
Corps of Engineers. (These sampling devices can be supplied by
any field equipment company or by agricultural equipment
companies.) The latter sampler is machined to a predetermined
volume and is designed to be handled and shipped as a soil-tube
unit. A number of similar devices are available for collecting
short cores from surface soils.
The soil punch is fast and can be adapted to a number of
analytical schemes provided precautions are taken to avoid
contamination during shipping and in the laboratory. An example
of how this method can be adapted would be to use the system to
collect samples for volatile organic chemical analysis. The
tubes could be sealed with a Teflon plug and coated with a vapor
sealant such as paraffin or, better yet, some non-reactive
sealant. These tubes could then be decontaminated on the outside
and shipped to the laboratory for analyses.
7.1.2 Ring Sampler
Soil engineers have a tool that can be purchased from any
engineering equipment supply house that can be used to collect
larger surface samples. A seamless steel ring, approximately 15
to 30 cm in diameter, is driven into the soil to a depth of 15 to
20 cm. The ring is extracted as a soil-ring unit and the soil
34
-------
removed for analysis. These large cores should be used where the
results are going to be expressed on a per unit area basis. This
allows a constant area of soil to be collected each time.
Removal of these cores is often difficult in very loose sandy
soil and in very tight clayey soils. The loose soil will not
stay in the ring. The clayey soil is often difficult to break
loose from the underlying soil layers thus the ring must be
removed with a shovel.
This device has not been used extensively for collecting
samples for chemical analysis but the technique should offer a
useful method for collecting samples either for area
contamination measurements or for taking large volume samples.
7.1.3 Scoop or Shovel Sampling
Perhaps the most undesirable sample collection device is the
shovel or scoop. This technique is often used in agriculture but
where samples are being taken for chemical pollutants, the
inconsistencies are too great. Samples can be collected using a
shovel or trowel if area and/or volume are not critical. Usually
the shovel is used to mark out a boundary of soil to be sampled.
The soil scientist attempts to take a constant depth of soil but
the reproducibility of sample sizes is poor; thus the variation
is often considerably greater than with one of the methods listed
above.
7.2 Shallow Subsurface Sampling
Precipitation may move surface pollutants into the lower soil
horizons or move them away from the point of deposition by
surface runoff. Sampling pollutants that have moved into the
lower soil horizons requires the use of a device that will
extract a longer core than can be obtained with the short probes
or punches. Three basic methods are used for sampling these
deeper soils:
• Soil probes or soil augers
• Power driven corers
• Trenching
The soil probe collects 30 or 45 cm of soil in intact,
relatively undisturbed soil cores whereas the auger collects a
"disturbed" sample in approximately the same increments as the
probe. Power augers can use split spoon samplers to extract
cores up to 60 cm long. With special attachments longer cores
can be obtained with the power auger if this is necessary.
35
-------
The requirement for detail often desired in research studies
or in cases where the movement of the pollutants is suspected to
be through very narrow layers cannot be met effectively with the
augers. In these cases some form of core sampling or trenchinq
should be used.
7.2.1 Soil Probes and Hand Augers
Two standard tools used in soil sampling are the soil probe
(often called a King-tube) and the soil auger. These tools are
designed to acquire samples from the upper two meters of the soil
profile. The soil probe is nothing more than a stainless steel
or brass tube that is sharpened on one end and fitted with a
long, T shaped handle. These tubes are usually approximately 2.5
cm inside diameter although larger tubes can be obtained. The
cores collected by the tube sampler or soil probe are considered
to be "undisturbed" samples although in reality this is probably
not the case. the tube is pushed into the soil in approximately
20 to 30 cm increments. The soil core is then removed from the
probe and placed in either the sample container or in a mixing
bowl for compositing.
The auger is approximately 3 cm in diameter and is used to
take samples when the soil probe will not work. The samples are
"disturbed"; therefore, this method should not be used when it
is necessary to have a core to examine or when very fine detail
is of interest to the scientist. The auger is twisted or screwed
into the soil then extracted. Because of the length of the auger
and the force required to pull the soil free, only about 20 to 30
cm maximum length can be extracted at one time. In very tight
clays it may be necessary to limit the length of each pull to
about 10 cm. Consecutive samples are taken from the same hole
thus cross contamination is a real possibility. The soil is
compacted into the threads of the auger and must be extracted
with a stainless steel spatula.
Larger diameter augers such as the bucket auger, the Fenn
auger and the blade augers can also be used if larger samples are
needed. These range in size from 8 to 20 cm in diameter.
If distribution of pollutant with depth is of interest,
the augers and the probes are not recommended because they tend
to contaminate the lower samples with material from the surface.
The probe is difficult to decontaminate without long bore brushes
and some kind of washing facility. One alternative is to take
several waste cores at each site prior to collecting the actual
samples. This allows the probe tube to be cleaned by the
scouring action of soil at similar concentrations to those found
in the sample taken. This should remove any contamination
36
-------
leftover from previous locations. Where there is a potential for
litigation, decontamination is essential to avoid any question
about cross contamination. The augers have some of the same
decontamination problems but the open thread surfaces allow
easier access to the collection surfaces; therefore, they are
easier to clean. See Section 7.8 for more detail on
decontamination procedures.
One final warning about the use of the hand augers and soil
probes. There are many soil scientists with back problems that
have resulted from trying to extract a tool that has been
inserted too far into the soil. A foot jack is a necessary
accessory if these tools are to be used. The foot jack allows
the tube to be removed from the soil without use of the back
muscles.
7.2.2 Power Augers and Core Samplers
These truck or tripod mounted tools are used for collecting
samples to depths greater than approximately 30 cm. Standard
ASTM methods for use of these tools are available from the
American Society for Testing and Materials or can be found at any
college or university library. The methods outlined in Section
7.3 are applicable in this case and will not be discussed
further.
7.2.3 Trenching
This method of soil sampling is used to carefully remove
sections of soil during studies where a detailed examination of
pollutant migration patterns and detailed soil structure are
required. It is perhaps the least cost effective sampling method
because of the relatively high cost of excavating the trench from
which the samples are collected. It should therefore be used
only in those cases where detailed information is desired.
A trench approximately 1 meter wide is dug to a depth
approximately one foot below the desired sampling depth. The
maximum effective depth for this method is about 2 meters unless
done in some stepwise fashion. Where a number of trenches are to
be dug, a backhoe can greatly facilitate sampling. The samples
are taken from the sides of the pit using the soil punch or a
trowel.
37
-------
The sampler takes the surface 15 cm sample using the soil
punch or by carefully excavating a 10 cm slice of soil that is 10
cm square on the surface. The soil can be treated as an
individual sample or composited with other samples collected from
each face of the pit. After this initial sample is taken the
first layer is completely cut back exposing clean soil at the too
of the second layer to be sampled. Care must be exercised to
insure that the sampling area is clear of all material from the
layers above. The punch or trowel is then used to take samples
from the shelf created by the excavation from the side of the
trench. This process is repeated until all samples are taken.
The resulting hole appears as a set of steps cut into the side of
the trench as is shown in Figure 7.1.
An alternate procedure that is also effective results from
using the punch to remove soil cores from the side of the trench
at each depth to be sampled (Figure 7.1). Care must be taken to
guard against soil sloughing down the side of the hole. A shovel
should be used to carefully clean the soil sampling area prior to
driving the punch into the trench side.
7.3 Sampling for Underground Plumes
This type of sampling is perhaps the most difficult of all of
the soil sampling methods. Often it is conducted along with
groundwater and hydrological sampling. The equipment required
usually consists of large, vehicle mounted augers and coring
devices although there are some small tripod mounted coring units
available that can be carried by several men using backpacks.
7.3.1 Usual Procedure for Underground Plume Sampling
The procedure listed here closely follows ASTM method
D1586-67 in many respects. The object of the sampling is to take
a series of 45.7 cm (18 in) or 61 cm (24 in) undisturbed cores
with a split spoon sampler. (Longer cores can be obtained by
combining several of the shorter tubes into one long split
spoon.) A 15.2 cm (6 in) auger is used to drill down to the
desired depth for sampling. The split spoon is then driven to
its sampling depth through the bottom of the augered hole and the
core extracted.
The ASTM manual calls for the use of a 63.5 kg (140 Ib)
hammer to drive the split spoon. The hammer is allowed to free
fall 76 cm (30 in) for each blow to the spoon. The number of
blows required to drive the spoon 15.2 cm (6 in) is counted and
recorded. The blow counts are a direct reflection of the density
38
-------
Figure 7.1 Trench amp!ing techniques
39
-------
of the soil and can be used to obtain some information on the
soil structure below surface. Unless this density information is
needed for interpretive purposes, it may not be necessary to
record the blow counts. In soft soils the split spoon can often
be forced into the ground by the hydraulic drawdown on the drill
rig. This is faster than the hammer method and does not require
the record keeping necessary to record the blowcounts. Most
commercial drilling companies have the equipment and the
experience required to conduct this type of sampling with some
supervision from the field scientist.
Samples should be collected at least every 1.5 meters (5 ft)
or in each distinct stratum. Additional samples should be
collected where sand lenses or thin silt and sand layers appear
in the profile. This sampling is particularly important when
information on pollution migration is critical. Soluble
chemicals are likely to move through permeable layers such as
sand lenses. This appears to be especially important in tight
clay layers where the main avenue of water movement is through
the porous sandy layers.
Detailed core logs should be prepared by the technical staff
present at the site during the sampling operation. These logs
should note the depth of sample, the length of the core and the
depth of any features of the soil such as changes in physical
properties, color changes, the presence of roots, rodent
channels, etc. If chemical odors are noted or unusual color
patterns are detected, these should be noted also. Blow counts
from the hammer should be recorded on the log along with the data
mentioned above.
The procedure using samples collected every 1.5 meters (5 ft)
is most effective in relatively homogeneous soils. A variation
in the method that is preferred by soil scientists is to collect
samples of every distinct layer in the soil profile. Large
layers may be sampled at several points if they are unusually
thick. A disadvantage of this approach is the cost for the
analyses of the additional samples acquired at a more frequent
interval. The soil horizons or strata are the avenues through
which chemical pollutants are likely to migrate. Some are more
permeable than others and are thus more likely to contain traces
of the chemicals if they are moving through the soil. Generally
speaking the sands and gravels are more prone to contamination
than are the clays because of increased permeability. This is
especially true out on the leading edges of the plume and shortly
after a pollutant begins to move. Low levels found in these
layers can often serve as a warning of a potential problem at a
later date.
Decontamination of the large equipment required for plume
sampling is difficult but necessary if the study is to be
useable. The staged sampling using the auger then the split
40
-------
spoon helps reduce the chances of serious cross contamination.
The auger carries considerable soil in the threads of the bit.
This can only be removed with high pressure hoses.
A disadvantage of this type of sampling is the impact of the
vehicle on yards and croplands. Special care must be taken to
protect yards, shrubs, fences and crops. The yards must be
repaired, all holes backfilled and all waste removed. Plastic
sheeting should be used under all soil handling operations such
as subsampling, compositing and mixing.
7.3.2 Variations in the Procedures
There are several variations for split spoon sampling.
Samples collected from soils below the water table or in very
soft soils may require the use of split spoons equipped with
retainers in the end of the spoon. The retainer is made with
flexible fingers that close over the end of the tube as the spoon
is retracted from the soil.
Samples collected for the analysis of volatile organic
chemicals pose a problem to the environmental scientist. The
volatile chemicals can be lost during transport and handling.
One option that may offer a solution to this problem is the use
of brass, stainless steel or Teflon liners in the split spoon.
Brass liners are available from most engineering and agricultural
supply houses. The liners are easily removed when the split
spoon is opened. The liner tube can be sealed with Teflon plugs
and some form of sealant applied over the plug. The method is
currently used for moisture determinations in agricultural and
research situations. This system avoids the problems of the loss
of chemicals that volatilize into the headspace of the sample
jars. The liners can be discarded after analysis if necessary
thus reducing the labor costs required to clean the tubes.
The main disadvantages of using this modified system is that
no core log can be prepared of the sample. The author was
informed that some laboratories are reluctant to develop methods
that can make use of samples acquired in this fashion.
7.4 Compositing
Many sample plans call for compositing of the soils
collected at a sampling location. This creates a problem from
the point of view of the soil scientist. The key to any
statistical sampling plan is the use of the variation within the
sample set to test hypotheses about the population and to
41
-------
determine the precision or reliability of the data set. As was
mentioned earlier, the composite sample provides an excellent
estimate of the mean but does not give any information about the
variation within the sampling area. Section 7.4.1 discusses one
alternative that is a combination of the compositing methods and
random sampling with duplication. Three methods that have been
used to composite soil are presented below.
7.4.1 Estimating Sample Variance.
The problem with the statistical analysis is found in the
lack of duplication within the sampling location. Each subsample
is combined into the composite therefore the data that is
contained in the subsample is averaged with all other subsamples.
The lack of a measure of the sampling error is the cause of the
problem confronting the statistician. Multiple samples taken at
each location would avoid this problem but costs usually preclude
this. A compromise is possible by only analyzing duplicates or
triplicates at a percentage of the locations. The exact location
is chosen by use of a random number table and should be
identified before the study begins. The duplicates should not be
two subsamples taken from the same composite sample but should be
made up of a second set of subsamples.
Large cores such as those collected by split spoon can be
split lengthwise in half. Each half is thus used as part of two
separate composite samples. This avoids the time required to
take the second set of cores but provides the duplication
necessary for calculating the sampling error.
7.4.2 Compositing with a Mixing Cloth
Soil scientists often use a large plastic or canvas sheet for
compositing samples in the field. This method works reasonablv
well for dry soils but has the potential for cross contamination
problems. Organic chemicals can create further problems by
reacting with the plastic sheet. Plastic sheeting, however, is
inexpensive and can therefore be discarded after each sampling
site.
This method is difficult to describe. It can be visualized
if the reader will think of this page as a plastic sheet. Powder
placed in the center of the sheet can be made to roll over on
itself if one corner is carfully pulled up and toward the
diagonally opposite corner. This process is done from each
corner. The plastic sheet acts the same way on the soil as the
paper would on the powder. The soil can be mixed quite well if
42
-------
it is loose. The method does not work on wet or heavy plastic
soils. Clods must be broken up before attempting to mix the
soil.
After the soil is mixed, it is again spread out on the cloth
into a relatively flat pile. The pile is quartered. A small
scoop, spoon or spatula is used to collect small samples from
each quarter until the desired amount of soil is acquired (this
is usually about 250 to 500 grams of soil but can be less if the
laboratory desires a smaller sample). This is mixed and placed
in the sample container for shipment to the laboratory. The
waste material not used in the sample should be disposed of in a
safe manner. This is especially important where the presence of
highly toxic chemicals is suspected.
7.4.3 Compositing with a Mixing Bowl
An effective field compositing method has been to use large
stainless steel mixing bowls. These can be obtained from
scientific, restaurant, or hotel supply houses. They can be
decontaminated and are able to stand rough handling in the field.
Subsamples are placed in the bowls, broken up, then mixed using a
large stainless steel scoop. The rounded bottom of the mixing
bowl was designed to create a mixing action when the material in
it is turned with the scoop. Careful observance of the soil will
indicate the completeness of the mixing.
The soil is spread evenly in the bottom of the bowl after the
mixing is complete. The soil is quartered and a small sample
taken from each quarter. The subsamples are mixed together to
become the sample sent to the laboratory. The excess soil is
disposed of as waste.
7.4.4 Laboratory Compositing
Small sets of samples can often be composited better in the
laboratory than in the field. A number of the small surface
cores discussed in Section 7.1.1 can be placed in the sample
bottle for shipment to the laboratory. These can then be placed
in a stainless steel laboratory mixer and mixed to the degree
needed by the analytical methods. This technique is the only
method that may be useful for obtaining composite samples for
some types of soils and may be the best method to use if
compositing of samples is necessary in a particular situation.
43
-------
7.5 Replicate Samples
The quality control program will require duplicate or
triplicate samples from a percentage of the sampling sites. These
may be collected from the composite or they may be comprised of
duplicate sets of samples. The latter is the preferred method.
A question often arises about how to handle the analytical
data for these multiple results. All analytical results for the
field replicates should be reported. Proper statistical designs
can use this data to increase the precision of the estimates
made. There is a tendency on the part of many scientists to
discard unusual results (outliers) and to average the remainder
of the samples. The discussion on soil variability given earlier
should point up the problem with this approach. The outliers are
probably part of the normal, wide variation seen in soils data.
Averaging the numbers in effect throws away data on the sampling
error that is needed to determine the reliability of the data
collected.
7.6 Miscellaneous Tools
Hand tools such as shovels, trowels, spatulas, scoops and pry
bars are helpful for handling a number of the sampling
situations. Many of these can be obtained in stainless steel for
use in sampling hazardous pollutants. A set of tools should be
available for each sampling site where cross contamination is a
potential problem. These tool sets can be decontaminated on some
type of schedule in order to avoid having to purchase an
excessive number of these items.
A hammer , screwdriver and wire brushes are helpful when
working with the split spoon samplers. The threads on the
connectors often get jammed because of soil in them. This soil
can be removed with the wire brush. Pipe wrenches are also a
necessity as is a pipe vise or a plumbers vise.
7.7 Record Keeping
One of the vital components of the protocol is to adequately
define the records required during the study. Good records
become extra important if litigation results from the data
collected. Every sample will be questioned in an attempt to
either discredit or verify the data depending upon the side of
the issue the attorney represents. Some of the records are
discussed below.
44
-------
7.7.1 Log Books
The sample teams should maintain an official log book of the
investigation. Observations of the field conditions, equipment
used, procedures followed and crew members involved are recorded
for each days sampling. These log books should be bound and all
data must be recorded in ink (preferably black ink) . Each log
book should be maintained by the crew leader and signed by him.
No erasures are allowed. When mistakes are made the data is
lined out with one line only and the corrected data entered above
the incorrect entry or on the next line of the log.
7.7.2 Site Description Forms
These serialized forms record the conditions at each site at
the time the samples are collected. A sketch map and photographs
of the site should be a part of the description. A Polaroid-type
camera should be used so that the pictures of the sites can be
checked before leaving the area of the sample collection. These
forms and the back of the photographs should be signed and dated
by the crew leader responsible for taking the samples. The NEIC
site description form should be used in most cases where the
USEPA is involved.
7.7.3 Sample Tags
Tags made up according to the specifications provided by NEIC
should be printed for use in the soils study. A tag must be
prepared for each sample. All data must be included on the tag
at the time the sample is collected. Wet samples should be
double bagged with the tag in the outer bag. The person
collecting the sample should sign the tag.
7.7.4 Chain-of-Custody Forms
This form is perhaps one of the most important as far as the
legality of the samples is concerned. Chain-of-custody traces
the possession of the sample from its origin through to data
analysis. Most field researchers are not accustomed to observing
the care needed to insure the safe custody of their samples. The
45
-------
samples must be in the the physical custody of the scientist
collecting the sample or else be secured in a facility with
controlled, limited access until the samples are signed for and
transfered to another responsible party. Samples must not be
left unattended in an unlocked vehicle for any reason. There is
nothing more disconcerting to technical representatives of the
regulatory agencies than to spend hours working with data
collected by field teams and then find the data is open to
question because the chain-of-custody had been violated. Samples
are a valuable resource and should be treated accordingly.
7.8 Decontamination
One of the major difficulties with soil sampling arises in
the area of cross contamination of samples. The most reliable
methods are those that completely isolate one sample from the
next. Freshly cleaned or disposable sampling tools, mixing
bowls, sample containers etc. are the only way to insure the
integrity of the data.
Field decontamination is quite difficult to carry out, but it
can be done. Hazardous chemical sampling adds another layer of
aggravation to the decontamination procedures. The washing
solutions must be collected for disposal at a waste disposal
site. The technique outlined below has been used under field
conditions.
7.8.1 Laboratory Cleanup of Sample Containers
One of the best containers for soil is the glass canning jar
fitted with Teflon or aluminum foil liners placed between the lid
and the top of the jar. These items are cleaned in the
laboratory prior to taking them into the field. All containers,
liners and small tools should be washed with an appropriate
laboratory detergent, rinsed in tap water, rinsed in distilled
water and dried in an oven. They are then rinsed in
spectrographic grade solvents if the containers are to be used
for organic chemical analysis. Those containers used for
volatile organics analysis must be baked in a convection oven at
105 C in order to drive off the rinse solvents.
The Teflon or aluminum foil used for the lid liners is
treated in the same fashion as the jars. These liners must not
be backed with paper or adhesive.
46
-------
7.8.2 Field Decontamination
Sample collection tools are cleaned according to the
following procedure.
• Washed and scrubbed with tap water
using a pressure hose or pressurized
stainless steel, fruit tree sprayer.
• Check for adhered organics with a
clean laboratory tissue.
• If organics are present, rinse with
the waste solvents from below.
Discard contaminated solvent by
pouring into a waste container for
later disposal.
• Air dry the equipment.
• Double rinse with deionized,
distilled water.
• Where organic pollutants are of
concern, rinse with spectroqraphic
grade acetone saving the solvent for
use in step 3 above.
• Rinse twice in spectrographic grade
methylene chloride or hexane, saving
the solvent for use in step 3.
• Air dry the equipment.
• Package in plastic bags and/or pre-
cleaned aluminum foil.
The distilled water and solvents are flowed over the surfaces
of all the tools, bowls etc. The solvent should be collected in
some container for disposal. One technique that has proven to be
quite effective is to use a large "glass or stainless steel funnel
as the collector below the tools during flushing. The waste then
flows into liter bottles for later disposal (use the empty
solvent bottles for this. A mixing bowl can be used as a
collection vessel. It is then the last item cleaned in the
sequence of operations.
The solvents used are not readily available. Planning is
necessary to insure an adequate supply. The waste rinse solvent
can be used to remove organics stuck to the tools. The acetone
is used as a drying agent prior to use of the methylene chloride
or hexane.
47
-------
Steam cleaning might prove to be useful in some cases but
extreme care must be taken to insure public and worker safety by
collecting the wastes. Steam alone will not provide assurance of
decontamination. The solvents will still have to be used.
7.9 Quality Assurance
Quality assurance in EPA is usually handled by someone other
than the sampling team. The field team leader is responsible for
insuring that the quality assurance program is carried out
correctly, however. The team will be required to take duplicate
samples at prescribed intervals and will be required to submit
field blanks of all materials used. It would be desirable to
prepare a bulk soil for use as a field blank for the soil
samples. This will have to be handled very carefully because of
the difficulty in finding "clean" soil for use as the blank.
Distilled water can be used in lieu of a soil blank. Additional
samples such as equipment swipes, rinse water and solvents should
be taken on a regular basis to verify the quality of the data
obtained from the samples. Procedures for handling quality
assurance have been outlined in an interim guideline prepared by
the EPA Office of Monitoring Systems and Quality Assurance of the
Office of Research and Development (OMSQA, 1980).
7.10 Safety
Toxic chemicals create a hazard for the soil sampling team.
The team often is operating above plumes containing mixtures of
highly toxic chemicals. The drillers and excavators are in an
especially hazardous position. An industrial safety specialist
should be consulted prior to undertaking a study of these highly
contaminated areas. Physical examinations should be given to the
crew on a regular basis unless the sampling team operates only on
rare occasions in which case they should have physicals before
and after the sampling effort.
Many of the field team members will not want to follow the
procedures outlined by the safety officer. This should not be
tolerated. This problem seems to be especially acute with the
drilling crews. Every effort should be made to provide the teams
with adequate training on the use of all safety equipment and
recovery procedures prior to going into the field.
48
-------
SECTION 8
DATA ANALYSIS
The final step in any sampling study is the analysis and
interpretation of the data that has been collected. It is not
necessary for the field scientist to conduct the data analysis,
but his input is necessary if any interpretation of the data is
to be made. Impressions and observations obtained during on site
activities are needed to adequately determine the actual behavior
of the pollutant.
The person doing the data analysis must keep in mind the
purpose for which the samples were collected. These purposes can
usually be grouped into one of the following categories.
o Estimate the level and variability of
a pollutant in a geographic area.
o Determine if the pollution measured is
above some standard or is higher than
the ambient levels found in the control
area.
o Define the areal extent and depth of
the pollution and map the pattern
of the distribution.
There are statistical tests available for handling data
collected by each method discussed in Section 6. Prior to
attempting to use any of the methods, a statistician versed in
sampling design should be consulted to assure that the
appropriate design is being used. An assumption has been made
that this was done at the beginning and is not being done at the
end of the study.
49
-------
8.1 Analysis of Data From a Simple Random Design
The simplest analysis of any of the designs is that for the
simple random analysis. This analysis can be done easily on a
desk calculator if the number of points is not too great. The
first three parameters that should be calculated are the mean,
the variance and the confidence interval. Where the results are
to be compared to other areas a number of tests are likely to be
used. These are discussed below.
8.1.1 Basic Parameters
The mean (Eq,
confidence interval
equations.
8.1.1), variance (Eq. 8.1.2) and the
(Eq. 8.1.3) are calculated by the following
y = 2 y,/n
1=1 1
(8.1.1)
n
- 2
V(y) = 2(y. - y)Vn(n-l)
(8.1.2)
where y. = ith observation
n = number of samples
y = mean of the samples
V(y)= variance of the mean
y ±
(8.1.3)
where L = confidence interval
ta= t-statistic for a two-tailed t-test
at the a significance level for
(n-1) degrees of freedom.
a = significance level
V (y) = variance of the mean
50
-------
8.1.2 Use of t-Test
The data for a control population and the data for the
contaminated areas can be compared statistically using either
the analysis of variance test or the Studentized t-test. The
control population often has less samples than the polluted area
thus some adaptation will have to be made in the data to
compensate for the inequality in the number of samples in the two
treatments. The t-test uses equation 8.1.4 (Li, 1959) to
determine if the two sets of samples are different. This
equation assumes homogeneity of variances which is most often the
case in soils work.
t =
-y2)
(l/n1 +
l/nj
(8.1.4)
where
Cni
T (
(i=i
t =
y =
J\.
2
n =
calculated t-value
mean for sample set x
pooled variance calculated
by formula 8.1.5
number of samples in x
n.
'K
- 2)
(8.1.5)
The two data sets are considered to be from the same
population (i.e., they are equal) if the calculated t-value (tc)
falls outside of the critical region (i.e., it falls within the
range -ta< tc
-------
8.1.3 Analysis of Variance of Simple Random Design
If the analysis of variance (ANOVA) is used to determine the
difference between the two sets of sample data, the following
table can be used to do the calculations for the ANOVA. The term
treatment is a term used by statisticians when handling different
data sets. This should be kept in mind when reading the sections
that follow. One treatment is the data from the polluted area
and the second is the data from the control area.
Table 8.1.1 Analysis of Variance for Simple Random Design
Source of Variation freedom* Mean s
-------
Occasionally a study will be conducted where no control
exists. This is not a recommended practice, but there are
situations where there is no alternative. The analysis in this
restricted case is limited simply to the mean and the variance of
the data.
8.2 Data Analysis for Stratified Random Design
The origin of the stratified random design is similar to the
simple random design with the exception that there is
stratification of the study area into subareas. The following
sections provide the procedures for the analyses of the data that
can be conducted with this design.
8.2.1 Basic Parameters for the Stratified Random Design
The calculation of the mean and the variance for the
stratified random design can take two forms. Only one of the
forms will be presented here. Petersen and Calvin (1965) note
that in cases where the variance is common between two strata and
proportional allocation is used to assign samples to each strata
the calculations for the mean and the variance are simplified.
As was mentioned earlier, proportional allocation is common in
soils work because the variance tends to be the same over strata
that are in close geographic proximity. Petersen and Calvin
(1965) and Cochran (1965) can be consulted for the other approach
to calculating these parameters.
The mean is calculated by use of equation 8.2.1 and the
variance by use of 8.2.2.
n
y = ( X y./n)
i=l 1 (8.2.1)
V(y) - s2 /n (8.2.2)
where y = mean of all strata
V(y) = variance of the mean y
n = total number of observations y.
2 i
s = pooled variance (see equation 8.1.5)
53
-------
8.2.2 Analysis of Variance for Stratified Random Design
The analysis of variance for the stratified random design is
similar to that for the simple random design. Table 8.2.1
presents the ANOVA table for handling the data from a soil
sampling study where this design has been used.
Table 8.2.1 Analysis of Variance for Stratified Random Design
Source of Variation
*
Strata
Within strata
Total
T /n = Sum of
Degree of
Freedom
(h-1)
(2n-h)
(In-1)
the total of
Mean Squares*
UIT2/n)-(G2/ln))/(h-l)
(2y2-( T2/n))/(2n-h)
(2y2-(G2/Zn))/(£n-l)
each stratum squared and
divided by the number of samples in the stratum.
2
G /£n = The square of the sum of all observations
divided by the total number of observations.
2
y = The sum of the square of each observation
h = The number of strata.
The similarity between Table 8.1.1 and 8.2.1 can be explained
by the fact that the simple random experiment with a control is
nothing more than a stratified random experiment with 2 strata.
The only difference is in the identifiers for treatment (k) and
strata (h) .
Tests and interpretation of the results of the ANOVA would be
the same as that presented for the simple random design. If the
strata are different, it may not be possible to determine which
stratum is causing the ANOVA to show the differerence if more
than two strata are used. This can be determined, however, by
using tests such as the single degree of freedom, regression
analysis or by calculating the least significant difference
(LSD). This latter test can be used to determine if there is a
difference between the stratum means (y) or the differences
54
-------
between stratum totals (Th). 'Inequality 8.2.3 can be used for
testing the differences between two stratum means.
(8.2.3)
where Y h = mean for the h stratum
ta = t value from statistical
tables at the ( Jn-h)
degree of freedom and
the a significance level
Jn = total number of observations
s = Pooled variance which is
P
also the within strata mean
square from Table 8.2.1
The absolute, |y - y |, is compared to the term on the right side
of the inequality; if larger, the difference is significant.
Each pair of strata means are compared in this fashion..
8.3 Data Analysis for Systematic Sampling Designs
This design provides uniformly spaced data for the entire
study area. The uniform distribution allows reliable maps to be
drawn, trend analysis to be conducted and facilitates kriging
calculations. The literature (Petersen and Calvin, 1965;
Cochran, 1946; Yates, 1948) indicates that the systematic
approaches provide an increased precision over the two random
designs under most conditions. As was mentioned earlier, cyclic
and periodic phenomena can create problems in data analysis. The
systematic design is in reality a stratified design with one
observation per stratum. The lack of multiple samples taken from
each of these stratum precludes the use of analysis of variance
because there are no degrees of freedom for use with the within
stratum mean square.
A number of tests have been devised for extracting
statistical data from these designs. One method analyzes the
differences between observations made at adjacent nodes on the
grid. Either 4, 6, or 8 differences are generated for each point
on the grid. The number depends upon the design of the grid and
the location of the particular pair of points. The set of
differences become the source of data for calculating a within
55
-------
stratum mean square. The analysis would be the same as that
shown in Table 8.2.1.
Regression techniques can also be used. Regression
effectively removes the effects of trends over the area and still
provides information about the concentration differences and the
variance. These tests do not allow the researcher to identify
the spatial relationship of the data points, however.
When systematic designs are used, an attempt should be made
to evaluate the effects of cyclic patterns that may be present in
the data. Autocorrelation analysis can be used to identify if
these patterns are present or not. Frequently examination of a
plot of the residuals from a regression analysis will also reveal
cyclic patterns in the data. If these patterns exist, some
effort should be made to remove their effects from the data
before analysis is done.
8.3.1 Kriging Analysis of Systematic Data
The technique called kriging was first developed by D.G.
Krige (1966) as a means for estimating gold ore reserves in South
Africa. The technique develped by Krige was based upon the use
of moving averages in handling systematic data. His techniques
were further developed by G. Matheron (1971) at the National
School of Mines in Paris and have since been expanded into a
whole body of knowledge called qeostatistics. Matheron called
the method kriging in honor of D. G. Krige. This technique has
been used in European and South African mining fields for some
time but it has only recently begun to be used in pollution
control work.
A number of soil scientists have explored its use and found
that in cases where a mapping effort involves significant
research, economic or political decisions or any kind of analysis
where spatial distribution is an important part of a decision
making process, the technique provides what is called the best
linear estimation of the distribution of a particula-r soil or
rock component. The estimate is unique in that it provides a
minimum estimation variance for the available data set and also
allows the researcher not only to develop an isopleth map but an
error map as well. The technique has advantages over other
methods of analyzing spatial data in that trends normally seen in
environmental data (especially pollution data) can be removed
from the analysis. (The term trend applies to those cases where
there is a change in some property along one or more axes of the
study area. Pollution from a leaking landfill would show such a
trend. The further away from the landfill the lower would be the
concentration thus a trend would exist.)
56
-------
Nielsen and his associates (Nielsen et al., 1973; Warrick et
al.r 1977; and Vieira et al., 1981) have used the method to
study the spatial variability of the soil system in an attempt to
develop reliable predictions of water movement through the soil.
Campbell (1978) used the technique in an attempt to delineate the
boundaries between soil types. Burgess and Webster (1980 a/ b)
and Webster and Burgess (1980) present an excellent review of the
technique. Olea (1975) has produced a number of monographs for
the Kansas Geological Survey that provide detailed information on
the technique. Associates of Dr. Olea have produced a computer
package that can be used to conduct kriging and plot the
resulting maps. This program ,called Surface II, is avialable at
a number of computer centers around the U.S.
In those cases where kriging has been used in pollution
control work, it has met with considerable favor. Madeline
Barnes (1978, 1980) has used this technique in a number of major
operations where radioactive materials were being removed from
the environment. The most notable example was the work done on
the Enewetak Atoll in the Pacific Ocean. Nielsen and his
associates (Nielsen , et. al. , 1973; Vieira et. al., 1981) have
used the techniques for studying the variability of a number of
soil water properties.
Kriging has an advantage over other statistical tools in that
it provides not only the means of evaluating the spatial
variability of the soil property but it also provides an estimate
of the variance at each point on the map surface. The main
disadvantage of kriging appears to be the complexity of the
mathematics involved. Although kriqing may be difficult to use
under some conditions, it often makes the most use of limited
available data and thus provides the best answer for the amount
of data available.
The statistical basis for kriging is the theory of
regionalized variables. Kriging attempts not only to estimate
the values of the regionalized variable (the spatially
distributed variable) but also to assess the probable error
associated with the estimates. A variable is considered to be
regionalized if "it varies from one place to another with
apparent continuity, but cannot be represented by an ordinary,
workable function" (Davis, 1973). The theory is not a new branch
of statistics but in reality is an extension of the part of
conventional statistics called time series analysis. It is used
primarily because "conventional statistical approaches are
inadequate for the description of any variable from a natural
phenomenon which has a spatial distribution" (Olea, 1975) — a
common occurrence with soil systems. The intrinsic theory behind
kriging is explained fully by Olea (1975) who reviews the current
status of the method and by Mather on (1969, 1971) who developed
much of the theory.
57
-------
Figure 8.3.1 presents an overview of the procedure for usinq
kriging. The process of the data flow begins with the selection
of sampling sites even though kriging is not a sampling method.
The first step in the use of kriging is to calculate the
semivariance (1/2 of the variance) by use of the following
equation:
li' (8.3.1)
where n = the number of pairs of points on a line
y.. = the semivariance
y.= the value at point x.
y.. + ..= the value at a point j distance from
x. along a line passing through x.
The V(h) values are plotted against the spacing h along the
grid line to give the variogram. There are a number of forms of
the variogram. Those shown in Figure 8.3.2 are found in a number
of the references listed in this report. A number of the
patterns are reported by Barnes (1978,1980). Figure 8.3.2 (a)
and (b) are the two classic forms of the varioqram. All of the
other examples are variants of these two forms. Figure 8.3.2 (a)
shows the range of influence which is the distance over which the
samples are correlated. When the curve fails to pass through the
origin as in (c) or (d) , a "nugget" effect is seen. This results
from sample grid spacing that is too wide to pick up the detail
of the system. (The term originated from the gold fields where
discrete nuggets of gold were found within blocks of ore.) The
limit of the curve in Figures 8.3.2 (a,c,f) is called the sill
and represents the variance of the entire system (i.e. there is
no longer any correlation or dependence of one sample upon
another). Figure 8.3.2 (e) shows pure nugget effect where all
samples are independent of each other. Classical random
statistics should be used in such situations. Olea (1975) notes
that there is an art to fitting the correct model to the
variogram. A number of the models can be tried to find the one
that best fits the type of data being evaluated.
A second important property of regionalized variables is that
of drift. Drift is the trend of the data over the geometric area
of the investigation. If Z (jc) is a regionalized variable and the
drift is m("x) then the residual y("x") is given by equation 8.3.2
(Olea, 1975).
58
-------
LAY OUT SAMPLING LOCATIONS I
COLLECT SAMPLES
ANALYZE SAMPLES
CALCULATE AND PLOT SEMI-VARIOGRAM |
[SELECT SEMI-VARIOGRAM MODEL
DETERMINE THE REGION OR NEIGHBORHOOD
COMPUTE KRIGING WEIGHTS
I KRIGE POINTS ON PLOTTING GRID
PLOT MAP
[CALCULATE ERRORS ON GRID POINTS
|PLOT ERROR MAP AT DESIRED CONFIDENCE LEVEL |
E MORE SAMPLES NEEDED
No
I WRITE REPORT |
Figure 8.3.1 Flow sheet for kriging analysis
59
-------
(b)
7
range
nugget
(e)
nugget
range
Figure 8.3.2 Examples of semi-variograms developed during kriging.
60
-------
y(~x) = Z("x) - m(T) (8.3.2)
A variogram of the residuals is plotted in a manner similar to
Figure 8.3.2. This represents the variogram with the drift
removed from the data. The object of this exercise is to
determine the true variance of the data. Where drift is a major
factor in the data, it may be necessary to go to some form of
multivariate analysis rather than use kriging.
Calculation of the semivariance of the grid data for several
directions will allow the researcher to determine if drift is
present in the data and also allow other anisotropy to be
identified. Anisotropy can often be removed from the data by
transposition of the data.
Kriging also can be used with data that was not obtained from
a systematic grid pattern but the mathematical manipulations of
the data increase considerably. The problem with data not
obtained on some linear pattern is that the variogram cannot be
calculated. If additional information will allow an estimate to
be made of the variogram, it is possible to krige missing points
and thus develop a grid for developing isopleths. This type of
situation is where the method is really an improvement over other
methods of data analysis. The areas not covered by samples are
partially covered by samples lying adjacent to the vacant areas.
The variance of the points kriged provides an estimate of the
precision of any data obtained in this way.
One of the methods that has been used for testing kriging has
been to krige a set of points where data is known. Each actual
data point is removed, a point is kriged then a comparison made
between the actual and the kriged points. Chi-squares can be
used to test the two distributions to determine the fit of the
data. This comparison is also used to calculate the error term
for the kriged values.
This technique is relatively new when compared to classical
statistics; therefore, the methods have not been refined to the
same degree as those of random sampling theory or analysis of
variance. The details of this method are too extensive for
inclusion in this paper. The reader is referred to Olea (1975) ,
Barnes (1980), Campbell (1978), Burgess and Webster (1980 a,b) ,
Webster and Burgess (1980) , and Rendu (1978) for the details of
the method. The Kansas Geological Survey computer programs can
be obtained for a fee at a number of computer centers in the U.S.
The names of those closest to the researcher can be obtained from
the Survey.
61
-------
REFERENCES
Barnes, M. G. 1978. Statistical Design and Analysis in the
Cleanup of Environmental Radionuclide Contamination.
Unpublished M. S. Thesis. University of Nevada, Las Vegas,
Nevada. 52 pp.
Barnes, M. G. 1980. The Use of Kriging for Estimating the
Spatial Distribution of Radionuclides and Other Spatial
Phenomena. In Trans-Stat No. 13. Battelle Memorial
Institute-Pacific Northwest Laboratory. Richland,
Washington. 22 pp.
Beckett, P. H. T., and R. Webster. 1971. Soil Variability: A
Review. Soils and Fertilizers. 34:1-15.
Burgess, T. M. and R. Webster. 1980a. Optimal Interpolation and
Isarithmic Mapping of Soil Properties. I. The
Semi-variogram and Punctual Kriging. The J. of Soil Science.
31(2): 315-332.
Burgess, T. M., and R. Webster. 1980b. Optimal Interpolation
and Isarithmic Mapping of Soil Properties. II. Block
Kriging. The J. of Soil Science. 31(2):333-342.
Campbell, J. B. 1977. Variation of Selected Properties Across a
Soil Boundary. Soil Science Society of America J.
41(3):578-582.
Campbell, J. B. 1978. Spatial Variation of Sand Content and pH
Within Single Contiguous Delineations of Two Soil Mapping
Units. Soil Science Society of America J. 42(3):460-464.
Campbell, J. B. 1979. Spatial Variability of Soils. Annals of
the Association of American Geographers. 69(4):544-556.
Campbell, J. B. 1981. Spatial Correlation Effects Upon Accuracy
of Supervised Classification of Land Cover. Photogrammetic
Engineering and Remote Sensing. 47(3):355-363.
Chapman, H. D., and P. F. Pratt. 1961. Methods of Analysis of
Soils, Plants and Waters. University of California,
Riverside, California. 309 pp.
62
-------
Cline, M. G. 1944. Principals of Soil Sampling. Soil Science.
58:275-288.
Cochran, W. G. 1946. Relative Accuracy of Systematic and
Stratified Random Samples for a Certain Class of Population.
The Annals of Mathmematical Statistics. 17 (2) : 164-177.
Cochran, W. G. 1965. Sampling Techniques. John Wiley. New
York, New York. 413 pp.
Davis, J. C. 1973. Statistics and Data Analysis in Geoloqy.
John Wiley. New York, New York. 550 pp.
Harrison, A. F. 1979. Variation of Pour Phosphorus Properties in
Woodland Soils. Soil Biology and Biochemistry. 11:393-403.
Hindin, E., D. S. May, and G. H. Dunston. 1966. Distribution of
Insecticide Sprayed by Airplane on An Irrigated Corn Plot.
Chapt. 11 in Organic Pesticides in the Environment. A. A.
Rosen and H. F. Kraybill (eds) . Advances in Chemistry
Series #60. American Chemical Society. Washington, D. C.
pp. 132-145.
Krige, D. G. 1966. Two-dimensional Weighted Moving Average
Trend Surfaces for Ore Evaluation. J. of South African
Institution of Mining and Metallurgy. 66:13-38.
Li, J. C. R. 1959. Introduction to Statistical Inference.
Edwards Brothers. Ann Arbor, Michigan. 553 pp.
Matheron, G. 1969. Le Krigeage Universal: Les Cahiers du Centre
de Morphologic Mathmatique de Fontainebleau. Fascicule 1.
1'Ecole Nationale Superieure des Mines de Paris. Paris. 83
PP.
Matheron, G. 1971. The Theory of Regionalized Variables and its
Applications. Les Cahiers du Centre de Morphologic
Mathmatique de Fontainbeleau, Fascicule 5. 1'Ecole
Nationale Superieure des Mines de Paris. Paris. 211 pp.
Mathur, S. P., and R. B. Sanderson. 1978. Relationships Between
Copper Contents, Rates of Soil Respiration and Phosphatase
Activities of Some Histosols in an Area of Southwestern
Quebec in the Summer and the Fall. Canadian J. of Soil
Science. 58(5):125-134.
Mausbach, J. J., B. R. Brasher, R. D. Yeck, and W. D. Nettleton.
1980. Variability of Measured Properties in Morphologically
Matched Pedons. Soil Science Society of America J.
44(2):358-363.
63
-------
Nielsen, D. R., J. W. Biggfar and K. T. Erh. 1973. Spatial
Variability of Field-measured Soil Water Properties.
Hilgardia. 42(7):215-260.
Office of Monitoring Systems and Quality Assurance. 1980.
Interim Guidelines and Specifications for Preparing Quality
Assurance Project Plans. QAMS-005/80. U. S. Environmental
Protection Agency. Washington, D. C.
Office of Occupational Safety and Health. 1979.
for Hazardous Waste Site Investigations. U.
Protection Agency. Washington, D. C.
Safety Manual
S. Environmental
Office of Pesticides Programs. 1976. Sample Collection Manual:
Guidelines for Collecting Field Samples. U. S. Environmental
Protection Agency. Washington, D. C. 39 pp.
Olea, R. A. 1975. Optimum Mapping Technique Using Regionalized
Variable Theory. Kansas Geological Survey. Lawrence,
Kansas. 137 pp.
Petersen, R. G., and L. D. Calvin. 1965. Sampling. Chapt. 5 in
Methods of Soil Analysis. C. A. Black (ed.). American
Society of Agronomy. Madison, Wisconsin. pp. 54-72.
Cuenouille, M.H. 1949. Problems in Plane Sampling.
of Mathematical Statistics. 20:355-375.
The Annals
Rao, P. V., P. S. C. Rao, J. M. Davidson, and L. C. Hammond. 1979.
Use of Goodness-of-fit Tests for Characterizing the Spatial
Variability of Soil Properties. Soil Science Society of
America J. 43 (2):274-278.
Rendu, J. M. 1978. An Introduction to Geostatistical Methods of
Mineral Evaluation. South African Institute of Mining and
Metallurgy. Johannesburg, South Africa. 84 pp.
Richards, L. A. (ed.). 1969. Diagnosis and Improvement of
Saline and Alkali Soils. U. S. Salinity Laboratory. Riverside,
California. Agriculture Handbook No. 60. U. S. Department of
Agriculture. Washington, D. C. 160 pp.
Soil Science Society of America. 1965. Glossary of Soil Science
Terms. Soil Science Society of America Proceedings. 29(3):
330-351.
Soil Survey Staff. 1951. Soil Survey Manual. Agriculture Handbook
No. 18. Soil Conservation Service. U. S. Department of Agri-
culture. Washington, D.C. 503 pp.
Vieira,S. R., D. R. Nielsen and J. W. Biggar. 1981. Spatial
Variability of Field-measured Infiltration Rates. Soil Science
Society of America J. (in-press).
64
-------
Warrick, A. W., G. J. Mullen, D. R. Nielsen. 1977. Scaling
FieId-measured Soil Hydraulic Properties Using Similar Media
Concept. Water Resources Research. 13 (2) : 355-362 .
Webster, R. and T. M. Burgess. 1980. Optimal Interpolation and
Isarithmic Mapping of Soil Properties III. Changing Drift
and Universal Kriging. The J. of Soil Science.
31(3) :505-524.
White, G. C. and T. E. Hakonson. 1979. Statistical
considerations and Survey of Plutonium Concentration
Variability in Some Terrestrial Ecosystem Components. J. of
Environmental Quality. 8 (2):176-182.
Ya te s, F. 1948. Systematic Sampling. Philosophical
Transactions of the Royal Society of London. Series A.
241:345-377.
65
-------
APPENDIX A
SOIL PROTOCOL DEVELOPMENT PROCESS
SECTION A-l
PURPOSE
This appendix is designed to be a guide to assist in the
development of a soil sampling protocol. Extensive discussions
with soil scientists in the field indicate that it is not
possible nor advisable to develop a single protocol that will
attempt to meet all situations. The material that follows allows
the environmental scientist to progress through a protocol
developmental process thus arriving at a protocol that will meet
his particular needs. The four major types of study situations
addressed in Section 4 are addressed here from the point of view
of developing a protocol for use in each type of a setting.
A-l
-------
SECTION A-2
PRELIMINARY BACKGROUND INFORMATION
Every field of research has a basic set of knowledge that the
worker must be able to handle. The follwing types of reference
materials should be read before any attempt is made to sample
soils.
Be conversant with terminology in the following areas:
Basic soil science
Sampling theory
Statistical designs
Environmental monitoring
Quality assurance
Basic toxicology
Basic chemistry
Read the following materials:
Federal Register. Vol. 144:69464. December
3,1979.
Interim Guidelines and Specifications for
Preparing Quality Assurance Project Plans
(OMSQA, 1980) .
Safety Manual for Hazardous Waste Site
Investigations (OOSH, 1979).
The three articles on kriging written by
Burgess and Webster (1980 a, b) and
Webster and Burgess (1980).
The article by Beckett and R. Webster (1971).
The article by Mausbach et al. (1980)
The main body of this report.
A-2
-------
SECTION A-3
DEVELOPMENT OF THE PROTOCOL
Any protocol that will be developed by the environmental
scientist will have the following basic outline:
1. Introduction
2. Objective
2.1 Goals to be met
2.2 Reliability that can be placed on the results
2.3 Precision of the estimates generated
2.4 Resources to be made available
2.5 Schedule
3. Definition of the Magnitude of the Problem
3.1 Known aereal extent of the contamination
3.2 Known vertical extent of contamination
3.3 Chemicals that have been identified
3.4 Known concentrations
3.5 Toxicity of chemicals
3.6 Attitude of community toward problem and
problem solvers.
3.7 Identification of type of pattern
4. Selection of Statistical Design
4.1 Number of samples needed to meet reliability
4.2 Distribution of sample sites
4.3 Frequency of sampling
4.4 Quality assurance
5. Selection of a Sampling Method
5.1 Type of pollution
5.2 Depth of contamination
5.3 Tools
5.4 Determine how samples to be collected
6. Data Analysis
6.1 Basic parameters
6.2 Statistical designs
6.3 Regression
6.4 Spatial distribution
A-3
-------
A suggested way to use this section is to make up a series of
sheets of paper with headings similar to those in the above
outline. When the questions are asked and answered make note of
the answers on the pieces of paper. When you are through with
the exercise you should have a good start on writing your own
protocol.
3.1 Background Information.
Read section 5 (pp. 14-19) of the main body of this report;
then, assemble all of the following kinds of information that are
available.
1. History of the particular study area; its
uses and misuses.
2. Any chemicals that have been placed on the
site.
3. The source of the pollution that you are
attempting to identify.
4. Ownership of the study area both past and
present.
5. Any analytical information on samples
collected in the area.
6. The results of all investigative reports.
7. All geological and soils information on
the study area or the surrounding area.
8. All information on the toxicology of the
pollutants studied.
9. The political and community situation.
10. All existing protocols pertinent to the
sampling program.
11. Copies of USGS quad maps of the area.
12. Copies of all archived aerial photography
of the area.
13. Copies of all current aerial photography.
14. Copies of all documentary photography of
the site.
15. Copies of all reports on the site.
In addition to the above information, visit the study area
and talk to members of the community that are knowledgeable about
the locality and also about the specific area of the study.
Attempt to determine likely control areas at this time. Remember
you are an investigator at this time and are attempting to learn
as much about the study area as you can in order to make your
sampling study not only meet the objectives but also to gain as
much insight into the pollution situation as is possible.
A-4
-------
3.1.1 Questions to Answer.
Attempt to answer the following questions prior to the study.
1. What is the pollution source?
2. What chemicals are in that source?
3. When was it placed there?
4. How was it placed there?
5. Who placed it there?
6. Why was it placed there?
7. What kind of containers was it in?
8. What levels have been found in the
environment?
9. How toxic is it and what are the known
effects?
10. Has anyone locally been hurt by the
pollution? (This is not an epidemiological
study but is to let you know what the
reactions of the community may be.)
11. How far has the pollution gone?
12. What data is available on its migration?
13. Who determined that migration had occurred?
14. Are the data on concentrations and on
migration reliable?
15. What other environmental studies have been
conducted in the area?
16. Is the data available for examination?
17. Will the researchers in the area cooperate?
18. Are there any court actions pending?
19. Are there any administrative actions pending?
20. Is there a potential for litigation in the case?
21. What is the public's awareness of the problem?
22. Are they frightened or angry?
23. Who are the spokesmen for the local population?
3.2 Develop an Objective Statement for the Study.
A. Has an objective statement been issued for the study?
YES - Record it and go to 3.3
NO - Review Section 3 of main report and
then go to 3.2.1
A-5
-------
3.2.1 What is this study attempting to accomplish?
The goals that are needed are explicit, clear statements of
what is to be accomplished by the study. It may be nothing more
than a question such as "How much of the area around the spill is
contaminated?". It may, on the other hand, go to considerable
length to determine the levels of contamination at various
locations within the study area.
Spell out a definite set of questions that you hope to answer
with the study. You should not be hesitant to ask for a more
definitive goal if it is not clear what is being asked.. There
are often cases where the person ordering the study does not know
what is really possible and what is really needed. That is where
the scientist must assist in setting the objectives of the study.
Decide upon your objectives and write them down. Get an
approval on them so that you are sure that you understand what is
being asked and to insure that your results will meet the needs
envisioned by the administrative order.
Write down your final objective under the appropriate part of
the outline then go to 3.2.2.
3.2.2 How close must the estimates be to the real Bean
value.
Scientists often measure the precision of the methods that
they use in a laboratory but they seldom attempt to do so in the
field. This is because of the difficulty in defining the
variation to the point that they can control it. A level of
precision should be delineated for the study. This is usually
expressed as a percentage of the mean.
A second way of evaluating the precision of the data is to
identify the smallest difference between two samples that you
would like to be able to detect. If you are attempting to
delineate variations in the pollution in a large area, you would
probably want a level that was quite small because of the
subtilties of the deposition pattern. On the other hand, in a
situation where you are attempting to determine the area
contaminated by a major spill, you may not need to be able to
detect small differences in order to make the determination that
one area is contaminated and another next to it is not.
A-6
-------
If a number cannot be decided upon a twenty percent value may
not be too far out of line. You are asking that your sampling
determine a value that is within twenty percent of the true mean.
This would be totally unacceptable for a laboratory method but it
reflects the true field situation.
Record your decision and go to 3.2.3.
3.2.3 How reliable must the answers be?
The question of reliability and precision may seem to be the
same. What you are attempting to determine is not how close your
answers are but what is the probability that the answer you get
is the correct one.
Read Section 6.1 of the main report before deciding your
answer to this question. You should also read the section on
simple random sampling (6.2). Statistical sampling requires that
a known confidence level be given. Three are often used by the
scientific community. These are the 68%, 95%, and 99% confidence
levels or significance levels. Another way to state this is to
say that the probability is 0.32 (or about 1 in 3) that the
absolute statistical error exceeds one standard deviation around
the true mean for the 68% significance level; 0.05 (or 1 in 20)
for the 95% confidence level; and .01 or {1 in 100) for the 99%
confidence level.
For the results to be absolute, a 99% confidence level should
be used. When resources are limited or when reliability is not of
paramount importance, the 68% confidence level can often be used.
The important point is to know the confidence level before the
study is conducted.
When there is no basis for a decision, the 95% confidence
level is often used. Record your answer and go to 3.2.4.
3.2.4 What resources are available for the study?
The answer to this question may be developed by an iterative
process. The answer can be a very definite limitation if it is
known before the planning is done but without it the planning is
often unrealistic. You may not be able to answer this question.
This depends upon the administrative structure you operate under.
Your final determination may not come until the planning is
complete and the schedule is known. Resources include not only
finances but also personnel, laboratory capacity and equipment.
A-7
-------
Record your answer and go to 3.2.5.
3.2.5.When must the study be finished?
The schedule is determined by the resources and vice versa.
You will have to set a target date and attempt to meet that date;
therefore, be realistic in setting the date.
Record the target date and go to 3.3.
3.3 Determine the Magnitude of the Problem.
You need to know the magnitude of the problem as it is known
at this time. The information that you must have at this point
covers not only the physical extent of the contamination but also
the concentration range, the toxicity of the chemicals involved,
and the public relations aspects of the problem. This is not a
major data gathering phase of the study but it is a reflection of
the needs that may have to be met if information is totally
lacking. Read Section 4 of the main report before you answer the
following questions.
A. Do you know the study category you will be working in?
YES - Go to Question B.
NO - Go to 3.3.1.
B. Which category will the study area come under?
Large area with pollutant in the surface layers.
Large area with pollutant at deeper layers.
Local area with pollutant in surface layers.
Local area with pollutant in a deep plume.
The category that the study falls into determines the
sampling and statistical designs that should be used. After
making your decision go to 3.4.
3.3.1 What is the known areal extent of the
contamination?
You need to determine the magnitude of the area to be covered.
This may be defined by the type of situation you are addressing.
A-8
-------
For example if you are looking at worldwide fallout from nuclear
testing you are working at one scale; but, if you are working
with a small spill site you are at the opposite end of the
spectrum. There are two categories that you are attempting to
identify:
o Large areas that range from several square miles
to a major portion of the U.S.
o Local areas, such as the area around an accident
site or a landfill site.
Make your determination on the basis of available data and
your own knowledge of the problem.
Record your answer and go to 3.3.2.
3.3.2 What is the vertical extent of the contamination?
Information on this part of the study may be more difficult
to answer than the question in 3.3.1. Examine drilling logs,
investigation reports and information on the behavior of the
pollutants in the soil. Those that are soluble are more likely
to have moved both horizontally and vertically. Examination of
the groundwater flow patterns should tell where the pollution may
have gone. Your designs will depend upon this fact.
Is the pollution a soluble, deeply penetrating chemical?
Is the source located deep in the soil profile?
Is the chemical migrating upward to the surface some distance
from the suspected source?
Has the chemical been at the site for a considerable period
of time?
If you answer yes to any of these questions, then place the
chemical in one of the two deeply penetrating categories. Record
your answer and go to 3.3.3 through 3.3.7.
3.3.3 What chemicals have been identified in the study
area?
3.3.4 What are the concentrations?
3.3.5 Are the chemicals toxic?
3.3.5.1 Are they toxic at the levels seen?
3.3.5.2 Have there been any suspected injuries in the
study area?
A-9
-------
3.3.6 What is the attitude of the community toward the
problem?
3.3.7 Place the study into one of the four classes of
studies listed in 3.3.B and go to 3.4.
3.4 Selection of a Statistical Design.
Each category of study design has an associated group of
statistical designs' that can best be used to address the
situation. This section will attempt to direct you to those that
have promise of working but the final decision should be done in
conjunction with the statistician.
A. Do you know the design that you will use on the study?
YES - Record the information and qo to 3.5.
NO - Go to 3.4.1.
3.4.1 How many samples do you need?
See Section 6.2.1 of the main report for answer. After
determining the number of samples go to 3.4.2.
3.4.2 What is the distribution of the sample sites?
At this point you may want to delay the final decision on
which statistical design you will use until you talk to a
statistician. Basically you are being asked to select between
the simple random design (Section 6.2 of the main report) , the
stratified random design (Section 6.3) and the systematic design
(Section 6.4). The simple random design locates the points at
any location on the study area. The stratifie'd design does the
same thing but it places a portion in each stratum. The
systematic covers the entire area with a grid or with a series of
line transects.
Based upon the objectives of the study and the decisions made
in the paragraphs above determine the sample distribution pattern
that you will use. If you are more interested in a uniform
distribution use a grid design. If you are interested in a
totally unbiased sample use some form of the simple random
design.
A-10
-------
There are a number of variants on each of the designs that
can be provided by your statistician. One that may be used is
what is called a nested simple random design. This is used when
simple random designs are used to locate the sample site but the
location of the samples within that site are fixed by geometry or
some other limit to randomness. This is the case with soils
where the location on the surface is randomly determined but the
below surface samples are fixed by the surface location. This is
not a problem for the field scientist but it is for the
statistician if this fact is not taken into consideration in the
analysis.
Record the information on the design chosen and include
information on the number of replicates that the laboratory wants
along with the number of duplicates your statistician tells you
that he needs for the data analysis in 3.6 below. Go to 3.4.3.
3.4.3 What is the frequency of sampling?
The frequency of sampling depends upon the nature of the
study. Any study attempting to determine seasonal patterns or
patterns in variables that are likely to be associated with the
seasons should be designed in such a manner that samples are
collected at least once each quarter. Long term trends in
ambient soil pollutant levels will only need to be sampled on an
annual basis; but, it must be sampled at the same time each year.
The more detailed the information needed the more often the
soils should be sampled. Those situations where the pollution is
changing rapidly will require more frequent sampling than those
situations where the pattern is stable or only slowly changing.
Select the frequency of sampling that meets the circumstances
of your study. Record the choice made and go to 3.5.
3.5 Select a Sampling Method.
Read Section 7.0 of the main report. From this select a
sampling method that meets your specific need. You must combine
the type of pollution with the depth of contamination to arrive
at the tools that you will use. Once this is done you select the
method by which the samples will be collected. This process
combines the tools with the statistical design and the laboratory
procedures. If the pollution is deep, use one of the methods in
Section 7.3 dealing with underground pollution. If it is shallow
then one of the methods in 7.2 or 7.1 will be useful. The
decision to use compositing will be determined by the data needs
A-ll
-------
of the statistician in order to answer the questions combined
with economics. Use Equations 6.2.1, 6.2.2 and 6.2.3 to assist
in making the decisions. Go to 3.6.
3.6 Data Analysis.
The data analysis will be determined by the design that was
chosen in 3.5. Section 8 of the main body gives a discussion of
the approaches. You need to determine the approach that best
fits your needs. Approach the problem from the point of view of
the user of the data. The statistician will probably use one of
the statistical computer program packages to analyze the data.
The outline that is now filled in should provide you with a
fairly detailed set of information that you can incorporate into
the final protocol.
A-12
-------
APPENDIX B
SAMPLING PROTOCOL FOR SURFACE SOILS
SECTION B-l
OVERVIEW
The purpose of this protocol is to provide an example of a
protocol for use in sampling surface soils. The procedures
outlined here are not intended to be the final solution to a soil
sampling protocol but are only a guide for use by those
scientists undertaking to develop a protocol for some particular
use. The protocol can be used in cases where the chemicals of
interest are expected to be located on the soil surface as would
be the case in an accident or chemical spill.
The procedure used in this protocol is simple and fast. Ten
15 cm. long subsamples are to be taken from the surface layer at
each sampling station with an Oakfield type tube sampler. The
ten subsamples will be composited and sampled for analyses. A
triangular grid is used to insure complete coverage of the area.
B-l
-------
SECTION B-2
SAMPLING DESIGN
A triangular patterned systematic sampling design will be
used to collect the samples. Ten subsamples will be collected at
each of the intersections of the grid lines. At ten percent of
the stations, three sets of samples will be taken for analysis.
The choice of ten percent was purely arbitrary and has no
particular significance. The number can be better selected when
input from the analytical laboratories is available. The data
collected will be analyzed by kriging, if possible. If it is not
possible to use kriging, the method of adjacent differences will
be used to determine the variation within the grid.
2.1 Minimum number of samples
Calculate the minimum number of samples to collect using
equation B-2.1. In order to use the equation you must have the
following pieces of information:
• Significance level you desire to use
• Coefficient of variation in %
• Percentage error you will allow
• t-tables
The values are entered into the equation. It is not possible
to obtain the t-value without knowing n. The approach used is to
assume a value for n then look in the t-table for the t-statistic
at (n-1) degrees of freedom. Calculate a new value for n. This
value is then used to pick a new t-value. The process is
repeated until the starting and ending values for n are the same.
Use the next highest integer if fractional values of n result
from the calculations.
B-2
-------
2.1.1 The Equation
n = (CV)2t2/p2 (B-2.1)
Where n = the number of samples; CV = coefficient of variation in
%; p = the percentage of error you will allow; and t is the
two-tailed t-value obtained from t-tables for the significance
level at (n-1) degrees of freedom.
If values are not known for the variables in the
equation, a first approximation can be used. Use 95%
significance level, 65% coefficient of variation and 20 % for the
error you will allow in the numbers. These latter two values may
seem high. They are intentionally chosen on the high side to
provide a margin for error. The 65% CV is not out of line with
data from a number of soils studies. The 20% margin of error is
also not unreasonable for soils work. Using these values gives
an n of 43. The closest triangular spacing to the n of 43 would
be 37 samples. This number is determined from Table B-2.1.
This number would reduce the significance level to approximately
93%. This value was determined by inserting the n of 37 into
equation B-2.1 and solving for the t value. This calculated t
value was then compared with the t-table values in order to
determine the level of significance.
If it is essential to have 95% confidence or better then the next
highest triangular grid would be chosen. This grid yields 61
sample points. If this number is used, the significance level
would be approximately 98%.
2.2 Grid Layout
The triangular grid consists of equilateral triangles laid
out to form a hexagon. The design can be elongated if desired
but the hexagon is the easiest to determine the number of samples
needed to meet a particular area configuration. The hexagon has
an equal number of line segments on each side. This fact can be
used to assist in determining the number of samples in the array.
Figure B-2.1 shows the grid and its coordinates. The number of
sample points for each grid size is shown in Table B-2.1.
B-3
-------
• Sample point
Figure B-2.1. Triangular grid design showing coordinate numbering.
B-4
-------
Table B-2.1. Number of Sampling Points in Triangular Grids,
Number of Points Number of Samples
on a Side
2 7
3 19
4 37
5 61
6 91
2.2.1 Selection of a Starting Point
Select a starting point for locating the lower left corner of
the hexagon by using either the UTM coordinates on a map or by
using a grid overlay. The procedures outlined in Section 6.2.2
of the main body of the report can be used to locate the starting
point for the grid. After the starting point is selected, the
triangular grid is oriented over the area to be studied in such a
manner that one of the nodes is superimposed on the starting
point selected above. Samples will be collected at each of the
nodes where the lines cross.
2.2.2 Grid Size
The length of the sides of the equilateral triangles is to be
determined by examination of the expected pollution area. This
may be determined by a pilot study or by examination of aerial
photographs and by site visits. The grid should extend beyond
the boundaries of the polluted area if it is possible to make a
determination of the approximate location of the boundaries be fore
the sampling is done. A set of grids drawn on plastic with
different map scales can greatly facilitate the use of this
design. The location of the starting point determined in Section
2.2.1 determines where the sampling points will fall but the
orientation and the size of the hexagon may have to be determined
by trial and error using the plastic overlays. Move the overlay
until the desired grid alignment is obtained keeping in mind that
one of the nodes must be located on the starting point.
B-5
-------
The size of the hexagon can be determined by an alternate
procedure. First, estimate the area of the study by placing a
square over the known or suspected contamination. This area (A)
is entered into equation B-2.2 to determine the length (1) of the
sides of the hexagon formed by the triangular grid.
A = 2.598 (I2) (B-2.2)
Use of this equation will allow the edges of the resulting
hexagon to extend beyond the boundaries of the contamination
because the border of the hexagon will fall outside of much of
the square used to estimate the area of contamination. Using the
size of grid calculated, prepare a plot plan on a map of the area
using the starting point as one of the nodes of the grid.
The grid nodes are identified by a number-letter combination
as is shown in Figure B-2.1. The east-west rows will be lettered
and the diagonal rows will be numbered.
2.2.3 Physical Location of the Grid in the Field
Locate the starting point identified in Section 2.2.1 on the
ground using a map or aerial photograph of the site. Locate an
east-west line through that point. This line should correspond
to one of the lines of the grid that you placed on the map of the
site in Sections 2.2.1 and 2.2.2. Measure east along the line
the distance (1) calculated with equation B-2.2 or scaled from
the map. This is the first sampling location. Continue this
process until the appropriate number of points are located along
this line. Mark each sample point with a stake. Using a metal
tape measure or surveyor's chain, strike an arc from each of the
points. Another point is located at the intersection of the
arcs. This point is staked. The process is continued until the
entire hexagon has been staked out on the ground. The number of
sample points located on each side of the hexagon is given in
Table B-2.1. Photograph the plot upon completion of staking.
2.3 Compositing of the Samples
At each grid node, take ten individual 15 cm cores with a
standard Oakfield tube sampler. These cores are to be placed in
a stainless steel pan, then mixed with a stainless steel scoop
until the soil appears to be relatively homogeneous. Take small
samples from each quarter of the pan. Remix the remaining soil.
B-6
-------
Take a second set of subcomposites adding these to the first
subcomposite. Continue this process until 250 g of soil is
obtained. Dispose of any waste at an appropriate disposal site.
2.4 Replicates
At 10% of the sampling points take triplicate sets of cores
(i.e. 3 groups of 10 each); composite these as separate samples
— DO NOT MIX AS ONE COMPOSITE. The same procedure will be used
for each composite as was described in Section 2.3 above. These
samples will be used to obtain a measure of the variation within
each grid. In order to insure uniform coverage with the
replicates, the hexagon will be divided into six segments
determined by running a line from each corner. This will give
six triangles from which to take the triplicate samples. Ten
percent of the samples in each small triangle will be randomly
selected for replication. Fractional numbers will be rounded to
the next higher number.
2.5 Controls and Background Samples
Collect enough samples from an uncontaminated area to equal
20% of the total samples collected in the sampling grid. These
samples will be used to establish an environmental baseline and
to test the limits of the contamination measured in the grid.
(This sets an analytical limit to determine which of the grid
samples are contaminated and which are not) .
2.5.1 Location of Controls
Locate the control samples upwind from the site in an area of
similar soil types and similar land use to the study area.
Aerial photographs available in the county SCS office can be used
as a starting point for identifying where the best control sites
are likely to be found. The candidate areas should be visited
and an at tempt made to determine the history of the area. You are
looking for an area that has as many characteristics in common
with the study area as is possible. The ideal control area would
be one with no source of pollution but all soil and land use
characteristics the same. This is seldom possible so the field
team leader must determine the best compromise for the area.
Where time permits, a number of pilot study samples could be
taken from the candidate control areas.
B-7
-------
Care must be exercised when taking these samples to insure
that the owners and neighbors understand that the control area is
not believed to be contaminated; on the contrary you believe that
it is not contaminated. Let them know that you are collecting
the samples for a baseline. This precaution avoids a lot of
unnecessary anxiety in the community where a spill has occurred.
In rural areas, the county extension agent or the soil
conservationist can greatly assist in finding cooperative land
owners and in reducing the fears of the local community.
B-8
-------
SECTION B-3
SAMPLE COLLECTION ON THE TRIANGULAR GRID
Sampling of shallow soils is simple and straightforward. The
approach presented here is designed to provide the researcher
with the basic tools for collecting the samples.
3.1 Sampling for Non-volatile Chemicals
At each sampling point lay out a circle with a diameter of
either 3 meters or 1/4 of the side of the small triangle,
whichever is smaller. Beginning at the northernmost point on
this circle, take ten, equally spaced, 15 cm soil cores around
the circle using the Oakfield tube sampler Model 22G. Place
these cores in the mixing pan and follow the procedure outlined
in Section 2.3. If ten cores do not provide an adequate sample,
increase the sampling density around the circle. The composited
samples are to be placed in precleaned, glass containers fitted
with precleaned teflon or aluminum foil lid liners. Use the
procedures outlined in Section 7.8 of the main body of this
report for cleaning the glassware and the tools used.
When rock or large roots are encountered, move the sampler
several inches to one side or the other. Select a direction and
follow it during the sampling effort. A new, decontaminated
sampling tube is to be used at each sampling location. The
mixing bowl, stainless steel scoop, spoon etc., are to be
decontaminated between each sampling point. See Section 7.8 of
the main report for details on this decontamination procedure.
3.2 Sampling for Volatile Chemicals
Volatile chemicals create a special problem. Samples
containing volatiles must be collected in a form that will
prevent loss from volatilization. Water samplers use what is
called a "headspace jar". These do not work well with soil
because it is impossible to exclude all air space with the cores.
B-9
-------
Compositing speeds up volatilization therefore it cannot be used
as a sampling method with these chemicals. The following
procedure is used in lieu of the "headspace jars".
3.2.1 Soil Sampler
One of two types of samplers can be used to obtain cores from
the soil. The U.S. Army Corps of Engineers tube density sampler
drives a 5.1 cm (2 in) diameter tube into the soil to collect a
soil sample. The tubes are machined to a predetermined weight
and inside diameter. These tubes can be reused and are small
enough for handling during surface sampling efforts.
The second sampler is the short 45.7 cm (10 in ) split spoon
with ring liners. The spoon with liners can be driven into the
soil with a sledge hammer or with a sample jack such as the
Soiltest Hydraulic Forta-Sampler . The split spoon retains the
soil core inside of a brass liner. A 15 cm long ring liner
should be used for the sample. (This same system can be used
where a depth profile is desired.) A series of ring liners of
desired thickness are included in the split spoon. The soil in
each ring is analyzed separately.
3.2.2 Collection of Sample
The tubes, either split spoon or density sampler are forced
into the soil to the desired depth. The tubes are extracted and
sealed with a precleaned teflon cap, then tightly wrapped with
duct tape and coated with a non-contaminating sealing compound.
Three samples will be taken at each location.
3.2.3 Transport and Analysis
The sealed soil-tube unit will be shipped to the laboratory
in locked ice chests cooled with dry ice. They will be stored in
a cooler maintained at 4 C. The tubes will not be opened until
time for extraction of the organic chemicals. The tube will be
opened and the soil extruded into the extraction vessel or else
placed in a tube furnace where the volatile organics can be
driven off at the appropriate temperatures. With these
relatively short tubes it may be possible to place the soil and
the split spoon liner into the tube furnace without having to
extrude the soil.
B-10
-------
SECTION B-4
RECORDS, SECURITY AND SAFETY
4.1 Records
Maintain a logbook of the operations alonq with all
appropriate site description forms, photographs and maps showing
the sampling locations and the collection sites. All samples are
to be tagged with NEIC tags and chain-of-custody forms are to be
filled out by the sample collector and accompany the samples
until the analysis is complete.
4.2 Security
Samples are to be in the physical'possession of the sample
collector or within his immediate view at all times. If they
must be left in a vehicle, the vehicle is to be locked at any
time the sample collector is away from the vehicle. The NEIC
chain-of-custody procedures are to be rigidly adhered to at all
t ime s.
4.3 Safety
The chemicals of concern to the environmental sampler are
usually toxic. Some are extremely hazardous. Prior to extensive
sampling under this protocol, the field team should determine the
nature of the chemicals present and provide adequate safety
measures. Gloves should be worn at all times. Make sure that
the gloves and safety clothing worn will stand up to the solvents
used in decontamination. If volatile chemicals are involved, it
may be necessary to wear charcoal filter masks such as the
American Optical Organic Vapor Respirator. NEIC and the EPA
Safety Officer can advise on the best equipment to use. OSHA and
NIOSH can also assist in determining the proper equipment if
there are any questions on what should be worn.
B-ll
-------
4.4 Site Restoration
Experience has shown that often there can be some terrain
damage to the study area from vehicle traffic. This damage must
be repaired by the sampling crew immediately following the field
work. This should be done within a couple of davs of the sample
collection in order to avoid problems with the land owners.
B-12
-------
SECTION B-5
DATA ANALYSIS
The results of the analysis will be subjected to a series of
statistical tests. The following is a suggestion. The final
tests should be determined after close consultation with the
statistician. Details of the methods are not discussed here
because standard reference materials and computer packages such
as Stat Pac can. be used to conduct the analysis. The steps are
listed below.
Calculate the mean and standard
deviation for the control and the
polluted area data.
Use the t-statistic to determine a
confidence interval for the background.
Use the 95% significance level.
Determine which samples are above the
background level.
Plot the location and concentrations of
the samples on a map.
Delineate the contaminated area.
If isopleths are desired, conduct
kriging.
Plot isopleth map.
Plot error map.
Calculate the error mean square for
differences between adjacent points.
B-13
-------
SECTION B-6
STAFFING, EQUIPMENT AND SUPPLIES
The sampling effort outlined in this protocol requires a
minimum of two people. A third person can greatly facilitate the
work and should be included if possible. Two members lay out the
grid and collect the samples. The third member takes the
photographs, prepares the site description and map and handles
all tagging and record keeping.
6.1 Equipment and Supplies.
10 to 12 Oakfield tube samplers,
Model 22-g obtained from Soil Test,
Inc.
Borebrush for cleaning.
10 to 12 ten-quart stainless steel
mixing bowls.
A U. S. Army Corps of Engineers tube
density sampling set with 30 to 40
six-inch sample tubes.
Safety equipment as specified by
safety officer.
One-quart Mason type canning jars
with Teflon liners (order 1.5 times
the number of samples. Excess is for
breakage and contamination losses.).
A large supply of heavy-duty plastic
trash bags.
Sample tags.
Chain-of-custody forms.
Site description forms.
Logbook.
Camera with black-and-white film.
Stainless steel spatulas.
Stainless steel scoops.
Stainless steel tablespoons.
Caps for density sampling tubes.
Case of duct tape.
100-foot steel tape.
2 chain surveyor's tape.
Tape measure.
Noncontarninating sealant for volatile
sample tubes.
Supply of survey stakes.
Compass.
Maps.
B-14
-------
Plot plan.
Trowels.
Shovel.
Sledge hammer.
Ice chests with locks.
Dry ice.
Communication equipment.
Large supply of small plastic bags
for samples.
B-15
-------
APPENDIZ C
SAMPLING PROTOCOL FOR CONTAMINANT PLUME
SECTION C-l
OVERVIEW
This protocol can be used for those situations where
contaminated groundwater has moved from the ooint of Deposition.
The contaminated soils are likely to be located at considerable
depth below the surface. The soil scientist is not only
interested in the horizontal spatial pattern but also the
vertical pattern. Sampling in this situation requires power
equipment such as truck mounted augers and coring devices. The
cost of sample collection becomes an overriding consideration.
Litigation is a definite possibility; therefore, extreme care
must be taken to insure the integrity of the samples.
Samples are to be collected with a 60.9 cm (24 in) split spoon
sampler operated in conjunction with a 20.3 cm (8 in) auger.
This procedure allows discrete samples to be collected from the
various strata found below ground level. The nature of the
subsoils is such that compositing is not used except in rare
cases. The layers of soil provide avenues for migration of the
pollutants. These should be sampled individually rather than in
a composite. Compositing can be used if the soil is very
homogeneous, a rare occasion in most soils. The decision to
composite should be made only after preliminary coring has
determined that the soils are in fact homogeneous. Sampling in
soft soils will require retainers in the end of the split spoon.
A rectangular or square grid is recommended because of the
plume. Random samples often fail to reveal the presence of
highly contaminated plumes unless a very high sampling intensity
is used. The grid overcomes the problems encountered with random
samples by covering the entire area of suspected contamination.
C-l
-------
SECTION C-2
SAMPLING DESIGN
A rectangular grid pattern is recommended. The long axis of
the rectangle should be located along the axis of the plume or
suspected plume. Investigation of well logs and consultations
with geologists will determine the direction of groundwater flow,
thus allowing the plumes general direction to be determined.
Samples should also be collected on the upstream side of the
source. The author has observed situations when two phases of
chemical pollutant would move in opposite directions.
Exploratory drilling prior to initiation of the study will
provide the necessary information to determine if there is a need
for a more complete grid pattern around the total circumference
of the source.
A preliminary study is recommended even when there is no
reason to suspect that there may be a large area of
contamination. This allows the statisticians to develop a
complete picture of the data needs before the study is begun.
The preliminary study also is a means for the laboratories to
prepare to receive the samples and to have a chance to obtain
standards and work out potential problems with the analyses
before the main load of samples comes into the laboratory.
2.1 Minimum Number of Samples
The large equipment needed for the deep sampling efforts
requires considerable cost; therefore, the total number of
samples will be controlled by the budget for the study. The
resources available for the study therefore must be committed
before the study planning is begun in any great detail. The
equation presented below (equation 6.2.3 of the main body of the
report) can be used to calculate the number of samples.
c ' co + nCs + nCa
C-2
-------
The costs are for the total cost, fixed costs, sampling costs and
analytical costs respectively. The equation can be rearranged to
yield the number of samples as follows:
K — i n — f* \ /1 i" * P ^
n — \ v. v, I / ( v- T v_ )
o s a
Once the number of samples is determined, the calculated n value
is entered into equation 6.2.2 which is then solved for the
t-value. The rearranged equation appears below.
t =-/(n p2/ CV2)
The terms are as described in the main body of the report on page
25. The calculated t-value is compared with the t-value in the
statistical tables. Interpolate to find the level of
significance or probability that the number of samples will
provide. The allowable error (or the percentage difference that
you desire to detect) , (p) , can be taken as the confidence
interval on a background sample set, or the detection level of
the method of analysis, or a known sampling error for the types
of situations encountered during plume sampling.
2.2 Grid Layout
The sampling grid should be aligned with the axis of the
suspected plume. Where information is not available for
determining this, the axis should be elongated with the direction
of groundwater flow. The number of grid points should closely
fit the number of samples calculated in Section 2.1 above. An
attempt should be made to determine the approximate limits of
contamination. This can usually be done in conjunction with the
hydrogeologists that are usually involved with any study dealing
with underground plumes. The rate and extent of migration can be
used as a rough estimate of the extent of contamination. The
size of the grid cells then will be determined by dividing the
total area of suspected contamination by the allowable number of
samples.
The coordinates of the grid will be those of any X - Y grid
system. An appropriate numbering scheme .can be assigned to the
grid lines to aid in locating sampling sites and in coding tags
and site description forms. The starting point for the grid is
located by the same random process that is discussed in Section
6.2.2.
C-3
-------
2.3 Control Area
An appropriate "control site located in an uncontaminated
portion of the area should be obtained for collecting control
samples. The soil structure should be as nearly identical to
that in the study area as can be found in the immediate vicinity.
Care must be taken to insure that there is no communication
underground between the control sites and that the control site
is not located down wind from any industrial source of the same
chemicals. Candidate control areas should be sampled prior to
the final selection to insure that there are no unsuspected
sources of pollutants likely to cause problems in the analysis
and interpretation phases. This preliminary sampling is
especially important around major industrial areas where past
practices may have contaminated an area.
2.4 Preliminary Study
Many "plume hunt" studies do not have site specific
historical data available for use in planning the study. The
systematic sampling grid used in this study allows the researcher
to work up to the optimum number of samples in stages. This
phased approach is recommended if the scheduled completion date
will allow the time for these phases to be carried out. Use of
equation 6.2.2 can be used to determine the reliability of some
percentage of the total number of samples. The particular qrid
nodes sampled can be randomly selected if there is a desire to
sample in a random fashion but a subset of data points can be
obtained by using a coarser grid made up of every other grid line
or every third grid line.
This use of the grid designed for the main study allows
repeated sampling to be done if there is a desire to determine if
there are seasonal variations in the data and also allows for
kriging analysis to be done on a portion of the samples. Where
kriging is done an error map can be generated that will show
where additional samples are needed to reach the precision
desired.
Where the preliminary study is truly exploratory (i.e., there
is no data available other than visual or olfactory evidence of a
problem) the use of some form of surrogate analysis is desirable.
For example, in some cases total organic halide analysis can be
used to locate the plume at a considerably reduced cost when
compared to an analysis like the gas chromatographic-mass
spectrographic analysis required for the standard Priority
Pollutant Analysis series. These surrogate tests can often allow
the outer limits of contamination to be found in the field..
C-4
-------
SECTION C-3
SAMPLE COLLECTION
Sampling for deep lying plumes of contaminated soil is
expensive and at times quite difficult to perform. The approach
used here is essentially the ASTM method D1586-67.
3.1 Sampling Equipment
A truck mounted drill rig equipped with A-rod drilling
equipment or adaptable to the A-rod connector will be used.
Standard 5.1 cm (2 in) split spoon samplers that are 61 cm (24
in) long will be used to collect samples. Samples collected for
volatile organic chemical analyses will be contained in brass
liners unless stainless steel liners can be obtained. The brass
thin walled liners will be used for all samples unless
compositing is done for some specific reason. A 20.3 cm (8 in)
diameter soil auger will be used to excavate to the sampling
depth. All samples will be placed in precleaned one-quart Mason
canning jars fitted with a precleaned teflon lid liner. Dry
sands, or soft noncohesive soils will require that the split
spoon be fitted with a retainer or pocket shoe.
3.2 Non-Volatile Pollutant Sampling
3.2.1 Sample extraction
The split spoon will be attached to the drill rod and forced
into the soil to the full depth of the spoon. In difficult
soils, a 63.5 kg (140 Ib) hammer will be used on the drill rig to
drive the spoon into the soil. If the soil is loose, wet or in
any way unconsolidated, use a basket retainer in the split spoon.
C-5
-------
The spoon is then extracted and turned over to the sampling
crew. The drill crew will then attach the auger and drill down
to the depth of the first sample's penetration ( i.e., to two
feet). The soil will be shoveled back from the hole face. The
hole will then be cleaned with the auger by increasing the
revolutions as the auger is lowered and returned to the surface.
A second split spoon will be attached and forced into the bottom
of the hole formed by the auger. This sample will be extracted
and given to the sampling crew. The process will be continued
until the desired depth of sampling is reached. If large rocks
are encountered it may be necessary to shift the sample point to
a second location. Move the rig one meter north. If this does
not work move one meter south of the original hole. The exact
location should be noted in the log book and on the core log.
The blow counts for the hammer will be recorded on the core log.
3.2.2 Sample Preparation
The split spoons will.be opened by the sampling crew. The
cores will be carefully split lengthwise with a stainless steel
spatula. The color, texture and any unusual features of the core
will be noted. Any evidence of chemical contamination will be
recorded in the logbook and on the core log sheet. The length of
the core should be measured as well as the depth of any textural
changes or unusual features present in the core. A standard tape
measure will be used for these measurements.
The sample will then be transfered to a precleaned, labeled,
glass canning jar fitted with a teflon liner placed next to the
sample. An NEIC sample tag will be filled out for each sample.
The tag will accompany the sample to the laboratory. The outside
of the jar will be cleaned then it will be double bagged with the
tag placed in the outer bag.
Any obvious potential routes of migration such as sand
lenses, silt layers or old root channels should be sampled
separately. These layers are often the first areas to become
contaminated and therefore provide an early warning of future
problems. Any marked changes in the texture should be sampled as
a separate unit. If the number of samples acquired is excessive,
every second or third split spoon sample can be taken for
analysis.
C-6
-------
3.3 Volatile Chemical Sampling
Samples collected for volatile chemicals are often difficult
to acquire in a condition suitable for analysis. The less
disturbance the greater the chance that the analysis will be
meaningful.
Split spoons can be fitted with liners made of brass,
stainless steel and in some cases, Teflon. The samples cannot be
described on the well log so a considerable amount of
interpretive information is lost. The core log will be made from
observations of the material removed by the auger. The procedure
for taking the sample is the same as that presented above in
Section 3.2. The extracted core will be left in the liner and
shipped to the laboratory for analysis. The ends of the liner
tube will be sealed with a teflon cap, then taped with duct tape
and sealed with a non-contaminating sealant. The samples will
then be transported to the laboratory in locked ice chests
maintained at 4 C with dry ice. The details of sample handling
are outlined in the Federal Register (Federal Register. Vol
44:69464. December 3, 1979).
3.4 Security
All samples will be logged, tagged and entered on the
standard NEIC chain-of-custody forms. The samples will be
maintained either under constant surveillance or locked in a
limited access storage area. If it is necessary to place the
samples in a vehicle, the vehicle will be kept locked when
unattended. The amount of time samples are left unattended in a
vehicle should be kept to a minimum. Tfie chain-of-custody form
will be signed when samples change hands. The team leader is
responsible for the samples and should insure that they are
turned over to a responsible party before relinquishing custody
of the samples.
3.5 Safety
Underground plume samples often contain highly toxic
chemicals. Extreme care should be taken when handling the
samples. Follow the EPA safety manual procedures during
sampling. It is better to overprotect than to have someone hurt
because of a lack of diligence to pursue safety procedures.
Protective gloves are a minimum protection for all members of the
crew that must handle the soil samples. Crews sampling around
sources of volatile organic chemicals should have proper fitting
vapor masks available at all times. There are some situations
C-7
-------
where the safety officer may require the masks to be worn at all
times. In areas where the chemical concentrations are known to
be low, the masks may be carried and donned only when chemicals
are detected by odor or sight. All auger holes are to be
refilled with either grout or clean soil.
3.6 Decontamination
Decontamination is a major problem. The augers and split
spoons require a pressurized hose and decontamination facility
that can handle hazardous wastes. The augers and split spoons
are to be pressure washed, scrubbed then rinsed with tap water.
If organics are involved, the split spoons are to be rinsed in
waste acetone, then methylene chloride. This is followed by a
distilled water rinse, a spectrographic grade acetone rinse, then
a spectrographic grade methylene chloride rinse. Inorganic
chemical sampling can eliminate the solvent rinses. All tools,
etc., must be cleaned following the same procedure as the
sampling equipment. Save all waste for disposal at a licensed
hazardous waste landfill.
C-8
-------
SECTION C-4
DATA ANALYSIS
The addition of another variable, depth, further complicates
the analysis of the data. Regression techniques often provide
the best method for handling this type of sampling data
especially in the early stages of an investigation. The
preliminary data should be used to determine the optimum number
of samples if cost is not a major factor. Kriging can help
determine the location of any additional samples that are needed.
Careful examination of the preliminary data will often allow the
researcher to exclude some samples because of the homogeneity
and/or thickness of a particular layer. (This statement is based
upon the fact that the more homogeneous the media, the fewer the
number of samples that are required to arrive at a conclusion
with a predetermined precision.)
The depth variable requires that kriging be conducted in
three dimensions. This has not been done although it should be
possible. The mathematics would be quite difficult. An
alternative would be to krige the data at each layer where
samples were collected. (This assumes that sampling was done at
a common depth throughout the entire soil mass.) By
superimposing one on top of each other the volume of the plume
can be observed.
Regression is to be used to determine the variables that are
influencing the migration of the pollutants and to make
comparisons between layers in the soil and between locations in
the study site. The t-test can be used to develop a confidence
interval for comparing the contaminated and uncontaminated areas.
Use kriging to develop isopleth and error maps for the area.
C-9
-------
SECTION C-5
STAFFING, EQUIPMENT AND SUPPLIES
5.1 Staffing
• Drill crew of three men.
• Sampling crew of three men (two sample
collectors and one record keeper).
• One runner to acquire supplies and handle
dec on tarn in a t ion.
5.2 Equipment and Supplies
• Drill rig equipped with A-rod fittings or
adaptors and an 8 inch auger.
• 25 two-foot sections of 2 inch split spoon
tubing.
Brass or stainless steel liners.
Teflon caps for liners.
Shovels.
Drums for hauling waste material.
Grout and clean soil for refilling holes.
Spatulas.
One-quart Mason jars ( use 1.5 times the
number of samples expected.)
Safety equipment.
Compass.
Surveyor's chain.
Survey stakes.
Hammer.
Case of duct tape.
Non-contaminating sealant.
All core logs, notebooks, etc.
Chain-of-custody forms.
Tags.
Ample supply of small plastic bags to hold
sample bottles.
Ice chests with pad locks.
Dry ice.
Heavy-duty plastic trash bags.
Communication equipment.
C-10
------- |