EPA/600/2-85/104
September 1985
PRACTICAL GUIDE FOR GROUND-WATER SAMPLING
"by
M.J. Barcelona, J.P. Gib"b, J.A. Helfrich, and E.E. Garske
Illinois State Water Survey
Department of Energy and Natural Resources
Champaign, IL 61820
Cooperative Agreement
*
Project Officer
Marion R. Scalf
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
Robert S. Kerr Environmen|al Research Laboratory
Office of Researehfand Development
U.S. Environmental Protection Agency
' Ada, Oklahoma 7-4820
Printed on Recycled Paper
-------
DISCLAIMER
The information in this document has been funded wholly or in part
by the United States Environmental Protection Agency under assistance
agreement number CR-809966-01 to the Illinois State Water Survey,
through the Board of Trustees of the University of Illinois. It has been
subject to the Agency's peer and administrative review, and it has been
approved for publication as an EPA document.
-------
FOREWORD
The U.S. Environmental Protection Agency was established to coordinate
administration of the major Federal programs designed to protect the quality
of our environment.
An important part of the Agency's effort involves the search for
information about environmental problems, management techniques and new
technologies through which optimum use of the Nation's land and water resources
can be assured and the threat pollution poses to the welfare of the American
people can be minimized.
EPA's Office of Research and Development conducts this search through a
nationwide network of research facilities.
As one of the facilities, the Robert S. Kerr Environmental Research
Laboratory is the Agency's center of expertise for investigation of the soil
and subsurface environment. Personnel at the laboratory are responsible for
management of research programs to: (a) determine the fate, transport and
transformation rates of pollutants in the soil, the unsaturated zone and the
saturated zones of the subsurface environment; (b) define the processes to be
used in characterizing the soil and subsurface environment as a receptor of
pollutants; (c) develop techniques for predicting the effect of pollutants on
ground water, soil and indigenous organisms; and (d) define and demonstrate
the applicability and limitations of using natural processes, indigenous to
the soil and subsurface environment, for the protection of this resource.
This report contributes to that knowledge which is essential in order
for EPA to establish and enforce pollution control standards which are
reasonable, cost effective and provide adequate environmental protection for
the American public.
Clinton W. Hall
Di rector
Robert S. Kerr Environmental
Research Laboratory
m
-------
-------
CONTENTS
Page
FIGURES Vll
TABLES IX
ACKNOWLEDGEMENTS X
EXECUTIVE SUMMARY XI
SECTION 1. INTRODUCTION 1
Literature Overview 2
Ground-Water Sampling and Quality Assurance . . 4
Elements of the Quality Assurance Program ..... 7
Objectives 8
Sampling Quality Control 8
Analytical Quality Control . ......... 11
Representative Ground-Water Sampling 16
Criteria for Documenting Representative Sampling ...... 19
Accuracy, Precision, Detection/Quantitation Limits
and Completeness 21
SECTION 2. ESSENTIAL ELEMENTS OF A GROUND-WATER SAMPLING PROGRAM 24
Hydrogeologic Setting and Sampling Frequency 25
Hydrogeologic Setting .... ....... 25
Sampling Frequency . 3,3
Information Needs and Analyte Selection ....... 37
Parameter Selection 39
General ground-water quality parameters 40
Pollution indicator parameters 41
Specific chemical constituents. . . 42
Minimal Analytical Detail for Ground-Water Monitoring
Programs 45
Detection monitoring data set 45
Assessment monitoring data set . . . . 46
Well Placement and Construction 47
Drilling and Well Completion Methods 48
Hollow-stem continuous-flight auger 49
Solid-stem continuous-flight auger 51
Cable tool 51
Air rotary 52
Air rotary with casing hammer 53
Reverse circulation rotary ... 54
Mud rotary 54
Bucket auger ..... 55
Jetting ............. 56
Driving ..................... 56
Monitoring Well Design 56
Depth of the well 58
Diameter of monitoring wells ..... 63
Size of screen 65
Grouts and seals 66
Multiple-completion wells . 67
Well or sampling point documentation ........... 67
-------
Page
Well Development, Hydraulic Performance and Purging Strategy. . 72
Well Development 72
Techniques for high hydraulic conductivity wells 73
Techniques for low hydraulic conductivity wells 76
Hydraulic Performance of Monitoring Wells 77
Water level measuring techniques 78
Steel tapes 78
Electric drop lines 79
Pressure transducers .... 80
Hydraulic conductivity testing methods 80
Slug tests • 80
Pumping tests 84
Analysis of water level data 84
Well maintenance procedures 85
Well Purging Strategies 89
Pumping rates 90
Evaluation of purging requirements 90
Sampling Mechanisms and Materials 95
Sampling Mechanisms . : 96
Recommendations for selecting sampling mechanisms 99
Sampling Materials , 100
Subsurface conditions and materials affects 100
Recommendations for selecting sampling materials 104
Sample Collection Protocol . . . . 105
Water Level Measurement 109
Purging Ill
Sample Collection . . 112
Filtration . 118
Field Versus Laboratory Determinations 120
Blanks, Standards and Quality Assurance . 122
Sample Storage and Transport 124
SECTION 3. RECOMMENDED SAMPLING PROTOCOLS 128
The Basis for Sampling Protocol Development . . 128
Sampling Protocol for-Detection Monitoring 130
Analyte .Selection and Sampling Procedures 132
Assessment Monitoring .... 135
Field Sampling Procedures 146
Sampling equipment setup, well inspection and water
level measurement 147
Verification of the well purging requirement 150
Sample collection/filtration •. 151
Field determinations 154
Sample storage and transport 154
SECTION l|. CONCLUSIONS 156
SECTION 5. RECOMMENDATIONS -. 161
SECTION 6. REFERENCES 163
V1
-------
FIGURES
Number
1 .1
1.2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
2.14
2.15
2.16
Steps in ground-water sampling and sources of error . .
Steps in water sample analysis and sources of error . .
Occurrence and movement of ground-water through
a) porous media, b) fractured or creviced media,
c) fractured porous media ,
Local and regional ground-water flow systems in
humid environments
Temporary reversal of ground-water flow due to
flooding of a river or stream
Typical ground-water flow paths in arid environments
Total porosity and drainable porosity for typical
geologic materials
Type of plume generated from: a) a slug source or
spill, b) an intermittent source, and c) a continuous
source
Resulting change in a capture area due to regional flow
Sampling frequency nomograph
Well placement and flow paths at low water levels ...
Drilling log sheet ......... < •
Monitoring well construction diagram
Page
10
14
27
28
29
30
32
35
36
38
62
68
71
Schematic diagram of an air-driven well development device 75
. 83
Hvorslev piezometer test (a) geometry,
(b) method of analysis
Effects of waste-handling activity on
ground-water flow paths
Percentage of aquifer water versus time for different
transmissivities
87
92
Generalized ground-water sampling protocol 110
vn
-------
Number
Page
2.17 A well-head instrumentation package for Eh, pH,
conductivity and temperature measurements ........ 113
2.18 Suggested recording format for well purging and
sample collection ................... . 114
2.19 Sample chain of custody form .............. 127
3.1 Generalized flow diagram of ground-water sampling steps . 129
3.2 Matrix of sensitive chemical constituents and various
sampling mechanisms ................... 131
3.3 Recommended sample collection methods for detective
monitoring programs ................... 136
-------
TABLES
Number
1.1
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
3.1
3.2
3.3
3.4
3.5
Data Requirements for Water-Source Definition and
Aquifer Representation of Ground-Water Samples
Page
18
Recommended Drilling Techniques for Various Types
of Geologic Settings ....... 50
Well Casing Material Specifications and Depth
Recommendations *..... 64
Data Needed for a Monitoring Well Construction Diagram . 70
Performance Evaluation of Ground-Water Sampling
Mechanisms 101
Relative Sample Contact Comparison for Selected
Materials 103
Recommendations for Rigid Materials in Sampling
Applications (In decreasing order of preference) .... 106
Recommendations for Flexible Materials in Sampling
Applications (In decreasing order of preference) .... 108
Inorganic Sample Log (Filtered Samples) 116
Organic Sample Log (Lab Filtered, If Necessary) .... 117
Field Standard and Sample Spiking Solutions 125
Recommended Analytical Parameters for Detective
Monitoring 133
Recommended Sample Handling and Preservation Procedures
for a Detective Monitoring Program 139
Metallic Species in RCRA Appendix VIII Which Require
Only Metal Determinations 144
Equipment for Field Sampling ....... 148
Sample Purging Parameter Readings 152
IX
-------
ACKNOWLEDGEMENTS
The authors appreciate the advice and support of the staffs of the
USEPA R. S. Kerr Environmental Research Laboratory (Ada, OK) and the
Environmental Monitoring Systems Laboratory (Las Vegas, NV). The
comments and suggestions of several reviewers were very helpful in the
preparation of this document. The work was also supported by the con-
tributions of effort and time of Steven Heffelfinger, Michael O'Hearn,
Mark Sievers, Pamela Beavers, and Pamela Lovett of the State Water.
Survey. The work has also been made possible by: the Campus Research
Board of the University of Illinois, a large number of material and pump
suppliers, as well as the support of our colleagues and families.
-------
EXECUTIVE SUMMARY
Ground-water monitoring is a complex undertaking. Cost effective
monitoring relies on careful planning and critical reading of the scien-
tific literature. These activities will insure that the application of:
well-placement, construction, sampling and analytical procedures, result
in the collection of high quality data. The information needs of each
program must be recognized and all subsequent monitoring network design
and operation decisions must be made in light of the available data. In
this sense, monitoring is an evolutionary process which should be
refined as the information base expands.
Routine monitoring efforts may be sustained for decades. There-
fore, it is unreasonable to follow preliminary guidance offered for
generalized monitoring activities as the data base for a specific situa-
tion is developed. Therefore, high quality hydrologic and chemical data
collected in the detection phase of monitoring are essential to planning
future activities. Effective monitoring efforts 'are both dynamic and
flexible. Our present understanding of natural and contaminated sub-
surface conditions is developing, but incomplete.
The practical elements of a viable long-term ground-water
monitoring effort include:
Evaluation of hydrogeologic setting and program information
needs
Proper well placement and construction
Evaluation of well-performance and purging strategies; and the
-------
Execution of effective sampling protocols which
include the appropriate selection of sampling
mechanisms and materials, as well as sample
collection and handling procedures.
Proven ground-water monitoring procedures are in a state of rapid
development at the present time. It is prudent to specify monitoring
methods and results which will permit the collection of high quality,
representative information for the most sensitive chemical constituents
of interest. All methods used in a specific situation should be care-
fully documented so that one can learn as the information needs and
dimensions of the monitoring effort mature.
Volatile organic compounds, redox or pH sensitive chemical con-
stituents are problematic chemical constituents which place significant
demands on monitoring efforts. It is clear that, given properly con-
structed and maintained sampling points, sampling and handling methods
which minimize sample disturbance are the most cost-effective means
available to provide high quality ground-water information. Positive
displacement, no gas-contact sampling mechanisms constructed of appro-
priate inert materials (Teflon(R) > stainless steel > other plastics or
ferrous materials) provide the basis for an effective monitoring effort.
Actual sampling and analytical performance (accuracy, precision,
detection and quantitation limits) which ensure the collection of water
originating from the formation of interest should be established in
every monitoring effort, regardless of the specific information needs of
individual programs. This can best be assured by the implementation of
quality assurance and quality control measures which are both checked
and documented carefully. The current state of our understanding of
xit
-------
effective monitoring procedures requires that common sense also play a
large part in planning ground-water sampling efforts.
If the practical recommendations of this guide are put into prac-
tice, we will have a much improved information base available in the
future. This will be essential to making wise decisions on ground-water
rehabilitation or other remedial actions as well as to improving our
knowledge of dynamic ground-water systems.
xi i i
-------
-------
SECTION 1
INTRODUCTION
The need for reliable ground-water sampling procedures has been
recognized for years by a variety of professional, regulatory, public
and private groups. The technical basis for the use of selected
sampling procedures for environmental chemistry studies has been devel-
oped for surface water applications over the last four decades.
However, ground-water quality monitoring programs have unique needs and
goals which are fundamentally different from previous investigative
activities. The reliable detection and assessment of subsurface con-
tamination situations require that minimal disturbance of geochemical
and hydrogeologic conditions occur during sampling. At this time
field-proven well construction, sampling and analytical protocols for
ground-water sampling have been developed for many of the more problem-
atic chemical constituents of interest. However, the acceptance of
these procedures and protocols must await more careful documentation and
strong Agency recommendations for monitoring program execution. The
time and expense of characterizing actual subsurface conditions places
severe restraints on the methods .which can be employed. Since the tech-
nical basis for documented, reliable drilling, sample collection and
handling procedures is in the early stages of development, conscientious
efforts to document method performance under real conditions should be a
part of any ground-water investigation.
This guide provides the elements of effective ground-water sampling
for routine applications. This is not to minimize the ongoing
development of specific sampling or in situ sample collection methods
for research purposes. It is important, however, that essential
-------
elements of reliable sample collection and handling are understood so
that the eventual development and application of more sophisticated
methods can be based on high quality data.
We proceed from the point of view that the placement of wells for
sampling access has been done appropriately and the task at hand is to
construct the wells and to collect water samples representative of the
formation of interest. The sampling procedures described in this guide
are recommended on the basis of long-term reliability in routine moni-
toring programs.
LITERATURE OVERVIEW
Much of the literature on routine ground-water monitoring methodol-
ogy has been published in the last ten years. The bulk of this work has
emphasized ambient resource or contaminant source monitoring rather than
case-preparation or enforcement efforts. General references which are
useful to the design and execution of sampling efforts are those of the
U.S. Geological Survey (1,2), the U.S. Environmental Protection Agency
(3.^,5,6) and those of a number of other groups (7,8,9). In large part,
these past works treat sampling in the context of overall monitoring
programs providing descriptions of available sampling mechanisms, sample
collection and handling procedures. The impact of specific methodol-
ogies on the usefulness or reliability of the resulting data have
received relatively little discussion (10,11).
Routine monitoring data is used most often to determine if any
deterioration in water quality has occurred over time. In principle,
this information will accurately represent hydrogeologic or geochemical
conditions at a site and enable an understanding of the dynamics of sub-
-------
surface systems. A certain level of knowledge must be achieved to
insure the success of a detective monitoring program and to plan modifi-
cations or refinements of the monitoring program if contamination is
indicated. Otherwise, poor decisions may result which will prove to be
far more expensive and time-consuming than the careful performance of
proper detective monitoring activities would have been.
High-quality chemical data collection is essential in ground-water
monitoring programs. The technical difficulties involved in representa-
tive sampling have been recognized only recently (10,12). It is clear
that the long-term collection of high quality ground-water chemistry
data is more involved than merely selecting a sampling mechanism and
agreeing on .sample handling procedures. Efforts to detect and assess
contamination 'Can be extremely unrewarding without accurate (e.g.
unbiased) and precise (e.g. comparable and complete) concentration data
on ground-water chemical constituents.
Gillham et al. (13) have published a very useful reference on the
principal sources of bias and imprecision in ground-water monitoring
efforts. Their treatment is extensive and stresses the minimization of
random error which can enter into well-construction, sample collection
and sample handling operations. They further stress the importance of
collecting precise data over time to maximize the effectiveness of trend
analysis, particularly for regulatory purposes. Accuracy is also very
important, since the ultimate reliability of statistical comparisons of
results from different wells (e.g. upgradient versus downgradient
samples) may depend on differences between mean values for selected con-
stituents from relatively small replicate sample sets.
-------
GROUND-WATER SAMPLING AND QUALITY ASSURANCE
Individual ground-water sampling and analytical events yield
results which provide a snapshot picture of hydrogeologic and chemical
conditions at a monitoring site. When the results of successive events
are assembled properly, they enable one to better understand the
nature, extent and degree of subsurface contamination. It is important
to remember that hydrologic and chemical conditions vary in both time
and space and that the subsurface environment of ground water is
dynamic. Therefore, sampling frequency and the location of discrete
sampling points must be considered carefully to resolve the temporal and
spatial distributions of ground-water contaminants.
Each ground-water sample must be collected so as to insure the
reliability of analytical determinations. Also, accurate and precise
measurements of water level and hydraulic conductivity must be made so
that the analytical results can be interpreted with consideration of the
hydrogeologic system.
Achieving the information needs of a ground-water sampling program
over a specified time period requires careful planning and execution of
the sampling design. Careful planning is particularly crucial to dis-
tinguishing between the actual hydrologic and chemical variability at a
site and that which may arise from errors in the sample collection,
handling, and analysis procedures. Each field measurement and water
sample collected for laboratory analysis should also be representative
of the discrete sampling point within the sampling network. Emphases
are often placed on quality control and quality assurance for chemical
analysis alone. One should keep in mind that there is no substitute for
high quality sampling and field measurements.
4
-------
A high quality set of hydrologic and chemical data is accurate,
precise, comparable, and complete. Also, data must be collected at a
minimum level of sensitivity and completeness to satisfy the information
needs of the sampling program. Accuracy, precision, sensitivity, and
completeness are measures of sampling and analytical performance. The
accuracy of each concentration datum is the measure of its closeness to
the true value. Accuracy is normally expressed as an average of a
number of measurements to the true value. The accuracy of analytical
procedures may be assessed by the use of standard reference materials.
In this case, accuracy is expressed as the percentage of the ratio of
the measured value to the true value. For environmental samples, where
the true value is frequently unknown, accuracy is reported as bias (or
the percent recovery minus 100) established by internal or surrogate
standard techniques. Generally, values of bias in excess of ±20%
indicate systematic error or a problem with sampling or analytical
procedures.
The precision of a data set is a measure of the probability that a
measurement will fall within certain confidence limits. Precision is
frequently expressed as the standard error (sx) of the mean value (x) of
a set of replicate determinations (n) at a stated mean (or true) value.
The standard error is related to the standard deviation (s) by the
expression: sx = s * /n- Increasing the number of replicates at an
established level of precision will generally improve the level of con-
fidence (reduce random error) in the data. Duplicate sample values
which differ by less than ±50% relative difference indicate good error
control.
-------
Sensitivity is a term which relates to both the limit of detection
(LOD) and the method detection limit for a particular chemical constitu-
ent. The method detection limit pertains to the lowest concentration of
a particular chemical constituent which can be measured reliably in a
sample. The LOD is the lowest concentration level which can be deter-
mined to be statistically different from a blank. A practical guideline
is to set the LOD for a specific constituent at a level equivalent to
three standard deviations (expressed in mass or concentration) above the
blank. This level establishes a threshold for qualitative or "trace"
detection sensitivity and provides a degree of confidence in values
reported as "less than" a detectable concentration. More stringent
criteria for quantitation set the limit of quantitation (LOQ) at 5 or 10
standard deviations above the blank to insure that quantitation is on a
sound foundation. Regardless of the convention used, it is important
that the LOD and LOQ be reported with all data sets at least for certain
problematic chemical constituents. Completeness of the total planned
data set should include the performance parameters defined above.
Sampling and analysis procedures contribute to the overall quality of
the data set and documentation of control over both systematic and
random .error is central to the effort.
The crucial elements of planning a ground-water sampling effort are
discussed in detail in this guide. High quality data collection
requires strict adherence to proven well construction, sampling and
analytical protocols developed with due precautions against bias, impre-
cision, contamination or chemical alteration of the water sample. In
this respect all field measurements attendant to water sample collection
-------
are considered part of the sampling protocol. Quality control proce-
dures built into sampling and analytical protocols will guard against
the loss of data by minimizing both systematic and random error.
ELEMENTS OF THE QUALITY ASSURANCE PROGRAM
A quality assurance (QA) program is a system of documented checks
which validate the reliability of a data set. QA procedures are used to
verify that field and laboratory measurement systems operate within
acceptable limits. These limits should be determined during sampling
program design for each measurement which the program requires. The
limits may be modified or refined as new information is gathered.
However, a documented basis for evaluating the need for modification
must be established if the expense and manpower involved in ground-water
investigations is to yield cost-effective, high quality data.
The QA program should be implemented as a set of basic measurement
procedures and corresponding quality control checks (6). The overall
effectiveness of the quality control checks in reducing errors should be
audited by a person or technique .outside of the normal sampling and
analytical operations. In this way the QA program will ensure that
quality control (QC) procedures are followed on a daily basis to:
reduce variability and errors, identify and correct measurement
problems, and provide a documented statistical measure of data quality.
The effectiveness of the overall program demands that all personnel are
aware of the QA/QC requirements for the investigation and that the
quality control objectives are understood.
-------
OBJECTIVES
Sampling Quality Control
An understanding of the specific characteristics of the study site
is required to plan effective QC checks. Generally, this understanding
is achieved in phases which must be recognized by the sampling program
manager. Each sampling datum represents a single opportunity to collect
data from a sampling point which can rarely be retrieved if errors are
not identified.
A minimal data set consisting of selected field measurements and
sample volume recovery must be agreed upon to comprise a "sample." Then
a minimum completeness or data recovery level should be defined which
will adequately characterize existing conditions and fall within
expected limits of future variability.
It must be kept in mind that even with adequate QA auditing of
sample results within control limits, there are system constraints on
the subsequent interpretation of sampling and analytical information.
Hydraulic and hydrologic properties are, to some extent, scale dependent
and ground-water monitoring is frequently conducted in geologic forma-
tions which are not aquifers. Further, solution chemical properties are
only part of the subsurface geochemical system. These and other unique
characteristics of ground-water systems may introduce systematic error
or bias into monitoring data sets. Gillham et al. (13) have addressed
many of these potential problems.
Effective QC procedures for ground-water sampling should be based
on proven field measurement and sampling procedures. The wide variety
of hydrogeologic and geochemical conditions of interest for contaminant
monitoring have been investigated by an equally diverse combination of
-------
procedures. Very few of these procedures have been standardized after
systematic development and controlled evaluation trials. Therefore,
tailoring QC procedures to the situation at hand is a complex task.
Well construction and development techniques as well as sampling proce-
dures, mechanisms, and materials all have the potential to introduce
errors into monitoring results. These sources of error should be con-
sidered in the development of QC checks.
Given that the ground water may be under relatively high partial
pressures of nitrogen or carbon dioxide, water samples need to be
handled very carefully. The samples also originate in geologic media
which are rarely isotropic at the regional to local scale. Frequently,
suspended solids accompany water sample collection which can seriously
affect analytical results. The discussions provided by Sisk (6) and
Brown and Black (14) are useful in planning general QC procedures for
ground-water sampling efforts.
A common challenge to effective ground-water data quality control
is that the accuracy of a sample result is difficult to judge, since the
true value is. unknown. Accuracy of individual measurements must there-
fore be judged by the analysis of a reference material or by spiking the
sample with a known quantity of analyte followed by reanalysis. The
results from field blanks and standards may then be compared to the
results of laboratory standards and spiked samples to gain confidence in
the accuracy of sample analyses. The precision of measurements within a
data set is thus defined as the average agreement between repeated
measurements on samples and standards. Quality control over the first
four steps involved in sample access and retrieval is difficult to
achieve. This is shown schematically in Figure 1.1. Therefore, it is
-------
Step
Sources of Error
In-Situ Condition
Establishing a Sampling Point
Field Measurements
Sample Collection
Sample Delivery/Transfer
Field Blanks, Standards
Field Determinations
Preservation/Storage
Transportation
Improper well construction/
placement; inappropriate
materials selection
Instrument malfunction;
operator error
Sampling mechanism bias;
operator error
Sampling mechanism bias;
sample exposure, degassing,
oxygenation; field conditions
Operator error; '
matrix interferences
Instrument malfunction;
operator error;
field conditions
Matrix interferences;
handling/labeling errors
Delay; sample loss
Figure 1.1. Steps in ground-water sampling and sources of error
10
-------
very important to choose well construction and sampling protocols which
are simple and minimize disturbance in order to collect accurate data.
One can readily observe that the integrity of both the sampling
point and sampling mechanism are as critical as operator expertise to
minimize the error or variance introduced into the sample results.
Decisions made in establishing a sampling point and the choice of
sampling mechanisms can introduce significant systematic error (bias)
into all subsequent sample results which may go undetected without care-
ful QA auditing of the data as soon as possible. Further, documented
sampling QC checks and QA audits are controlling factors in the useful-
ness of the analytical data. The laboratory can only be expected to
reliably report data based on the samples, field standards, and blanks
as received.
The potential sources of error noted in Figure 1.1 define essential
elements of sampling quality control. These are:
1) Proper calibration of all sampling >and field measurement
equipment
2) Assurance of representative sampling, particularly with respect
to site selection, sampling frequency, well purging and
sample collection
3) Use of proper sample handling precautions
Analytical Quality Control
Laboratory quality control is necessary to ensure valid analytical
results. Analytical QC procedures must be developed in parallel with
those involved in the sampling operation. Whether the laboratory
analyses are made by an in-house or' contract lab, the value of blind
11
-------
control samples and blanks submitted as "normal" samples is enormous.
Blind control samples may be prepared solutions or ground water spiked
with the contaminants of interest at known concentrations. Blind
controls provide the only true check on the accuracy of analytical
results. Effective QC procedures provide daily checks that the analyti-
cal system is in statistical control. Blind control samples and
multiple determinations should be emphasized wherever possible. Repeat
sampling and analysis is a poor second choice to performing the tasks
adequately in the first place. The variables involved in sampling must
be controlled to the maximum extent possible for the rigors of labora-
tory QC procedures to be meaningful. Three useful references for
planning QA and QC for ground-water data collection are contained in
reviews by Nacht (15) and Keith et al. (16), and Kirchmer (17,18).
The need to establish a measure of confidence in the analytical
results is underscored in a formal laboratory QA program. The program
should address three main functions: the control, determination and
documentation of data quality. These are minimal criteria for effective
laboratory QC, which should extend to field determinations. Regardless
of the analytes of interest and the degree of sensitivity required by
the information needs of the ground-water sampling program, every labo-
ratory should adhere to well-documented control procedures. These pro-
cedures have been reviewed in general by Dressman (19) and Dux (20).
The expectations which may be anticipated from contract laboratory
services are no less rigorous than those of in-house laboratories.
Specifics of such cooperative sampling/analytical arrangements have been
covered by Kingsley et al. (21) and Kingsley (22).
12
-------
In contrast to the steps in the sampling protocol, analytical
quality control is straightforward, provided that the analytical labora-
tory staff is made aware of any unusual attributes of the samples. This
.type of feedback can substantially improve the validity and interpretive
value of measurement results.
The steps of an analytical protocol are normally quite specific to
the individual analytes of interest. The planning, of comprehensive QA
procedures should be done carefully with each individual step in the
analytical protocol taken into account. In general, the analytical
protocol can be depicted as shown below in Figure 1.2. Appropriate QA
audits of the QC measures at each step should serve to keep potential
analytical errors in control.
Instrument malfunctions, analyst errors, and the use of "aged", old
or deteriorated standards pose problems that can be detected and cor-
rected with good QA/QC procedures. More difficult obstacles arise from
the application of "standard" methods to the analysis of highly contami-
nated samples. Matrix or direct interferences are among the most diffi-
cult sources of error to bring under control (23). Thoroughgoing QC
requires that standard methods be validated for the most difficult
sample matrix encountered within a particular set of samples. Valida-
tion by internal standardization techniques should be done over the
entire range of concentration represented in the sample results (24,25).
The necessary elements of an effective laboratory QA program are:
1) Adherence to documented laboratory QC procedures, including:
proper calibration of instrumentation, verification of daily
standardization and analytical performance parameters
(accuracy and precision) for all procedures, daily analysis
13
-------
Stec
Samples, from Storage
Field Blanks and Standards
Sources of Error
"Aged" samples; loss of
analytes; contamination
Subsampling
Procedural Standards
Analytical Separation
Analysis
Reference Standards
Sample aging/contamination
in lab; cross-contamination;
mishandling/labeling
"Aged" standards;
analyst error
Matrix interferences;
inappropriate/invalid
methodology; instrumental
malfunction/analyst error
Matrix interference;
inappropriate/invalid
methodology; instrumental
malfunction/analyst error
"Aged" standards
Calculations
Results
Transcription/machine errors;
sample loss in tracking system;
improper extrapolation/inter-
polation; over-reporting/
under-reporting errors
Figure 1.2. . Steps in water sample analysis and sources of error
14
-------
of sample replicates, standards, spiked samples and blanks
by approved methodologies and the use of QC charts to docu-
ment the validity of laboratory results
2) Participation in round-robin or interlaboratory studies
3) Prompt recording, storage and retrieval of laboratory results
with the corresponding analytical performance parameters
The development of a total QA/QC program for ground-water sampling
and analysis must be approached carefully. The care exercised in well
placement and construction, and sample collection and analysis, however,
can pay real dividends in the control of systematic errors. Repeated
sampling and field measurements will minimize the' effect of random
errors induced by field conditions or system malfunction.
The responsibility for the selection of reliable sampling and
analytical methods is to ,some extent shared by the sampling program
director and the client or agency in need of the information. As more
high quality data become available, QA/QC planning will be facilitated
for environmental sampling programs. The American Chemical Society
Committee on Environmental Improvement has published a valuable refer-
ence for reporting data quality (e.g. accuracy, precision, LOD, LOQ,
sensitivity) in a consistent format for monitoring purposes (26). This
guide contains recommendations based on experience and published
results. It will be revised and modified accordingly as the information
base grows. Therefore, it should be used in conjunction with the future
amendments of existing standard procedural documents (27,28,29).
15
-------
REPRESENTATIVE GROUND-WATER SAMPLING
Representative sampling is probably most difficult in situations
where reliable data is needed most (30). Chemists have struggled for
decades with the difficulties involved in obtaining representative ana-
lytical results from bulk solid or natural water samples. Scientists
who have worked with environmental samples fully appreciate these diffi-
culties. Statisticians, on the other hand, hold exact views concerning
the characteristics of representative samples. Statistically, a repre-
sentative sample is a subset of a set (or universe called the popula-
tion) which has the average characteristics of the set. For
ground-water samples, one must assume that such a sample is representa-
tive of the aquifer or geologic formation from which it came. It
follows then that the results of representative sampling and controlled
analytical determinations provide an accurate measure of the in situ
condition at the time of sampling. Claassen (31) has demonstrated that
an approximation of the representativeness c>f a ground-water sample
alone is achievable given the complexities and costs involved in exhaus-
tive investigations of the subsurface. Verification of the extent of
representativeness is thus the responsibility of project staff.
The goal of representative sampling is a relatively straightforward
undertaking in materials' analysis or investigations of well-mixed homo-
genous surface-water bodies. Sources of error or variance in sampling
or analysis should be independently verifiable if the measurement
systems are in statistical control. This is possible if truly random
sampling can be conducted and invalid samples can be identified through
the use of controls and blanks.
16
-------
Representative ground-water sampling, however, is limited to repli-
cate discrete samples from established sampling points which may accu-
rately and precisely reflect the average properties of the measured
system. Sampling accuracy, however, cannot be unequivocally verified in
the field. It is vitally important that the limits of the measured
system are understood by the project personnel responsible for the
interpretation of the data (12). In this way, the interpretation of
"high" or "background" levels of specific chemical constituents will be
consistent with the hydrogeologic system description. Statistical
theory and manipulations applied to data on hydrogeologic or geochemical
systems cannot substitute for expert judgment.
Claassen (31) pointed out that there exists a marked scale depend-
ency of the heterogeneity of aquifer systems. He suggested that most
aquifers are microscopically (-100 urn) heterogenous, some are homogenous
on a somewhat larger scale, while all are probably heterogenous on a
regional scale (km). His publication details suggested guidelines for
evaluating aquifer representation which should be carefully considered
in planning ground-water investigations of all types. Data requirements
for water source definition and aquifer representation of ground-water
samples are listed in Table 1.1. This data should be recorded for each
sampling point and updated after each scheduled well maintenance (e.g.
redevelopment operation). The well pumping history, in particular,
should be updated on each sampling date to insure that any deterioration
in well performance can be fully documented.
Hydrologists and geochemists have made progress towards the resolu-
tion of these problems of scale for aquifer representation. The work of
Ingamells (32) and Ingamells and Switzer (33) is notable in this area of
17
-------
Table 1.1. Data Requirements for Water-Source Definition and
Aquifer Representation of Ground-Water Samples
(Modified after Claassen, reference 31)
A. Drilling history
1. Well depth and diameter
2. Drill-bit type and circulating fluid
3. Lithologic data from cores or cuttings
1i. Well-development before casing
5. Geophysical logs obtained
B. Well-completion data
1. Casing sizes, depths and leveling information relative to
both land surface and top of casing
2. Casing material(s)
3. Cemented or grouted intervals and materials used
1. Plugs, stabilizers, and so forth, left in hole and
materials used
5. Gravel packing: volume, sizes, and type of material
.6. Screened, perforated, or milled casing or other intervals
which allow water to enter the borehole
7. Pump type, setting, intake location, construction
materials, and pump-column type and diameter
8. Well maintenance record detailing type of treatment and
efficiency
C. Well pumping history
1. Rate
2. Frequency
3. Static and pumping water levels
D. Estimation of effect of contaminants introduced into aquifer
during well drilling and completion on native water quality
E. Effect of sampling mechanism and materials on the composi-
tion of ground-water sample
1. Addition of contaminants
2. Removal of constituents
a. Sorption
b. Precipitation
c. Degassing
18
-------
research, however its practical application to hydrogeologic problems
has been limited. An inventive technique for resolving scale and heter-
ogeneity problems in aquifer representation has been reported by Keely
(34). Briefly, a combination of pumped wells or pumping wells and moni-
toring wells are sampled over a time series simultaneous with water
level and yield measurements. The combined chemical time series samples
and the drawdowh results provide a data set which describes the spatial
variability of dissolved chemical constituents, as well as aquifer
transmissivity and storage values. The application of this technique
to a contamination problem in Washington State yielded encouragement for
its use and refinement for future work (35). Multi-level sampling point
arrays also hold promise for the resolution of scale problems. However,
most of the published reports are limited to demonstrations of tech-
niques (36,37,38). Systematic evaluations of the performance of sam-
pling protocols for chemical constituents are rare.
Criteria for Documenting Representative Sampling
It should be evident that representative sampling in the strict
statistical sense is a challenging undertaking. To some extent the
criteria for "representativeness" depend on the level of detail required
in the program. The requirements for documenting representative samples
from the measured system will vary from site to site and perhaps from
sampling point to sampling point, depending on the situation under
investigation. This document defines representative sampling a priori
as representative for the specific purposes of the ground-water investi-
gation. In the case of regulatory compliance studies, the criterion for
19
-------
representativeness may be that which will be considered by the appro-
priate agency to be representative of the regulated facility. For
example, charge balance considerations and minimum acceptable accuracy
and precision limits for the determination of the contaminants of
interest are useful criteria for representative samples.
There are two sets of essential requirements for representative
sampling. The first set of criteria must be based on some knowledge of
the measured system and the experience of project planning staff. Close
attention must be paid to the requirements listed in Table 1.1, as well
as the potential impacts of: well placement, sampling frequency, the
mobility and persistence of chemical constituents and natural sources of
variability in the hydrogeology and geochemical characteristics of the
site. These criteria are subjective to some extent and evaluation of a
data set's "representativeness" may only be possible after extensive
preliminary investigation. As the level of detail involved in a sampl-
ing program increases, one must be careful to avoid excesses in borings
for core collection or well installation. Every disturbance of the
subsurface has the potential to contribute to contaminate migration and
confound data interpretation. Good detective work on site character-
istics and operational history can minimize the cost and disturbance of
extensive sampling activities.
The second set of criteria addresses the details of the sampling
and analytical protocols. They are based on the assumption that a
properly designed and executed ground-water sampling plan will enable
documented evaluation of the significance of the sample mean and the
variation between the mean and other members of the set. Basically,
20
-------
reliable protocols provide a known level of confidence in the represen-
tativeness of the sample.
Accuracy, Precision, Detection/Quantitation Limits and Completeness
The critical performance parameters common to both the sampling and
analytical protocols are accuracy, precision, minimum detection limits
and completeness. Proper planning of a comprehensive sampling program,
which includes QC check and QA auditing procedures to insure high
quality results, requires that each step in the protocols is evaluated
for each of the performance parameters. The most direct way to meet
this requirement is to specify and document the sampling protocol for
the most sampling error prone class of chemical constituents of
interest. In each class, certain constituents may require refinement of
the protocol for reliable sampling. Detailed documentation of accuracy,
precision and minimum detection limits for the corresponding analytical
procedures should be provided as well. In this manner sampling errors
can be evaluated independently from those involved in the analytical
work.
Establishing the performance of the sampling protocol to achieve
error control requires the execution of a controlled sampling experi-
ment. If possible, one should seek to verify sampling accuracy and
precision over the potential concentration range of the most sensitive
chemical constituent of interest. This type of experiment could estab-
lish the lowest practical level of a chemical constituent which can be
sampled within certain accuracy and precision limits. This minimum
"collectable" concentration would correspond to the LOQ for analytical
operations. However, sampling accuracy cannot be verified in the field,
21
-------
since the "true" or in situ value is unknown and it is most, unlikely
that any single (or average) value for a particular chemical constituent
could be considered as the "true" even for very localized sites.., There-
fore, the accuracy of the sample retrieval and collection steps, which
involve both the sampling mechanism and materials, must be evaluated in
controlled laboratory experiments. These experiments should simulate
field conditions and maintain a known concentration source of the most
sensitive chemical constituent of interest. Precision, on the other
hand, can be evaluated in the field or the laboratory if a sufficient
number of replicate determinations can be performed.
There have been few controlled sampling experiments reported which
provide supporting data for the evaluation of representative sampling
performance. Field experiments have been limited to documenting
apparent discrepancies in accuracy by different sampling techniques
(11t39), or studies which establish the precision of developing sampling
techniques (40). Since it is extremely difficult to maintain control
over sampling performance which may be largely operator dependent, the
choice of a specific sampling mechanism must be made very carefully. If
a sampling mechanism is chosen which has not been subjected to con-
trolled performance testing, the user should provide documentation which
assures control over mechanism related error. It may be that evalua-
tions of the accuracy of sampling mechanisms must be inferred by com-
parisons with published data and the precision should be established for
each study with a well designed sampling experiment. Thorough consid-
eration must be given to sources of systematic (bias) and random
22
-------
(imprecision) error at each step in the sampling protocol. The sampling
mechanism is of particular importance in this regard as it largely
determines the complexity of the sampling protocol.
23
-------
SECTION, 2 . , .«,
ESSENTIAL ELEMENTS OF A GROUND-WATER SAMPLING PROGRAM
The technical literature on ground-water sampling provides a great
deal of information on selected aspects of an efficient sampling
program. However, valid data on reliable methods-for drilling, well
completion/development, and sampling, reactive or organic chemical, con-
stituents in ground water are scarce.
Recommendations for conducting . ground-water sampling programs
stress the use of "appropriate'1 drilling - and sampling methods or
material's choices which will permit the collection of representative
samples. This leaves many critical decisions open to discretion when
data on the hydrogeologic setting or dissolved chemical constituents may
be incomplete. This section provides specific recommendations for
establishing a sampling point and conducting a sampling effort which
should be sufficient to the needs of most routine ground-water investi-
gations. In many cases, the detail and precautions which must be con-
sidered in planning a representative sampling effort cannot be predicted
until a substantial amount of high, quality data is made available by
preliminary sampling.
Due care to insure the collection of unbiased, precise hydrologic
and chemical data should be exercised from the outset in all monitoring
efforts. The data set should then be subjected to constant scrutiny and
reevaluation as the situation becomes better defined. This approach is
logical and cost-effective. Poorly conceived or "cook-book" .sampling
programs will ultimately end up generating poor data at considerable
24
-------
long-term expense. The logical, phased-'approach also facilitates regu-
latory review of the data and decision-making for assessment or remedial
actions.
HYDROGEOLOGIC SETTING AND SAMPLING FREQUENCY <
The hydrogeologic conditions at each site to be monitored must be
evaluated for the potential impacts the setting may have on the develop-
ment of the monitoring program and the quality of the resulting data
(41). The types and distribution of geologic materials, the occurrence
and movement of ground water through those materials, the location of
the site in the regional ground-water flow system, the relative perme-
ability of the materials, as well as potential interactions between
contaminants and the geochemical and biological constituents of the
formation(s) of interest must all be'considered. •
Hydrogeologic Setting
There are three basic types of geologic materials normally
encountered in ground-water monitoring programs. These are: 1) porous
media; 2) fractured media; and 3) fractured porous 'media. In porous
media, the water and contaminants move through the pore spaces between
individual grains of the media. These media include sand and gravels,
silt, loess, clay, till, and sandstone. In fractured media, the water
and contaminants move through cracks or solution crevices in otherwise
relatively impermeable rock. These media include dolomites., some
shales, granites, and crystalline rocks. In fractured porous media, the
water and contaminants move through both the intergranular pore spaces
as well as cracks or crevices in the rock or soil. The occurrence and
movement of water through the pores and cracks or solution crevices
25
-------
depends on the relative porosity and degree of channeling from cracks or
crevices. These media include fractured tills, fractured sandstone, and
some fractured shales. Figure 2.1 illustrates the occurrence and move-
ment of water and contaminants in these three types of geologic
materials.
The distribution of these three basic types of geologic materials
is seldom homogeneous or uniform. In most settings, two or more types
of materials will be present. Even for one type of material at a given
site, large differences in hydrologic characteristics may be
encountered. The heterogeneity of the materials can play a significant
role in the rates of both tracer and contaminant transport, as well as
the optimum strategy for monitoring a site.
Once the geologic setting is understood, the site hydrology must be
evaluated. The location of the site within the regional ground-water
flow system also must be determined. Piezometric surface data or water
level information of each geologic 'formation at properly selected ver-
tical and horizontal locations.is needed'to determine the horizontal and
vertical ground-water, flow paths at the site of interest. Figures 2.2
and 2.3 illustrate two geohydrologic settings.commonly ericountered in
eastern regions of the United States where ground-water recharge', exceeds
evapotranspirational rates, Figure 2.4 illustrates, a .common geohydro-
logic setting for the arid western regions of the United Stated.
In addition to determining the directions of ground-water flow, it
is essential to determine the approximate rates of ground-water movement
to properly design a monitoring program. Hydraulic conductivity and
gradient data are required to estimate the Darcian or bulk flow rates of
ground water. Hydraulic conductivity data should be determined using
26
-------
(a)
Figure 2.1. Occurrence and movement of ground water through
a) porous media, b) fractured pr creviced media,
c) fractured porous media
27
-------
LOCAL AND REGIONAL GROUND WATER
FLOW SYSTEMS IN HUMID ENVIRONMENTS
Figure 2.2
28
-------
TEMPORARY REVERSAL OF GROUND-WATER FLOW DUE TO
FLOODING OF A RIVER OR STREAM
Temporary
reversal of
groundwater flow
Figure 2.3
29
-------
TOTAL POROSITY AND DRAINABLE POROSITY FOR TYPICAL
GEOLOGICMATERIALS (After Todd, 1980)
50
45
40
35
30
25
20
15
10
5
0
T T
T 1 T
I I
T T
Porosity
V
U
Specific yield
(drainable porosity)
u
>.
•o
ro
«/>
a
c
•a
c
ro
VI
01
c
iZ
•a
c
<0
a>
•a
a
§
o
•o
a
a
S
CO
01
c
g
n>
a
I
i,
S
i
O
CO
1/16 1/18 1/4 1/21 2 4 8 16 32 64 128 256
Maximum 10% grain size, millimeters
(The grain size in which, the cumulative total, beginning with the coarsest material
reaches 10% of the total sample.)
Figure 2.5
32
-------
artefacts into the results. Physical and hydrologic conditions will
determine whether or not evidence for chemical or biological inter-
actions can be collected. If the potential for these reactions or
transformations exist, consideration should be given to screening for
likely intermediates or transformation products.
The importance of understanding the hydrogeologic setting of the
site to be monitored cannot be overemphasized in developing an effective
sampling program. Similarly, the effects of the hydrogeologic setting
on the samples to ,be collected should be evaluated in detail and
considered in developing the sampling protocol.
Sampling Frequency
Traditional determinations of optimum frequencies for ground-water
sampling have been made by regulation or from statistical arguments in
analogy with surface water monitoring experiences (43,W. Sampling
frequencies determined by these methods emphasize data needs and the
economics of sample collection and analysis. A more reasoned approach
is to first evaluate the type of source that is being monitored, a
spill, slug, intermittent, or continuous source. Then one should con-
sider the likely pulse or continuous plumes of contaminants to be moni-
tored; determine the minimum desired sampling frequency in terms of
length along the ground-water flow path and use hydrologic data to cal-
culate the required frequency to satisfy these goals.
The type of potential pollution source has a direct influence on
the resulting plume that may be created. In the case of a spill or slug
source of pollution, discrete plumes may result. The size, shape, and
rate of plume movements will be dependent on: source characteristics,
33
-------
the hydrologic and geologic nature of the site in question and the
chemical reactivity and biological interaction of individual contami-
nants with the subsurface environment. Figure 2.6a illustrates this
type of phenomena. Intermittent releases of a pollutant may result in a
series of discrete plumes that may or may not overlap depending on the
relative frequency of the releases and the factors mentioned above.
Figure 2.6b illustrates this type of phenomena.
Continuous sources of pollution result in the development of plumes
that may approach steady state conditions for nonreactive conservative
chemical species. The size and shape of this type of plume can be esti-^
mated using a relationship described by Todd (42). Todd analyzed the
effects of regional ground-water flow on the circular cone of depression
in the water surface developed by pumping a well. For the purposes of
evaluating the effects of a pollution source on the regional flow
system, the pollution source can be treated as an injection well. The
expression describing the boundary of the affected downgradient region
(ignoring dispersivity) is as follows:
-(y/x) = tan (2KbI/Q)y (Eq. 2.1)
where K - hydraulic conductivity, in liters per day per square meter
b = aquifer thickness, in meters
I - hydraulic gradient, in meters per meter
Q - leakage rate from the source, in liters per minute
The rectangular coordinates (x and y) are as shown in Figure 2.7
with the origin at the center of the source.
Based on the expected type of plume, a decision can be made con-
cerning how often in the flow path samples are required for adequate
definition of plume dynamics. This decision can then be translated into
34
-------
TYPE OF PLUME GENERATED FROM (a) A SLUG SOURCE OR SPILL,
(b) AN INTERMITTENT SOURCE, AND (c) A CONTINUOUS SOURCE
A
(a)
(b)
(0
Figure 2.6
35
-------
RESULTING CHANGE OF A CAPTURE AREA DUE TO
REGIONAL FLOW (After ref. 42)
O Ground surface
tm
w<
Original piezometric surface-
, Slope = /
Drawdown curve
Impermeable
7— It
1
Confined aquifer
< ^
h
b
f
Impermeable
Figure 2.7
36
-------
a sampling frequency using the hydrogeologic parameters measured at the
site. The velocity of ground-water flow is described using Darcy's
equation and the effective porosity of the materials being monitored:
v = KI/7.48N (Eq. 2.2)
where v = velocity of ground-water flow, in meters per day
K = hydraulic conductivity, in liters per day per square meter
I = hydraulic gradient, in meters per meter
N = effective porosity, in percent
Figure 2.8 presents a nomograph for translating the hydraulic data
into sampling frequencies at.various flow path lengths.
INFORMATION NEEDS AND ANALYTE SELECTION
The information needs of a ground-water sampling program determine
both the scope and details of field and laboratory efforts. The needed
chemical information, in particular, will drive the selection of tech-
niques, procedures and' methodologies which will constitute integral
sampling and analytical protocols. All of the steps in these protocols
must be tailored to the analytes of interest by a well conceived plan
for field and laboratory operations. Detailed data on source composi-
tion and the type or extent of contamination available to most initial
investigations is usually limited. This is particularly true of ground-
water investigations at waste management facilities. Regardless of the
state of the information base, the planning effort must incorporate
flexibility to meet a variety of contingencies.
It is often more cost-effective and reasonable to plan the effort
for the maximum long-term return on the investment of fiscal and human
37
-------
SAMPLING FREQUENCY NOMOGRAPH
(K)
10°
'
ID'2-
-4-
10-6-
.
io-«-
•1
.E
o
V
_N\
s \ ,, _
« \ 01 s
i Nj|
i • l\
1 10-2. S \
i S \
8 - D \
a 2 \
3 10° J? \
< x \
g
X
^
w
<
(9
£
QUENCY OF SAN
Ul
oc
UL
• 10'
106
• 105
10"
. 103
10
10 .
1
.1
s
£
*•*,
(D)
100 —
80
60
40
20
10 _
8
6
4
2
1.0 —
.8
.6
.4
"^ .1 —
(N)
ta
£ ,os -i
t™
UJ
S
X .08-
2 .10-
O
"- .15-
(9
0 .20-
«t „
ui -26 -
O
g .30"
K ^x^*
5 ^"^ •"-
— '"^ .50"'
: POROSITY
S
Ul
UL
Ul
F = DN
— = 13.8 days
864 Ki
Example (clean sand)
K ." 10"1
i - ID"4
N • 0.30
D = 0.4 meters
Figure 2.8
38
-------
resources. Therefore, the planning effort should anticipate difficul-
ties and allow for refinement of the sampling and analytical protocols
as new data becomes available.
The basis of a successful monitoring program is a robust, integral
sampling protocol, coupled to proven'analytical schemes. Both field and
laboratory personnel should be involved in planning, once the minimum
information needs of the program are identified. In this way, the
potential impact of seemingly minor details of the program protocols can
be judged more appropriately. .
Parameter Selection ,
Parameter selection for chemical measurements is very important to
the effective planning of sampling and analytical protocols. For
exploratory efforts, it is useful to obtain slightly more chemical and
* • •.
hydrologic data than that required by the immediate information needs of
the program. . The added data can normally be put to good use as the site
conditions become better defined. For example, in a situation where
essentially no chemical .data for a site exists, a complete mineral
analysis should be included. The results provide an internal consis-
tency check on major ionic constituents, field determinations (e.g.
alkalinity) and the potential effects of unusually high levels of metals
or nutrient anions (16,23). Reliable analytical methods for ionic con-
stituents and routine field determinations (pH, Eh, temperature, con-
ductance and alkalinity) are well referenced for ground-water samples by
the USEPA (27,28,29) and various other groups (45,46). The results of
the complete mineral analysis and field determinations define the major
ion solution chemistry which is quite valuable to obtaining an overall
39
-------
picture of the subsurface system of interest. The major ion chemistry
determines the inorganic background and potential for matrix effects in
sampling and analysis. Chemical speciation of many specific inorganic
constituents of interest (e.g. Fe, Cu, Pb) may be controlled by the
inorganic solution chemistry. In turn, the speciation of the chemical
constituents of interest effects subsurface transport behavior and
sensitivity to either handling disturbances or recovery in analytical
separations.
With a complete mineral analyses and a' clear view of information
needs, one can then select the additional chemical parameters of
interest. These parameters may be characterized as general ground-water
quality parameters, pollutant indicator parameters and specific chemical
i
contaminants.
General Ground-Water Quality Parameters
Parameters which give a general overview of ground-water quality
relate to total dissolved solids content (e.g. Na+, Cl~, S0ij=) and tra-
ditional water treatment difficulties of ground water. Taste or odor
removal needs associated with the presence of dissolved iron, manganese
and total phenols vary substantially among ground-water supplies. Beyond
The determination of ground-water quality parameters may also provide an
indication of severely contaminated conditions. The choice of sample
collection and handling methods should be given careful consideration.
Degassing (e.g. loss of C02) and oxygenation (e.g. loss of Fe, trace
metals) can markedly effect analytical results, even for water quality
constituents at the ppm (mg-L~1) level (11). The sensitivity of the
results for these water quality parameters to sampling procedures is a
40
-------
function of the major ion chemistry and chemical speciation. Therefore,
complete mineral analyses should be included in most sampling programs,
if only on a limited basis. •
Pollution Indicator Parameters ...
Contaminant monitoring program requirements for parameter selection
reflect the following objectives: to detect whether or not the operation
of a facility results in the contamination of ground water, to determine
whether concentrations of specific chemical constituents are within
prescribed limits, and to measure the effectiveness of corrective
actions. In general, contaminant monitoring program approaches are of
two types. ' •.'..:.-. (
The generic approach requires the determination of parameters indi-
cative of gross disruption of the inorganic or organic chemistry of sub-
surface conditions [e.g. pH, solution conductivity (Q~1), total organic
carbon (TOG) and total organic halogen (TOX)]. It is a low cost ana-
lytical alternative, generally applied in detective monitoring situa-
tions. The rationale is that these surrogate parameters will indicate
the impact of waste releases to ground-water systems and suggest the
identity of the major classes of: the chemical constituents involved.
The usefulness of pH and, Q*"1 have been .mentioned above in relation to
their importance to total dissolved solids content and major ion
chemistry of ground-water samples.,.
Prior to the detection of water quality changes and in the absence
of a complete mineral analysis, the usefulness of the indicator
parameter approach is limited. This is especially true for TOG and TOX
determinations which are nonspecific and are limited in sensitivity.
-------
Sample collection: and handling precautions must be optimized to
insure that the volatile and nonvolatile fractions of both TOC and TOX
are recovered quantitatively (47). Otherwise, the significance of these
generic parameters may, be misrepresented and systematic errors in
sampling or analysis will negate their utility as diagnostic tools. It
should be pointed out .that the use of TOC and TOX as pollution indicator,
parameters can "enhance the interpretative power of observed data on
specific contamination distributions at substantially lower cost. The
trade off, of course, is that transformations of specific volatile or
nonvolatile contaminants may go unobserved. The second contaminant
monitoring approach focusses on a more specific set of chemical con-
stituents. ••••..,.•
Specific Chemical Constituents
Several alternative' approaches to generic contaminant 'monitoring
program emphasize the sampling and determination of specific mobile or
persistent chemical constituents. The selection of parameters may be
limited to those identified by law (e.g. Interim Primary Drinking Water
Standards or Resource Conservation and Recovery Act—Appendix VIII
parameters, etc.) or may be based on the actual composition of a regu-
lated facility's waste streams. i
The use of a''specific ' list of chemical constituents should be
approached cautiously. The determination of a legally mandated suite of
parameters tends to fdciis primarily on specific classes of compounds in
wide usage as starting"materials for manufacturing or commercial product
formulations. 'This type of program has' definite advantages, particu-
larly in situations Where the spill' or release ofa product occurs (48).
42
-------
However, detailed investigations of organic compound distributions in
environments contaminated by organic mixtures disclose that by-products
or substituted congeners of "priority-pollutants" may be the major
mobile and persistent constituents, while those parameters mandated by
compliance programs may be present only as minor trace components (49).
in situations where the original waste components or contaminant
mixtures are known, it is preferable to consider the relative mobility
and persistence of the known components, as well as potential transfor-
mation products. This mode of parameter selection demands a reasonable
understanding of the situation under investigation. Most of the stan-
dardized procedures for sample collection, handling and analysis which
function well in the initial phases of an investigation may have to be
modified to insure control of errors when applied to specific contami-
nants (-18,50,51). Once the likely suite of target chemical constituents
has ,been developed, the sampling and analytical protocols should be
thoroughly reviewed and modified appropriately. .
It is important to keep in mind that sampling errors will be
carried over into the analytical operations which follow. Generic
sampling protocols recommended for use in, ground-water investigations
(52) should be proven to be compatible with the analytical procedures by
careful consideration of accuracy, precision, sensitivity and complete-
ness performance guidelines (26).
In order to maximize the cost-effectiveness and flexibility of the
initial planning of a ground-water sampling program, it is useful to
anticipate that the degree of analytical detail required will increase
as .the investigation proceeds. Therefore, it is wise to prepare the
sampling protocol for the most troublesome chemical parameters which may
A3
-------
be of interest and maintain close control over the sampling operations.
Volatile organic compounds (e.g., benzene and trichloroethylene) which
are soluble and frequent early indicators of more persistent contami-
nants are a good candidate group of chemical constituents on which the
sampling protocol should be based. The principal errors introduced by
the sample collection mechanism, materials' exposures and sample
handling are due to degassing or volatilization and sorption or leaching
effects. These errors are common to those involved in accurately deter-
mining major ion chemistry, TOC, TOX, trace inorganic and nonvolatile
organic constituents to varying degrees, depending on the speciation and
analytical sensitivity for the chemical contaminants of interest. In
general, sample collection errors are systematic and directly affect the
accuracy of all subsequent analytical results.
An inappropriate sampling mechanism (e.g. air lift mechanisms for
volatile or gas sensitive parameters) can yield consistently inaccurate
and useless results. The literature provides valuable guidance in the
choice of appropriate sampling mechanisms and materials once the param-
eters of interest are idehtified with an emphasis on the more challeng-
ing problems posed by organic compounds (52,53). It is clear that
sampling mechanisms which minimize gas exchange or materials' effects
and permit well head determinations of pH, Eh, fi~1 and temperature are
those of choice for most detailed sampling programs. Sampling protocols
are an active area of research, but one should keep in mind that it will
cost less over the long-term if the investigation is planned correctly
to meet information needs. High quality data merit the time and expense
44
-------
of detailed interpretation. Invalid or biased data, on the other hand,
are expensive to evaluate and ultimately damage the credibility of the
program.
Minimal Analytical Detail for Ground-Water Monitoring Programs
The minimum data set, sufficient to the information needs of the
monitoring program, is defined by both geochemical and hydrologic con-
siderations. Once the set of routine data elements necessary to define
the situation at hand have been established, sampling frequency and
completeness requirements will dictate the dimensions of the data set.
For optimum data recovery and facile data interpretation, it is impor-
tant to define the size of the data set and allow for expansion of the
elements of interest. Computer assisted sample tracking procedures
incorporated into the overall data management system (including analyti-
cal data handling) can facilitate data validation and trend analysis.
The following recommended data sets have been developed to coincide
with detective, assessment and remedial action evaluation program goals.
They provide a degree of analytical detail which can be checked for
internal consistency. This is important to assure that the highest
quality data are produced which are commensurate with the manpower and
fiscal investments that high quality data collection demands.
Detection Monitoring Data Set
The minimal data set for a monitoring program designed for future
detection of contamination should provide the base level of information
on hydrologic and chemical conditions at a site. The parameters iden-
tified below will permit mass and charge balance checks on the consis-
45
-------
tency of the data and will provide valuable information on ground-water
chemistry. In this manner, the ability to identify "missing" charged
constituents, which may be contamination related, can be established.
Chemical Parameters
pH, fl-1, TOG, TOX, Alkalinity, Total Dissolved Solids
Eh, Cl~, N03~, SOU", P0n=, Si02
Na+, K+, Ca++, Mg++, NHz, + , Fe, Mn
Hydrologic Parameters
Water Level, Hydraulic Conductivity
This level of detail provides the basis for solution chemistry
composition calculations which are important for predictions of contami-
nant speciation, mobility and persistence.
Assessment Monitoring Data Set
The minimal data set for a monitoring program designed to assess
the type and extent of contamination incorporates the level of detail
noted in detective monitoring situations and indentifies potential con-
taminants of concern. The actual suite of potential contaminants may be
stipulated by regulation in some instances.
Chemical Parameters
pH, fl-1, TOC, TOX, Alkalinity, Total Dissolved Solids
Eh, Cl~, N03~, S0n = , P0i|-, Si02, B
Na+, K+, Ca+\ Mg++, NHi,*, Fe, Mn
Fe(II), Zn, Cd, Cu, Pb, Cr, Ni
Ag, Hg, As, Sb, Se, Be
46
-------
Hydrologic Parameters
Water Level, Hydraulic Conductivity
The realm of potential organic contaminants in ground-water systems must
be delimited based on the nature of the likely contaminant source. The
priority pollutant analytical scheme or selected categories of RCRA
Appendix VIII parameters should be a good starting point when other data
are unavailable.
WELL PLACEMENT AND CONSTRUCTION
The placement and construction of monitoring wells .are among the
most difficult decisions involved in developing an effective monitoring
program. The preliminary locations and depths of monitoring wells
should be selected based on the best available pre-drilling data. Then
as the actual installation of these wells progresses, new geologic and
hydrologic data should be incorporated into the overall monitoring plan
to insure that the finished wells will perform the tasks for which they
are designed. In most instances, it is probably advisable to select a
minimum array of monitoring wells for the collection of geologic and
hydrologic data. Then additional wells can be designed and constructed
to more effectively meet the goals of the monitoring program.
The positioning of a monitoring point in a contaminant flow path
must be determined on the basis of hydrologic data.; Therefore, the
contaminant flow path must be clearly defined in three dimensions.
Special emphasis must be placed on the collection of accurate water
level data as well drilling and construction progress. For example, the
level at which sand heaves up into the borehole is often related to the
depth at which the vertical movement of ground water is upward as
47
-------
opposed to the normally assumed downward migration. Accurate measure-
ments of stabilized water levels from an established reference elevation
is essential to understanding the flow paths of ground water and dis-
solved constituents. .
The construction of monitoring wells should be accomplished in a
manner that minimizes the disturbance of the materials in which the well;
is constructed (3). If the monitoring program calls for determinations
of organic compounds, care should be taken to steam clean the drill ,rig
and all other equipment and well components prior to mobilization to the
site. Repeated cleaning of drilling equipment and well-construction
materials at the site also is necessary. The drill rig should be
checked for hydraulic fluid and oil leaks prior to the initiation of
drilling. These preliminary precautions are essential to insure that
artefacts of the drilling process are not detected later in the program
and considered to be the result of actual conditions at the monitored
facility.
The selection of the type of drilling equipment should depend on
the type of geology present-, the expected depths of the wells, and the
availability of equipment in the location of interest. However, the
availability and relative costs of different types of drilling equipment
should not be used as the primary selection criteria. The use of
specialized drilling techniques may have real advantages for even the
most preliminary site investigations (50).
Drilling and Well Completion Methods
The selection of drilling and well completion methods for monitor-
ing well construction has been approached traditionally from considera-
48
-------
tions of the type of geologic materials to be penetrated, the
anticipated depth of drilling, and the availability of construction
equipment and materials. Little attention has been given to the poten-
tial adverse chemical effects of the drilling and well, construction
procedures on the samples produced from the monitoring well. This guide
discusses several drilling methods in terms of the suitability of their
application for monitoring well construction. Detailed discussions of
drilling procedures and rigs are presented in other references (3,54).
The selection of an appropriate drilling method for constructing
monitoring wells should be based on minimizing both the disturbance of
the geologic materials penetrated and the introduction of air, fluids,
and muds. The use of organic drilling muds or additives should be
avoided. The introduction of any foreign material has the potential for•
interfering with the chemical quality of water obtained from the moni-
toring wells. Based on these factors and the physical limits of the
various drilling methods and rigs, the following evaluations have been
made of the more commonly used types.
A summary of recommended applications for various drilling tech-
niques is presented in Table 2.1.
Hollow-Stem Continuous-Flight Auger
The hollow-stem continuous-flight auger rig is among the most
desirable drill rigs for the construction of monitoring wells. The rigs
are generally mobile, fast, and inexpensive to operate in unconsolidated
materials. No drilling fluids are used and disturbance to the geologic
materials penetrated is minimal. However, augers cannot be used in
consolidated rock and most rigs are limited to drilling to approximately
49
-------
Table 2.1. Recommended
Various Types of
Geologic Environment
Glaciated or unconsolidated
materials less than 150 feet
deep
Glaciated or unconsolidated
materials greater than 150 feet
deep
Consolidated rock formations
less than 500 feet deep (minimal
or no creviced formations)
Consolidated rock formations
less than 500 feet deep (highly
creviced formations)
Consolidated rock formations
more than 500 feet deep (minimal
or no creviced formations)
Consolidated rock formations
than 500 feet deep (highly
creviced formations)
Drilling Techniques for
Geologic Settings
Recommended Drilling Technique
(1) Hollow-stem continuous-flight
auger
(2) Solid-stem continuous-flight
auger
(3) Cable tool
(1) Cable tool
(1) Cable tool
(2) Air rotary with casing hammer
(3) Reverse circulation rotary
(1) Cable tool
(2) Air rotary with casing hammer
(1) Air rotary with casing hammer
(1) Air rotary with casing hammer
50
-------
45.5 m (150 feet) (3). In formations where the bore hole will not stand
open, the well is constructed inside the hollow-stem augers prior to
their removal from the ground. This limits the diameter of the well
that can be constructed with this type of drill rig to about 10.16 cm(4
inches). 15.24 cm (6-inch inside diameter augers are available for this
purpose. The use of hollow-stem auger drilling in heaving sand environ-
ments also presents some difficulties for the drilling crew. However,
with care and the use of proper drilling procedures, this difficulty can
be overcome.
Solid-Stem Continuous-Flight Auger
The use of solid-stem continuous-flight auger drilling techniques
for monitoring well construction is limited to relatively fine grained
unconsolidated materials that will maintain an open bore hole. The
method is similar to the hollow-stem continuous augers except that the
augers must be removed from the ground to allow the insertion of the
well casing and screen. This method is also limited to a depth of about
45.5 m (150 ft) and does not lend itself to collection of soil or for-
mation samples. This type of drilling method is a poor second choice to
the more desirable hollow-stem auger methods.
Cable Tool
The cable tool type of rig is relatively slow but still offers many
advantages that make it the second choice for monitoring well construc-
tion in unconsolidated formations and the method of choice for rela-
tively shallow consolidated formations. The method allows for the
collection of excellent formation samples and detection of even rela-
tively fine grained permeable zones. The installation of a steel casing
51
-------
as drilling progresses also provides an excellent temporary host for the
construction of a mbnitoring well once the desired depth is reached.
As stated earlier, the method is slow and small amounts 'of water
must be added to the hole as drilling progresses until the water table
is encountered. However, 'the' quantity of water -added to the hole and
into the formation to be sampled is minimal. A drive pipe diamet'er of
10.16 cm (^ inches)' may be too small for the easy construction of a
5.08 cm (2-inch) diameter well. It is recommended that a minimum
15.24 cm (6-inch) diameter drive pipe be used to'facilitate the place-
ment of the well casing, screen, and gravel pack, and a minimum 152.4-cm
(5-foot) long bentonite seal prior to beginning the removal of the drive
pipe. The placement of a bentonite seal in the drive pipe prior to
pulling will assist in holding the gravel pack, well casing, and screen
in place. The seal will also' isolate the gravel pack and screen from
the cement seals above. The drive pipe'is pulled in small increments to
permit the bentonite seal to flow outward and fill the annular space
vacated by the drive pipe. The drive pipe also is pulled' in< small
increments as cement grout material is added to ensure that. a satis-
factory seal is obtained.- -
Air Rotary
Rotary drilling methods operate on the principle of circulating
either a fluid or air to remove the drill cuttings and maintain an open
hole as drilling progresses. The different types of rotary drilling are
named according to 'the type of fluid and the direction of fluid flow.
Air rotary drilling forces air down the drill pipe and back up the bore
52
-------
hole to remove the drill cuttings. The use of air rotary drilling tech-
niques is best suited for use in hard rock formations. In soft uncon-
solidated formations a casing is driven to keep the formations from
caving. Similarly, in highly creviced formations it is often difficult
to maintain air circulation. Air rotary drilling appears to have poten-
tial ,for constructing monitoring wells without adversely affecting the
quality of water from monitoring wells in hard rock formations with
minimum unconsolidated overburden. The successful construction of moni-
toring wells using this drilling techniques is dependent on the ability
to maintain an open bore hole after the air circulation ceases. If the
wells are, intended to monitor for organic constituents, the air from the
compressor on the rig must be filtered to insure that oil from the com-
pressor is not introduced into the.geologic system to be monitored. The
addition of foam to the circulating air is often employed to increase
the effectiveness of air drilling techniques. Most of the foam addi-
tives contain organic materials which may interfere .with both organic
and inorganic constituents in samples collected from the constructed
monitoring wells. The use of air rotary drilling techniques should not
be used in highly polluted or hazardous environments. Contaminated
solids and water are blown out of the hole which are difficult to con-
tain. Protection of the drill crew and observers is correspondingly
very difficult.
Air Rotary With Casing Hammer -..-••
Air rotary drilling with casing driving capability increases the
utility of this type of drilling method. The problems associated with
53
-------
drilling in soft unconsolidated and highly creviced formations are mini-
raized. The utility of constructing monitoring wells in the casing prior
to pulling it also makes this type of drilling technique more appealing.
However, the same concerns about the oil in the circulating air and the
addition of foam additives must be considered. Grouting and casing
pulling procedures similar to those described for cable tool drilling
methods should be employed.
Reverse Circulation Rotary
Reverse circulation rotary drilling has limited application for the
construction of monitoring wells. Large quantities of water are circu-
lated down the bore hole and pumped back to the surface thru the drill
stem. The hydrostatic pressure of the water in the bore hole is used to
maintain an open bore hole. If permeable formations are encountered,
large quantities of water will infiltrate into those formations
altering in situ water quality. Similarly, water bearing units with
differing hydrostatic heads will have the opportunity for free inter-
change of waters altering the quality of water in the unit of lower
hydrostatic head. Because of the large quantities of water normally
required for this type of drilling and the high potential for water to
enter 'the formations to be sampled, this type of drilling is not
recommended.
Mud Rotary
Mud rotary drilling operates in the same fashion as the air rotary
drilling technique except water and drilling mud are circulated down the
drill pipe and back up the bore hole to remove the drill cuttings. The
bore hole is held open by the hydrostatic pressure of the circulating
54
-------
mud and a mud cake that develops on the bore hole wall during the
drilling process. The viscosity of the drilling mud is controlled to
minimize the infiltration of the.drilling fluid into .porous formations
penetrated by the drilling equipment.
The construction of monitoring wells using, mud rotary drilling
techniques is very difficult. The well must be constructed in the bore
hole which is still filled with drilling mud. This makes it difficult
to determine where gravel pack materials terminate and the well seal
begins. After monitoring wells are constructed, they must be developed
to produce visually clear water which will facilitate field filtration.
Breaking down the mud cake and removal of all mud introduced by this
drilling technique is extremely difficult when small diameter monitoring
wells are being constructed. Experience has shown that drilling mUds
not effectively removed from the well bore opposite the screen and
gravel pack will interfere with the chemical and biological quality of
samples from those wells (55,56,57). Many clay or synthetic drilling
muds contain organic matter (e.g., polymers, pqlyacrylamide or starches)
which can also greatly effect the organic content of water obtained from
mud rotary drilled wells (23,W. For these reasons, the use of mud
rotary drilling methods is not recommended, particularly for investiga-
tion of organic contaminant situations.
Bucket Auger
Bucket auger drilling rigs are usually employed for the construc-
tion of shallow large diameter wells or caissons. Their use is limited
to fine grained formations that are capable of supporting an open bore
hole. The large diameter created by this type of drilling technique is
55
-------
not usually warranted. The use of hollow-stem continuous-flight auger
techniques can be more effectively employed in appropriate geologic
environments.
Jetting
Jetting of monitoring wells is not a common practice in most of the
United States. Little or no information is obtained on the materials
thru which the well is jetted. As with the reverse rotary drilling
technique, water used in the jetting process also enters the formation
to be monitored and alters the in situ water quality. This type of
drilling technique is not recommended for monitoring well construction.
Driving
Driving of well points and casing may be acceptable in certain
hydrogeologic environments. As with jetting, little or no information
is obtained on the materials through which the well is driven. This
type of well construction should be limited to relatively shallow (less
than 15.17 m (50 feet)) homogenous sand and gravel formations. Due to
the nature of this geologic environment, no well seals are normally
required.
Monitoring Well Design
The effective design of monitoring wells requires careful consid-
eration of the hydrogeology and subsurface geochemistry at a site. The
information obtained from preliminary borings or well drilling can be
most useful in making logical decisions on the drilling, construction
and development methods which are appropriate for the program's goals.
The design of a monitoring well should not be based on the most readily
available types of drilling equipment or that used by the favorite
56
-------
driller in the area where the project is located. Cost considerations
alone should be'secondary to the retrieval of valid data which will meet
the goals of the program. Apparent cost savings realized by expedient
well construction may have a serious impact on the quality of the hydro-
logic and chemical data produced from the monitoring effort. The well
design goal should be to construct wells that will produce depth and
location specific hydrologic and chemical data.. Precautions must be
taken to insure that well completion .and development procedures mini-
mize the, disturbance to the geologic environment and the water samples.
Wells constructed for the production of large quantities of water
normally are not satisfactory for use as monitoring wells in detective
or assessment type monitoring programs. These wells are constructed
1 • • .' ; ' "'••.. - , -1 ! - ' ' •
with long sections of well screen or open bore holes designed to produce
water from large vertical and horizontal segments of the aquifer mate-
rials tapped. The resulting chemical quality of water pumped from the
wells represents an integrated chemical quality from all sections of the
aquifer contributing water to the well. Without knowledge of the
vertical and horizontal contributions of water to the well, these
chemical data have little value aside from indicating the quality of
water produced by that well. A potentially large amount of dilution of
any relatively small plumes (relative to the size of the pumping cone)
intersected by the pumping cone of the well could effectively mask the
' • ' - •' ' •" r.... •• . . ." ,"-• - •' ." - - • '• -
presence of the plume. Similarly, the hydrologic data obtained from
these types of wells represents an integrated water level for the
57
-------
vertical segment of the aquifer open to the well. The hydraulic conduc-
tivity data represent integrated values for the segments of the aquifer
influenced during the course of the pumping test.
ij
Depth of the Well
The depth of a monitoring well should be determined based on the
geology and hydrology of the site and the goals of the monitoring pro-
gram. In most monitoring programs the goal is to monitor the potential
effects of near surface activity. Therefore, it is essential to docu-
ment and monitor the downward migration of potential pollutants that may
be leaking from the facility. As percolating water and their solutes
move into the saturated zone, local and region flow systems are
encountered that will impart a horizontal component to the migration of
the pollutants.
To properly define the movement of pollutants,, vertically and hori-
zontally, it is essential to collect depth discrete water level data.
The uppermost relatively permeable zone will provide part of the data
needed to determine the vertical direction of ground-water movement.
The shallowest monitoring wells in the monitoring system .should be
finished in these materials. Water levels from these wells, if finished
in the same geologic materials, will provide information on theihorizon-
tal directions of shallow ground-water flow. In unconfined aquifer
systems this will represent the "water table." In confined aquifer
systems it represents the piezometric surface of the shallowest per-
meable zone.
58
-------
Additional wells at the same locations but at greater depths are
needed to complete the data set needed to determine the vertical direc-
tion of ground-water movement. These wells should be finished in the
next deepest relatively permeable zone in a geologic setting where
interbedded permeable and nonpermeable zones are present. In geologic
settings where the materials are: relatively permeable and uniform with
depth, the screens of adjacent wells should be staggered at an interval
equal to one to three times the selected screen depending on the
vertical detail necessary to define contaminant distributions. This
vertically-nested well depth approach should be continued at each well
location; until water level data indicates that the potential for deeper
migration of surface derived pollutants is minimal.
The required number of vertically nested wells and their depths
also will be a function of the relative horizontal to vertical permea-
bilities of the formations beneath the site and the hydrologic setting
in which they are located. The optimum approach is to ensure that the
vertical locations of the well screens are at the most likely depth to
intersect pollutants from the facility being monitored. An example is
presented below to illustrate the application of this type of monitoring
well design approach.
Example 2.1. Selecting depths for vertically nested wells in an
alluvial river valley setting—
Site background:
The site to be monitored lies on the banks of a
major river. Regional information indicates that
the unconsolidated materials are sand and gravel
from the surface to the underlying bedrock, about
120 feet. Regional water levels vary from about 15
to 25 feet below land surface. The activity to be
monitored is a small metal plating facility that
59
-------
uses a lagoon for disposal of its wastes. The rela-
tive specific gravity of the wastes is similar to
that of the native ground water. No hydrocarbons
are associated with the wastes.
Preliminary well construction:
Locations for vertically nested wells were selected
at one upgradient and three downgradient locations.
Two wells, one about five feet below the seasonally
low water table elevation and one approximately ten
feet deeper, were constructed at each location. The
wells were all equipped with two-foot long screens.
The wells were developed, elevations of the casing
tops (water level measuring reference point) were
surveyed to the nearest 0.01 feet, and water levels
were measured to the nearest 0.01 feet.
Preliminary water level analysis:
The following list presents the construction,
elevation, and water level data obtained .from this
preliminary and a subsequent second drilling effort.
Well no.
BG-1
BG-2
DG1-1
DG1-2
DG2-1
DG2-2
DG3-1
DG3-2
Depth
below
land
surface
32.0
42.0
26.0
36.0
27.0
37.0
26.0
36.0
Land
surface
elevation
349.27
349.27
344.11
344.11
343.42
343.42
339.73
339.73
Elevation
midpoint
screen
318.27
308.27
319.11
309.11
316.42
306.42
313.73
303.73
Measuring
point
elevation
352.00
351.85
346.69
346.53
345.97
345.78
342.71
342.59
Water
level
below
MP
22.00
22.65
23.29
23-73
23.89
24.18
22.61
22.61
Water
level
elevation
330.00
329.20
323.40
322.80
322.08
321.60
320.10
319.98
Second drilling effort:
DG1-3 46.0 344.11
DG2-3 47.0 343.42
DG2-4 57.0 343.42
DG3-3 46.0 339.73
DG3-4
56.0
339.73
299.11
296.42
286.42
293.73
283.73
346.37
345.53
345.29
342.31
342.17
23.62
24.11
24.19
22.26
21.97
322.75
321.42
321.10
320.05
320.20
At each of the vertically nested well pairs, the
direction of ground-water movement is downward.
Plotting the total hydraulic head (water level ele-
vations) at the midpoint of the well screens and
constructing flow path lines from the proposed
lagoon facility suggests that deeper wells are
required at DG1, DG2, and DG3.
60
-------
Second drilling effort and analysis:
One additional well was constructed at DG1 and two
additional wells were constructed at DG2 and DG3.
Data from those wells are included in the above
table. Plotting the total hydraulic heads at the
midpoints of the well screens and constructing flow
paths suggests that these wells should be adequate
for monitoring potential leakage, from the lagoon.
Figure 2,9 illustrates the analyses of data and
plotting of vertical and horizontal flow paths.
From this preliminary data, the appropriate wells
for sampling and chemical analysis can be selected.
As a final word of caution, this planning and con-
struction effort was accomplished during a period of
low water levels; data from periods of high water
levels should be examined to determine if the same
well configuration is adequate. Similarly, these
analyses were conducted prior to the influences of
leakage from the lagoon. The same type of water
level analyses should be performed periodically to
insure that the monitoring program remains effective
in meeting the intended goals.
In addition to the general guidelines noted above, wells intended
for use in monitoring hydrocarbon pollutants that are less dense than
water and likely to float on the water table surface should be con-
structed so the well screen is always open to the water table. If the
water table is known to fluctuate several feet over the course of the
year, the screen will have to be long enough to accommodate those fluc-
tuations.
The design of monitoring wells for sampling sites contaminated with
immiscible hydrocarbons or hydrocarbon products more dense than water
also warrants special consideration. A well screened throughout the
entire thickness of the aquifer potentially affected appears to offer
the best potential for adequately addressing this type of monitoring
problem (58). Wells constructed in this manner and properly sampled to
61
-------
BG
Proposed
Lagoon
DG1
DG2
—; Li j
An ^ —--^^ —*— —.
30.00 /
/ S~
29.20 /
' / '
A /
/
Seasonal low
.water table
— jL ___
/
->£ 122.E
, / >,
/ ^^122.75
/
80 ^ ^^ / 121.60 /
DG3
F22.08
/
•v.
/
^ <^?
-------
minimize the vertical migration of non-aqueous phases within the well
should provide reasonable indications of the vertical distributions of
hydrocarbons in the aquifer system.
Diameter of Monitoring Wells
The diameter of a monitoring well casing should be held to the
minimum practical size which will meet both the strength requirements
for the anticipated well depth and the size of the sampling pump
required to deliver water samples to the surface. Studies by Gibb (11)
have documented that the water held in storage in the well casing under-
goes chemical change while in the well casing. When pumping begins for
sample collection, some of this chemically altered water will be brought
to the surface along with water from the formation being sampled. The
relative quantity is related to the hydraulic conductivity of the forma-
tion being sampled,, the rate at which the well is pumped and the size
(diameter.) of the well casing. The amount of water removed from the
' - - '.!"'
well casing is a function of the formation hydraulic properties and the
pumping rate. Therefore, as the diameter of the well increases, larger
portions of altered, unrepresentative, water samples are delivered to
the surface to create the same amount of drawdown. Based on the availa-
bility of ground-water sampling' pumps capable of lifting water from
depths as great as 15.17 to 227.50 m (500 to 750 feet), 5.08-cm (2-inch)
diameter wells should .be used in all situations except where depth
requirements call for added material strength. Table 2.2 presents
general depth recommendations for various sizes of PVC, stainless steel,
and Teflon(R) well casings.
63
-------
Table 2.2. Well Casing Material Specifications
Depth Recommendations
PVC
Stainless steel
Nominal
casing
diameter
Wall
thickness
Weight
Ibs/ft
Type of
thread
Maximum
recommended
hang length*
(ft)
Schedule 40
2"
0.154"
0.716
square
3,100
Schedule 80
2 »
0.218"
0.932
square
3,300
Schedule 40 Type 304
2"
0.065"
1.732
fine
11,500
,2"
0.065"
1.732
square
Not
available
A ^J. _U^S* A
Schedule 40
. 2"
• 0.080"
0.9
square
320
* Length refers to total of single material. Depth range of 'Teflon(R) can be
extended by casing only the saturated zone with this material using another
material above. The hang lengths were calculated on the baiss of the shear
strength of the threads and the weight of the suspended casing.
64
-------
Size of Screen
The screen in a monitoring well should be long enough to permit
entry of water from the vertical zone to be monitored. In most geologic
settings a 60.96-cm (2-foot) long screen is adequate. The length of the
screen should be held to a minimum so that water level data obtained
from the well will represent relatively depth discrete information. In
wells where the length of the screen is long (152.40 cm (5 feet) or
more) the resulting water level is an integrated value representing an
average water level of the materials opposite the screen.
Tne slot size of the screen also should be selected to retain the
formation materials yet permit free entry of water into the well. Since
most monitoring wells are not pumped at high flow rates the available
open area of the screen is not usually an issue as in production well
screen design. In very fine grained deposits a gravel pack material is
often placed between the screen and the formation to be monitored. The
grain size of the pack material should be three to five times the
average grain size of the formation materials. The screen slot size
should be selected to retain 90% of the gravel pack materials. When
Teflon(R) casing and screen are used in deep formations, it is recom-
mended that a slightly larger screen slot size be used since the
Teflon(R) will tend to compress and reduce the effective slot size. The
gravel pack materials should be thoroughly cleaned and composed princi-
pally of quartz sand. Materials containing fine grained clay or silt
sized particles should be avoided. The chemical nature of the pack
material should be as inert as possible. Silica sand or glass beads are
recommended.
65
-------
For wells where no gravel pack is used, the screen slot size should
be selected to retain 60 to 70% of the materials opposite the screen.
The finer portions of the aquifer materials are removed from the forma-
tion during well development and a natural gravel annul us is created
around the well screen.
Grouts and Seals
The selection of grouts and seals for monitoring wells is an essen-
tial consideration to obtain water samples that are representative of
in-situ conditions. First, the seal must be adequate to prohibit.the
entry of surface water down along the well casing. Similarly, a good
seal must be maintained along the entire length of the well casing to
ensure that water from overlying formations does not migrate downward,
Effective seals are obtained by using expanding materials that will not
shrink away from the well'casing after setting. Expanding neat cement
and bentonite clay or a mixture of neat cement and bentonite clay are
among the most effective materials for this purpose. >
The selected seal also must not interfere with the water chemistry
results. Bentonite clay has appreciable ion exchange capacity which may
interfere with the chemistry on collected samples when the grout seal is
in close proximity to the screen or well intake. Similarly, expanding
cement which does not harden properly may affect the pH of water from
monitoring wells when in close proximity to the well screen or intake.
To minimize these potential interferences, a 30.48-cm (1-foot)
layer of fine Ottawa or silica sand should be place above the selected
gravel pack. Then, if possible, 30.48 to 60.96 cm (1 to 2 feet) of
bentonite pellets should be placed in the hole to prohibit the downward
66
-------
migration of bentonite slurry or neat, cement.- The upper 152.40 to
304.80 cm (5 to 10. feet) of the well casing should .be sealed with
expanding neat cement to provide for security and an, adequate surface
well seal. The exact depth of the upper seal should be slightly deeper
than the probable deepest frost penetration. This will protect the well
from frost heaving. . .
Multiple-Completion Wells
The use of multiple-completed wells in a singl:e bore hole has
received much attention in the literature. However, the effectiveness
of the well .seals between intervening monitoring .points is often
suspect. Advocates of multiple completed wells in the same bore hole
suggest .that pump tests can be used to verify the integrity of indi-
vidual seals. These verification procedures can only be used in situa-
tions where the well completions are not in hydraulic, connection. The
care and time necessary to properly seal these types of • wells are, not
justified when compared to the straightforward procedures, for sealing
separate holes for vertically nested wells.
Well or Sampling Point Documentation
The details of the construction of each well or sampling point
should be documented by both a drilling log and a well construction
diagram. The drilling log should contain descriptions of the general
texture, color, size and hardness of the geologic materials encountered
during drilling. Figure 2.10 illustrates a typical log containing these
types of information in an easily understandable format.
Geophysical (earth resistivity or seismic)' data should be mapped
and correlated with data from the soil boringsV Neutron'or: beta logging
67
-------
RECORD OF SUBSURFACE EXPLORATION
PROJECT-
JOB NO-
— : : _ BORING H"5
Monitoring Wells SHEET j_QF j_
:
Z
a.
u.
a
•10
•15-
•20-
•25-
•30-
•35-
DRILL
DATE
DRILL
LOQG
PIEZO
SAMPLE
NUMBER
1
2
3
4
5
ING
DRI
EDE
EDE
MET
-j
a:P
Un
t-
SS
ss
ss
ss
ss
ss
s
^•••M
METI
-LED
IV
ADVANCED /
RECOVERED (In)
18/16
18/li)
18/10
18/12
18/10
8/12
0/20
^™™™™^»»
DESCRIPTION OF MATERIALS
(Color Modifier MATERIAL. Claislllcallon)
Soil Classification Sy.fsm Un i f j i;d
Surface Elcvillnn -
Dark Gray CLAY, with Si It
Brown Silty CLAY, Trace Fine San
Brown Fine SAND, Dry ' •
Brown Fine Silty SAND
Brown Fine, to Medium SAND
36' TOB
BLOWS
(per 6 In;
5-7-10
4-5-6-'
5-6-6
2-3-8
6-8-11
6-8-1 1
-11-12
DRY UNIT WEIGHT (pel) I
4on HOI iipw Auoers ^^^.
12-17-82 ~ GR°l
• _ —El*
~ — • • •'• Ho
EH^ See Sketch
Shear Strength, tsf
SVA QP/jD OU/iO
?.'(', 1 . 1'A 2 2V.
PL NMC LL
* • Y
0
t^
50 100
laj^aji Rock Quality Dttlgnilion
.
JNDWATE
countered, al
tin after con
after con
after con
R LEVELS
19.0
nplctlon
npletlon
npletion
Feet
Feet
Feel
Feel
Figure 2.10. Drilling log sheet
68
-------
results may also be included on the logs of the bore holes investigated.
Natural gamma ray, gamma-gamma density (Cesium-MST source);, and electro-
magnetic induction logs can be run inside plastic casings as small as
5.08 cm (2 inches) in diameter and in some cases may be adequate and
more cost effective than collection of core samples for describing geo-
logic conditions. Use of these techniques should be compared or truthed
with a minimal number of core samples for visual and laboratory exami-
nation. In all cases, the dates of all activities should be recorded to
permit the reconstruction of the development of- site understanding.
Data summaries in the form of geologic cross-sections are often
very useful in developing a visual presentation of the subsurface condi-
tions. However, caution must'be exercised in interpolating between data
points (soil borings). In very homogeneous geologic environments,
extrapolations of data for tens to hundreds of feet may be acceptable.
In more heterogeneous environments, extrapolation of data should not
exceed a few tens of feet. To assess the relative homogeneity of the
geologic environment, site specific data should be evaluated with
respect to regional geologic information. No site description should be
considered complete without an indication of the geologic variability of
site conditions.
Once the bore hole is completed and well construction is underway,
the data necessary for documenting well completion should be collected.
The data items shown in Table 2.3 should be used to prepare a well con-
struction diagram.
This information on well construction can be summarized on a one-
page diagram similar to that shown in Figure 2.11. Geologic and pre-
liminary water level data also should be included for, completeness. The
69
-------
Table 2.3. Data Needed for a Monitoring Well Construction Diagram
Date/time of construction
Drilling method
Well location (±0.5 ft)
Bore hole diameter
Well depth'(±0.05 ft)
Casing material ;
Screen material
Screen slot size/length
Gravel pack type/size (depths from to )
Grout/sealant used (depths from to )
Backfill material (depth from to
Surface seal detail (depth from
Well protector type
Ground surface elevation (±0.01 ft)
Well cap elevation (±0.01 ft)
to
70
-------
4" Well Protector
Concrete Cap
Soil Backfill from Cuttings
Granular Bentonite
Filter Gravel
2" PVC Riser Pipe with Cap
Sand Cave-in
2" PVC Well Screen with 0.006" Slots
Figure 2.11. Monitoring well construction diagram
71
-------
water level data should indicate the length of time the bore hole was
open prior to the water level measurement. This information should not
be considered to be representative of the final level water reflected by
the finished well. The effects of well trauma and gradual equilibration
of water levels in newly-constructed wells limit the value of initial
water level measurements (59).
WELL DEVELOPMENT, HYDRAULIC PERFORMANCE AND PURGING STRATEGY
Once a well is completed, the sampling point must be prepared for
water sampling and measures must be taken to evaluate its hydraulic
characteristics. • These steps provide a basis for the maintenance of
reliable sampling points over the duration of a ground-water monitoring
program.
Well Development
The proper development of monitoring wells is essential to the
ultimate collection of "representative" water samples. During the
drilling process, fines are forced through the sides of the bore hole
into the formation, forming a mud cake that reduces the hydraulic con-
ductivity of the materials in the immediate area of the well bore. To
allow water from the formation being monitored to freely enter into the
monitoring well, this mud cake must be broken down opposite the screened
portion of the well and the fines removed from the well. This process
also enhances the yield potential of the monitoring well, a critical
factor when constructing monitoring wells in low-yielding geologic
materials.
More importantly, monitoring wells must be developed to provide
water free of suspended solids for sampling. When sampling for metal
72
-------
ions and other dissolved inorganic constituents, water samples must be
filtered and preserved at the well site at the time of sample collec-
tion. Improperly developed monitoring wells will produce samples con-
taining suspended sediments that will both bias the chemical analysis of
the collected samples and cause frequent clogging of field filtering
mechanisms (60). The additional time and money spent for well develop-
ment will expedite sample filtration and result in samples that are more
representative of water chemistry in the formation being monitored.
The development procedure used for monitoring wells are similar to
those used for production wells. The first step in development involves
the movement of water at alternately high and low velocity into and out
of the wellscreened gravel pack to break down the mud pack on the well
bore and loosen fines in the materials being monitored. This step is
followed by pumping to remove these materials from the well and the,
immediate area outside the well screen. This procedure should be con-
tinued until the water pumped from the well is visually free of sus-
pended materials or sediments.
Techniques for High Hydraulic Conductivity Wells
Successful development methods for relatively productive wells
include the use of a surge block, bailing, and surging'by pumping. A
surge block is a plunger device that fits loosely inside the well
casing. It is moved .forcibly up and down, causing water to surge in and
out of the well screen. After surging; the well must be pumped to
remove-the fines carried into the well screen and casing. The use of
surge blocks for monitoring well development has not been widely used.
However, if the surge block is sized to fit loosely in the monitoring
73
-------
well (0.64 cm (1/4 inch) total clearance) it can be operated effectively
by hand in relatively shallow wells, less than 15.17 m (50 feet) deep.
Care must be taken to avoid damage to the casing or screen when surging
a monitoring well.
A bailer also may used to obtain the same surging effect created by
a surge block. The bailer must be sufficiently heavy to quickly fall
through the water forcing some water to flow out of the well into the
surrounding formations. The upward movement of the bailer will then
pull the lessened fines into and remove them from the well. The use of
bailers for development of monitoring wells is more common than the use
of surge blocks. Bailing is generally less effective than using surge
block through the potential for well damage is minimized. , •
Alternately pumping and allowing a well to equilibrate for short
intervals is another method for developing monitoring wells. .Pumping
procedures have had limited application in very high conductivity
wells. This is because it is difficult to draw-down these wells suffi-
ciently to create the high entrance velocities necessary for the removal
of fines in the aquifer and well bore. This type of development is more
often attempted by using air lift pumping mechanisms.
When pumping with air, the effectiveness of the procedure depends
on the geometry of the device injecting air into the well. Figure 2.12
illustrates a.simple device that diverts air through the well screen to
loosen the fines and forces air, water and fines up the well casing and
out of the well. This device is particularly effective for developing
monitoring wells in very productive geologic materials.
Several important factors should be considered when developing
monitoring wells with air. First, the air from the compressor must be
74
-------
SCHEMATIC DIAGRAM OF AN
AIR DRIVEN WELL DEVELOPMENT DEVICE
Flattened nozzle with
1/8" opening
3/8" OD stainless or
copper pipe
Overall dimension
less than 1-1/i"
1/8" diameter hole (both sides)
Figure 2.12
75
-------
filtered to ensure that oil from the air compressor is not introduced
into the well. High volume carbon filters can be used successfully to
filter the air from compressors. Secondly, in highly contaminated
ground-water situations air development procedures may cause the expo-
sure of field personnel to hazardous materials. Precautions must be
taken to minimize personnel exposure. Finally, air development may
perturb the oxidation-reduction potential of the formation of interest
with effects on the chemistry of initial water samples. Experience
shows that in permeable sand and gravel situations, the effects do not
persist for more than a few weeks.
Techniques for Low Hydraulic Conductivity Wells
Development procedures for monitoring wells in relatively unpro-
ductive geologic materials is somewhat limited. Due to the low
hydraulic conductivity of the materials, it is difficult to surge water
in and out of the well casing. Also, when the well is pumped, the
entrance velocity of water can be too low to remove fines effectively
from the well bore and the gravel pack material outside the well screen.
In this type of geologic setting, clean water should be circulated
down the well casing out through the screen and gravel pack, and up the
open bore hole prior to placement of the grout or seal in the annulus.
Relatively high water velocities can be maintained and the mud cake from
the bore hole wall will be broken down effectively and removed. Flow
rates should be controlled to prevent floating the gravel pack out of
the bore hole. Because of the relatively low hydraulic conductivity of
geologic materials outside the well, a negligible amount of water will
penetrate the formation being monitored. However, immediately following
76
-------
this procedure, the well sealant should be installed and the well pumped
to remove as much of the water used in the development process as
possible.
Hydraulic Performance of Monitoring Wells
The importance of understanding the hydraulics of the geologic
materials at a site cannot be over emphasized. Collection of accurate
water level data from properly located and constructed wells provides
information on the directions, horizontal and vertical, of ground water
flow (61). The success of a monitoring program also depends on knowl-
edge of the rates of travel of both the ground water and solutes. The
response of a monitoring well to pumping also must be known to determine
the proper rate and length of time of pumping prior to collecting a
water sample. Finally, the required sampling frequency should be deter-
mined based on the rate of ground-water travel, the mobility and persis-
tence of the chemical constituents of interest, and the goals of the
monitoring program.
It is recommended that "field" hydraulic conductivity test be con-
ducted to avoid the unresolved issues involved in laboratory testing.
Conductivity tests should be performed on every well in the monitoring
system to provide maximum understanding of the hydraulics of the site
being monitored, provide information for recommended sampling proce-
dures, and to determine appropriate sampling frequencies for the wells.
Traditionally, hydraulic conductivity testing has been conducted by
collecting drill samples which were then taken to the laboratory for
testing. Several techniques using laboratory permeameters are routinely
used. Falling head or constant head permeameter tests on recompacted
77
-------
samples in fixed wall or triaxial test cells are among the most common.
The relative applicability of these techniques is dependent on both
operator skill and methodology since calibration standards are not
available. The major problem with laboratory test procedures is that
one collects data on recompacted geologic samples rather than geologic
materials under field conditions. Only limited work has been done to
date on performing laboratory tests on "undisturbed" samples to improve
the field applicability of laboratory hydraulic conductivity results.
Water Level Measuring Techniques
There are three common water level measurement techniques or
devices used for measuring water levels in monitoring wells, steel
tapes, electric drop line, and pressure transducers. General descrip-
tions of their use and their relative accuracy are discussed in the
following sections.
Steel Tapes
The use of relatively narrow (o.64- to 0.95-cm (1A- to 3/8-inch)
wide) steel tapes is among the most accurate and straight forward
methods for making water level measurements. Tapes that are graduated
throughout their entire length in feet, tenths of a foot, and hundredths
of a foot with raised lettering and divisions are preferable. The
raised surface of the tape will permit the observation of color changes
when the chalk or other material is wetted. The bottom few feet of the
tape is chalked and lowered into the well to the anticipated water level
depth so that the chalked portion of the tape is in the water. The tape
is held at an even foot mark, making sure that the tape has been con-
tinuously lowered into the water and not raised back to the foot marker.
78
-------
The tape is then withdrawn and the reading from the wetted portion of
the tape subtracted from the foot reading held at the measuring point.
The resulting value is the depth to water from the measuring point.
Measurements taken in this manner are generally accurate to the
nearest 1/100 of a foot. Three readings should 'be taken for each
measurement to ensure that reproducible results are obtained.
Electric Drop Lines
Commercially purchased and home-made drop lines are often used for
measuring water levels in monitoring wells. Two conductor electrical
wire is 'fitted with a probe to hold the two wires apart and marked at
30.48- to 152.40-cm (1- to 5-foot) intervals through out its entire
length. Drop lines are generally powered by flashlight batteries and
equipped with a milliammeter. The drop line is lowered into the well
until the probe contacts the water closing the circuit between the two
wires and the meter indicates a current flow. The drop line is pulled
back and a ruler used to measure the distance between the nearest 30.48-
or 152.40-cm. (-1- or 5-foot) markers on the drop line.
After repeated use the markings on drop lines often have a tendency
to become loose or to slide along the wires. Drop lines may also become
kinked and don't hang straight in the well. These among other potential
problems can limit the accuracy of drop lines to about 1/10 of a foot.
They are, however, very convenient to use particularly in deep wells and
don't need to be be totally reeled out of the well to get multiple
readings.
79
-------
Pressure Transducers
Pressure transducers have not been used in monitoring wells until
the last four to five years. Their use does, however, offer advantages
over the steel tape and electric drop line. Transducers can be,lowered
into a monitoring well to a known distance below the measuring point and
by indicating the amount of pressure exerted on it, measures the height
of water above the transducer. This amount of "submergence" is
subtracted from the depth below the measuring point at which the
transducer is located to obtain the depth to the water. Transducers are
particularly useful for making water level measurements in a well during
pump or slug tests. The transducer is left in the well and transmits a
continuous record of water level data to a strip chart or digital
recording device during the course of the test. Permanent installations
of transducers into individual wells normally cannot be justified
because of their relatively high costs.
The accuracy of transducers depends on the type and sensitivity of
device used. Most transducers are rated in terms of a percent of their
full scale capability. For example, a 0 to 5 psi transducer rated at
0.01 percent will provide readings to the nearest 0.30 cm (0.01 foot).
A 0 to 25 psi transducer rated at 0.01 percent will provide reading to
the nearest 1.52 cm (0.05 foot).
Hydraulic Conductivity Testing Methods
Slug tests
Slug or bail tests are described in detail in Freeze and Cherry
(62). Two tests, one suitable for a point piezometer and one suitable
for a well screened over the entire saturated thickness of an aquifer
80
-------
are 'presented. Both tests are initiated by introducing a sudden change
in water level and measuring the resulting response of the well or
piezometer. The change in water level can be accomplished by intro-
ducing a known quantity of water, slugging the well, or by removing a
known quantity of water with a bailer. These methods are suitable for
relatively low conductivity settings where the resulting changes in
water levels take place slowly and accurate measurements 'can be made.
However, for wells where hazardous contaminants are suspected, removing
water from the well may not be desirable. In the case of the slug test,
water of a different quality than that in the aquifer also is introduced
into the system and must be removed prior to sampling the well.
Prosser (63) described a method of depressing the water level by
pressurizing the well casing and then rapidly releasing the pressure to
allow the water level to recover. This technique minimizes the distur-
bance of the well and has the least potential for compromising the
integrity of water quality samples. This method also can be used for
conducting tests on wells with very high hydraulic conductivities when
pressure transducers are used for the water level measurements.
The analyses of slug or bail test data has been described by
Hvorslev (64) and Cooper et al. (65). The Hvorslev method is for a
point piezometer, while that of Cooper is for a confined aquifer. In
most instances the method described by Hvorslev can be used. Hvorslev
analysis assumes a homogeneous, isotropic, infinite medium in which both
the fluid and soil are incompressible. The rate of inflow to the
piezometer (q) is defined by equation 2.3:
81
-------
q(t) = IIR2 (dh/dt) = FK(H - h) (Eq. 2.3)
where R = radius of the well
F = shape factor determined by the dimensions of the piezometer
K = hydraulic conductivity
H 3 initial water level above a reference point (±0.01 ft)
h = water level above the reference point at time t (±0.01 ft)
See Figure 2.13a. flvorslev defines the basic time lag, To, as
equation 2.4:
T0 = (JIR2/FK) (Eq. 2.4)
By substituting this parameter into equation 1 , the solution to the
resulting ordinary differential equation, with the initial condition,
h = H0 at t = 0 is:
H - h
H - Hr
-t/T,
A plot of field recovery data, H - h / H - Ho versus t on a logarithmic
scale results in a straight line. Note that for H - h / H - H0 = 0.37,
ln(H - h / H - H0) = -1, and from equation 2.4, To = t. This describes
the basic time lag.
To interpret field data, the data are plotted as shown on Figure
2.13b. The basic time lag is graphically measured and K is determined
using equation 2.3. For a piezometer intake of length L and radius R,
with L/R > 8, Hvorslev has evaluated the shape factor, F, and the
resulting expression for K is equation 2.5:
K = r2ln(L/R) (Eq. 2.5)
2LTn
82
_
-------
HVORSLEV PIEZOMETER TEST, (a) GEOMETRY; (b) METHOD OF ANALYSIS
(from ref. 64)
J_
T
dh
H h HoV-
-t = co (and t<0)
1\
•t+dt £
•t §
u
CO
(E
•t=0 —L-
-Datum
(a)
0.5 -
0.37
0 2 4 6 8 10
(b)
Figure 2.13
83
-------
Pumping Tests
Pump tests on monitoring wells are often difficult to perform.
Relatively low pumping rates, 100 to 1000 milliliters per minute com-
monly are required to produce data suitable for analysis. Problems of
disposing of the water pumped and making accurate water level readings
also must be addressed. Constant rate pump tests for periods of two to
four hours are normally required. Traditional analyses of pump test
data use equations derived by Theis (66) and Jacob (67). One of the
basic assumptions made in deriving those equations is that all of water
pumped from a well during the pumping test comes from the aquifer and
none comes from storage within the well. This condition is seldom
encountered in monitoring wells. Therefore, the methods presented by
Papadopulos and Cooper (68), which take into account the water removed
from storage in the well casing, should be used. This method as applied
to monitoring wells is described by Gibb (11). it should be noted that
the well construction procedures, particularly "smearing" of the well
bore or infiltration of drilling muds, can significantly impact hydrau-
lic conductivity calculations (69). Therefore, well development is
essential prior to hydraulic conductivity testing.
Analysis of Water Level Data
In settings where slug tests or pump tests can not be performed,
historical water level data can be analyzed using the procedures
outlined by Stallman (70). Reasonable estimates of hydraulic conduc-
tivity can be made by selecting appropriate well arrays and periods of
time when little or no recharge has occurred. Successive selections of
84
-------
various well arrays will permit the determination of hydraulic conduc-
tivity values for most wells in a monitoring program.
In all of the above described procedures there are significant
sources of error. Water levels should be measured to the nearest ±0.30
centimeters (0.01 foot),flow rates for pump tests to ±5 mL per minute,
and time to the nearest 2 seconds. Hydraulic conductivity values deter-
mined by the various methods should not be considered to be more precise
than ±20 percent. To minimize the potential error and quantify the
degree of variance, 3 to 5 slug or pressure tests should be conducted on
each well. The time and expense required to perform multiple pump tests
do not normally warrant the effort.
In addition to the above sources of measurement and interpretive
errors, wells that have not been properly constructed or developed will
not provide accurate data for determining hydraulic conductivity values
of the materials in which the well is finished. Care also must be exer-
cised when performing these tests to insure that pumping or injection of
water by nearby wells does not affect the results of these tests.
Well Maintenance Procedures
A plan for well maintenance and performance reevaluation should be
prepared to insure that the sampling point remains reliable. As a
minimum, high and low water level data periods for the site should be
examined once every two years to insure that the well locations (hori-
zontally and vertically) are still acceptable. It is also particularly
important to note that the exposure of the screened interval, to the
atmosphere due to low water levels can compromise the integrity of water
samples. Hydraulic conductivity tests should be performed once every
85
-------
five years or whenever significant amounts (7.62-15.24 cm (0.25-0.50
feet)) of sediment have accumulated in the well. Deficiencies in well
locations, delivery of decreases in hydraulic conductivity, or turbid
samples should be corrected by well redevelopment, the installation of
new wells or the rehabilitation of existing wells.
The operation of wells in the vicinity of the site under investi-
gation also may affect' changes in the hydrologic setting and resulting
flow paths. Biannual evaluation of the high and low water level condi-
tions at a site under evaluation is recommended to ensure that the well
locations and depths are still appropriate. Piezometric surface maps
for horizontal flow direction determination and vertical cross sections
of equipotential lines for vertical flow determination should be plotted
and reviewed. The example below illustrates how site operation often
causes failure of the original monitoring well design.
Example 2.2. Effects of waste disposal activities on site flow regime--
Figure 2.14 showed a relatively flat site with a slight water table
gradient prior to the placement of a waste impoundment. Background and
down gradient wells were constructed to determine the hydrologic and
chemical nature of the site prior to waste disposal. Water level data
from the nested monitoring wells were used to determine the horizontal
and vertical components of ground-water flow. In the pre-disposal
situation water passing the upgradient shallow well BG-S was expected to
flow past the deep downgradient well DG-D2.
After the installation of the disposal system a ground-water mound
was created beneath the impoundment. The increased head beneath the
impoundment also resulted in the reversal of ground-water flow in its
immediate vicinity. Background wells BG-D and BG-S are both now likely
to receive leachate from the source. Similarly, the increased head
beneath the impoundment increases the vertical component of flow and
causes the downgradient flow of ground water to move deeper into the
regional flow system. This shift in flow patterns indicates the need to
construct a deeper well at the site DG-D2 and DG-S2.
This type of analysis should be performed for high and low water
level periods once every two or three years to insure that the designed
monitoring system is still applicable to possible changes in the hydro-
logic system.
-------
-------
In addition to determining that the monitoring wells are still
properly located, documentation must be presented to ensure that the
wells are physically intact and capable of yielding water samples as
designed. Chemical encrustation or bacterial growths on the well screen
may result in decreased well performance and possible alteration of the
chemical quality of pumped samples. Well depth measurements . should be
reported on an annual basis to document that the well is still
physically intact and not filling with sediment. Turbid water samples
are an indication that the well intake or screen is not functioning as
designed and is likely to accumulate sediments.
Another recommended procedure for documenting the integrity of
monitoring wells is to require that slug or pump tests be conducted on
each well once every five years. Comparisons of these test data with
thatcollected originally provides documentation on the presence and
degree of well deterioration. This data can then be used to determine
if and when new wells or well rehabilitation is needed. The example
below illustrates common problems often encountered as the age of moni-
toring wells increase.
Example 2.3. Well deterioration and plugging—
ino*. ti *«ir 2~inch diameter PVC monitoring well was
iS i? in September 1979 within 250 feet of a petrochemical plan?
waste disposal impoupndment site. Upon completion, the well was tested
and found to have a hydraulic conductivity of 5.2 x 1
-------
The problems associated with this well are two-fold. During the
first two years, sediment had accumulated due to improper development
procedures. The cessation of sediment accumulation could have been due
to the ultimate development of the well from repeated pumping during the
sampling of the well. It also could have been due to the slow plugging
of the well screen or aquifer reflected by the drop in hydraulic conduc-
tivity Due to the nature of the possible leachate from the disposal
site being monitored, the life of the well could be threatened by attack
of the well casing materials. Careful monitoring of the well perfor-
mance and chemistry is recommended. The well may need to be replaced
with a new well constructed of more suitable materials for this type of
environment.
Well Purging Strategies
The number of well volumes to be pumped from a monitoring well
prior to the collection of a water sample must be tailored to the
hydraulic properties of the geologic materials being monitored, the well
construction parameters, the desired pumping rate, and the sampling
methodology to be employed. There is no one single number of well
volumes to be pumped that is best or fits all situations. The goal in
establishing a well purging strategy is to obtain water from the geo-
logic materials being monitored while minimizing the disturbance of the
regional flow system and the collected sample. To accomplish this goal
a basic understanding of well hydraulics and the effects of pumping on
the quality of water samples is essential. Water that has remained in
the well casing for extended periods of time (i.e. more than about two
hours) has the opportunity to exchange gases with the atmosphere and to
interact with the well casing material. The chemistry of water stored
in the well casing is unrepresentative of that in the aquifer and it
should not be collected for analysis. Purge volumes and pumping rates
should be evaluated on a case by case basis.
89
-------
Pumping Rates
The rate at which wells are p.urged of stagnant water should be kept
to a minimum. Purging rates should be maintained below the rates at
which well development was performed since well damage can result.
High purging rates can also cause additional development to occur with
resulting increased turbidity of water samples. Well hydraulic perform-
ance evaluation is essential to the determination of effective well
purching rates and volume requirements.
Evaluation of Purging Requirements
When a well is pumped a certain amount of drawdown is created in
the well and the surrounding aquifer system to induce flow of water to
the well. Traditional well analysis techniques described by Theis (66)
and Jacob (67) can be used to predict the amount of drawdown experienced
by wells under water table and piezometer conditions. The reader should
recall from above that the basic assumptions made in deriving these
relationships require that an insignificant amount of the water pumped
comes from the well bore. This condition is seldom experienced in the
case of small diameter monitoring wells, particularly for wells finished
in low hydraulic conductivity geologic settings. Popadopulos and Cooper
(68) presented an equation that describes the discharge from a pumped
well which takes into account the volume of water removed from casing
storage.
Well test data for six monitoring wells studied, in Illinois have
been analyzed using these equations (11). At all of the sites studied,,
the nonpumping water levels were significantly above the top of the
aquifers tapped, suggesting artesian conditions. In this case, a
90
_
-------
storage coefficient of 0.0001 was used in the analysis of the drawdown
data. The storage coefficient values selected should have little effect
on the predicted drawdowns for most aquifer systems using this equation.
Using the Popadopulos and Cooper equations (68), the percentages of
aquifer water pumped for a two inch diameter well pumping at a rate of
500 mL/min for a range of transmissivities were calculated, see Figure
2.15. These calculations give an indication of the sources of water at
various times for a well that is being pumped with the pump intake at
the top of the well screen. Different percentages would result if the
pumping rate, well diameter, or aquifer properties were different. These
types of calculations should be used as guidelines for the selection of
the appropriate pumping rate and numbers of well volumes to be pumped
prior to sample collection. However, they are only guidelines and
should be verified by the measurement of indicator parameters at the
well head during pumping collection. Two examples of pumping rate
selection and appropriate well purging volumes are given below.
Example 2.H. Well purging strategy based on hydraulic conductivity
data—
Given:
J48-foot deep, 2-inch diameter well
2-foot long screen
3-foot thick aquifer
static water level about 15_feet below land surface
hydraulic conductivity = 10~2 cm/sec
Assumptions:
A desired purge rate of 500 mL/min and sampling rate of 100
mL/min will be used.
Calculations:
One well volume = (48 ft - 15 ft) x 613 mL/ft (2-inch diame'
well)
=20.2 liters
91
-------
= 500mL/min
DIAMETER = 5.08 cm
10 15 20
TIME, minutes
Figure 2.15. Percentage of aquifer water versus
time for different transmissitivites
92
-------
Aquifer Transmissivity
= hydraulic conductivity x aquifer
thickness
= 10"^ m/sec x 1 meter
= 10"1* m2/sec or 8.64 m2/day
From Figure 2.15
at 5 minutes -95$ aquifer water and
(5 min x 0.5 L/min)/20.2 L
0.12 well volumes
at 10 minutes ~t005& aquifer water and
(10 min x 0.5 L/min)/20.2 L
=0.24 well volumes
It therefore appears that a high percentage of aquifer water can be
obtained within a relatively short time of pumping. The indicator
parameters should be monitored and the pumping rate slowed to the
desired 100 mL/min for sampling as soon as they have stabilized. The
indicator parameters should be monitored at very close intervals, every
1 or 2 minutes from the time pumping begins.
Example 2.5. Well purging strategy based on hydraulic conductivity
data—
Given:
48-foot deep, 2-inch diameter well
2-foot long screen
3-foot thick aquifer
static water level about 15_feet below land surface
hydraulic conductivity = 10~4 cm/sec
Assumptions:
A desired purge rate of 500 mL/min and sampling rate of 100
mL/min will be used.
Calculations:
One well volume = (48 ft - 15 ft) x 613 mL/ft (2-inch diameter
well)
= 20.2 liters
Aquifer Transmissivity = hydraulic conductivity x aquifer
thickness
= 10~6 m/sec x 1 meter
= 10~6 m2/sec or 0.0864 m2/day
93
-------
From Figure 2.15
at 5 minutes -20% aquifer water and
(5 min x 0.5 L/min)/20.2 L
-0.12 well volumes
at 10 minutes ~30% aquifer water and
(30 min x 0.5 L/min)/20.2 L
-0.72 well volumes
Based on these results, it appears that it may be more desirable to
pump this well down to the top of the screen and allow it to recover
Dewatering the screen and the gravel pack should be avoided to minimize
aeration effects on water chemistry. The samples can then be collected
at the desired 100 mL/min while the well is recovering. Calculations
using the equations developed by Papadopulos and Cooper suggest that the
well should recover at a rate of about 250 mL/min when the water level
is near the top of the screen. Therefore, the samples can be collected
within five minutes after dewatering pumping stops and can continue
until the desired volume of sample is collected.
If the well was not capable of recovering at a rate in excess of
100 mL/min, the sample would have to be collected in small aliquots
The amount of water that could recover in two hours should be collected
and another recovery period would be required to collect the next sample
segment. The recovered water should not be allowed to remain in the
well casing for more than about two hours prior to collection or it is
likely to be chemically altered for several parameters.
The selection of purging rates and volumes of water to be pumped
prior to sample collection can also be influenced by the anticipated
water quality. In hazardous environments where purged water must be
contained and disposed of in a permitted facility, it is desirable to
minimize the amount of purged water. This can be accomplished by pump-
ing the wells at very low pumping rates (100 mL/min) to minimize the
drawdown in the well and maximize the percent aquifer water delivered to
the surface in the shortest period of time. Pumping at low rates, in
affect, isolates the column of stagnant water in the well bore and
negates the need for its removal. This approach is only valid in cases
where the pump intake is placed at the top of, or.in, the well screen.
In summary, well purging strategies should be established by
1) determining the hydraulic performance of the well; 2) calculating
94
-------
t-easonable purging requirements, pumping rates, and volumes based on
hydraulic conductivity data, well construction data, site hydrologic
conditions, and anticipated water quality; 3) measuring the well purging
parameters to verify chemical "equilibrated" conditions; and 4) docu-
menting the entire effort (actual pumping rate, volumes pumped, and
purging parameter measurements before and after sample collection).
SAMPLING MECHANISMS AND MATERIALS
The selection of appropriate sampling mechanisms and materials are
vital to the success of any ground-water investigation. A situation may
be very thoroughly evaluated as to the hydrogeologic conditions, opti-
mized sampling frequency and analyte selection. Also, the sampling
points may be constructed and evaluated properly. Nonetheless, if poor
sampling mechanisms and materials are incorporated into the program, all
the preceding effort may be futile. Minimally disturbed samples must be
carefully collected and analyzed if the program is to meet its informa-
tion needs. In many cases, the results of preliminary investigations
can be reinterpreted, even if inappropriate choices of sampling mecha-
nisms or materials have been used prior to the execution of a sampling
experiment. These experiments should include simultaneous sample col-
lection by both the previous mechanism and sampling components, as well
as those which are more appropriate for the current situation based on
the available data. For example, an initial set of monitoring results
collected with a conventional bailer may show a trace of volatile
organic contaminants. In order to substantiate these observations and
improve the reliability of the results, a sampling experiment should be
run, including both bailed and bladder pumped samples on at least two
95
-------
successive sampling dates. This approach will permit the objective
evaluation of the effect of sampling procedures on the quality of the
results and hopefully will put an end to the generation of poor data.
Tradition is a very weak basis for the selection or continued use of
inappropriate mechanisms or materials.
Sampling Mechanisms
Sampling mechanisms for the collection of ground-water samples are
among the most error prone elements of monitoring programs. Several
useful sources have reviewed the range of available sampler designs and
should be consulted for specific information (3,11,13,52,71,72). Unfor-
tunately, the documentation of field sampling performance for many of
the available devices is lacking. Many of the sampling designs may be
expected to provide adequate performance for conservative chemical con-
stituents which are not (or minimally) affected by aeration, gas-
exchange and degassing. Among these constituents are Na"1", K+ and Cl~.
The chemical constituents which can provide the most useful information
to the investigation frequently are effected by the improper choice of
sampling mechanisms. Evaluations of sampling performance based on the
recovery of conservative, unreactive chemical constituents are simply
not reliable for planning effective monitoring efforts.
The introduction of bias into ground-water data sets by sampling
mechanisms has been investigated by several groups (13,53). in a con-
trolled laboratory evaluation (53), results disclosed that sampling for
dissolved gases or volatile organic compounds is prone to severe nega-
tive bias of the same order as analytical error. Further, the precision
which may be achieved in these cases is limited by both operator skill
96
-------
ana sampler design. The magnitude of these errors corresponded to the
extent to which control over conditions during sample transfer steps
(i.e. flow rate, atmospheric exposure, turbulence) could be maintained.
Positive displacement bladder pumps were found to be the most reliable
sampling mechanism evaluated since they are simple in design and opera-
tion and operational variables are easily controlled.
Similarly, Korte and Kearl (73) recommended positive displacement
bladder pumps over bailers, suction-lift and air-lift devices due to the
bladder pumps' range of utility, minimal disturbance of the samples and
overall simplicity of operation. They also noted that bladder pumps
permit efficient in-line filtration of samples in the field.
Modifications of selected sample collection mechanisms are being
developed to improve the reliability and applicability of ground-water
chemical data. Armstrong and McLaren (7*0 have refined pump/packer
arrangements to optimize the isolation of the sample intake as well as
sample recovery. Pankow et al. (40) have investigated the application
of in situ sample collection techniques for organic compounds which have
the advantage of minimizing sample exposure to either the atmosphere or
foreign materials. The routine application of these and other refine-
ments for ground-water sampling efforts (75) must await further develop-
ment .
Work to date has established that there is a great need for the
field evaluation of sampling mechanisms. This work has also identified
the capabilities which a reliable sampling mechanism should provide .
Important characteristics of ground-water sampling devices which
should be considered are:
97
-------
1. The devices should be simple to operate in order to
minimize the possibility of operator error.
2. The device should be rugged, portable, cleanable and
repairable in the field.
3- The device should have good flow controllability to permit
low flow rates (<100 mL/min) for sampling volatile
chemical constituents, as well as high flow rates (>1
L/min) for large-volume samples and purging stored
water from monitoring wells.
4. The mechanism should minimize the physical and chemical
disturbance of ground-water solution composition in
order to avoid bias or imprecision in analytical
results.
The scientific literature is somewhat inconsistent in descriptions
of the types of samplers and their primary mechanisms of operation. In
this regard, gas-lift mechanisms are exemplified by down-hole dual tube
arrangements, which employ violent gas/water mixing to force water up
and out of the well bore (or large diameter tube). Gas lift devices are
proven to be biased sampling mechanisms for a range of chemical con-
stituents. They are not recommended for any type of ground-water
investigation. Gas-drive devices, on the other hand, rely on controlled
displacement of water from the sampler body by either controlled gas
pressure applied across an interface or by gas pressure on a membrane
which permits for no gas contact with the sample.
Many sampling devices are designed for either deployment in a well
bore or as devices which are buried at discrete depths, gravel/sand
98
-------
packed and sealed from other formations analogous to a. properly com-
pleted screened interval in conventional wells. There are advantages to
the use of dedicated samplers, particularly for complex monitoring
situations which demand large arrays of sampling points. The corres-
ponding disadvantages include difficulties in assessing proper placement
or malfunction. The collection of hydrologic data is severely limited
by most of these devices. The choice of a sampler design, appropriate
for the situation of interest, should be made carefully after a compre-
hensive review of the scientific literature.
Recommendations for Selecting Sampling Mechanisms
It should be recognized that the purchase of a suitable sampler for
most ground-water investigations is usually a very small portion of the
overall program cost. It is further obvious that the choice of the
right sampler will determine the usefulness of the chemical data. A
sensible approach is to make the choice of a sampler on the basis of the
most troublesome parameters which may be of interest. Typically,
samples for dissolved gases and volatile organic compounds are the most
difficult to collect and handle..
Negative bias (loss of constituent) is the most common reason for
poor sampling performance for gas-sensitive or volatile compounds. In
general, sampling precision may be poorer by a factor of two or more
than that involved in analytical methodologies alone. Sampling bias
problems may be far worse under field conditions, particularly for
suction mechanisms and those devices which involve careful operator
attention or control (e.g. bailers and gas drive devices). Positive
displacement bladder pumps meet all of 'the important characteristics for
99
-------
sampling mechanisms noted above. These pumps have been found to be very
reliable, efficient sampling mechanisms which exhibit excellent overall
performance in all reported evaluations to date.
Table 2.4 contains general recommendations for ground-water
sampling mechanisms. It should be noted that it is the responsibility
of the monitoring program director to build in sampling performance
checks into the QA/QC program to verify actual efficiency for the
chemical constituents of interest.
Sampling Materials
There exists a wide range of biological, chemical and hydrologic
conditions which may be encountered in ground-water sampling programs.
Even if personnel safety is assured, ground-water sampling activities
must be approached cautiously. There are many chemical and physical
unknowns which must be accounted for, if the monitoring data is to be
truly useful. Sampling mechanisms are only an element of sampling
protocols, the materials which contact the samples must be chosen care-
fully as well.
Subsurface Conditions and Materials' Effects
The "Guide to the Selection of Materials for Monitoring Well
Construction and Ground-Water Sampling" (52) and the thorough treatment
of sources of sampling bias by Gillham, et al. (13) provide very useful
recommendations for materials' selection and error minimization for
ground-water investigations. Both of .these publications review the
potential obstacles to materials' related error control.
Well casing materials, well construction and completion procedures
and sample handling precautions all enter into the ultimate quality of
100
-------
•o
C
8
o
g
aluati<
>
0)
o
§
i
t
4n
!U
a.
•
OJ
s_
0}
S
CQ
1
C
0)
z:
60
J^
&
1
CO
Remarks
de both efficient well purging
e samples over a range of conditions
iculty in field operations.
•rH > 4-4
> — 1 4-1
O 4J -H
t. CO T3
0. J-J
C rH
O 'CD CO
JJ 03 S
13 i. 'c
CD Q. 'H
i) CD E
O L
CD £
O. 13 JJ
X C "H
Cd tfl 3
CO
C
c
CO
rH £•
rH
CO CD
t. 0
CO C
> to
o s
!_
o
r_
S
'
eg
oi
0)
rt
CD
o
.0
>>
I-
o
M
cd
o
•H
ct
f-
C
d
4-3
C
0)
o
03
rH
Q.
m
*G
Q) t.
> 0)
-P T3
•H^OJ
rn r-1
O -O
11 purging; require very careful
pie handling precautions under
field performance open to question.
CD
3
b
4-4
CD
rH
Unsuitab
CO
CO
to
CO
4J
^-^
i.
CD
rH
a
CO rH
t, to
CD C
rH C
E ^
c3 c
CO tt
Q C
CO C
o — •
03 C
o
•a — i
C J->
tfl tH
13
c c
operatio
field co
CO
£
CD
CO
CD 3
CO O
CO rH
t, co
CO £>
> 1
«S 1
CD
£
CD
^_^
t_
CD
r-t
CO
CD
rH
to *~*
> 2-
O Q.
CO
s: CD
o co
1 C
in
•o to
purging; sampling performance very
ific design and operational details.
rH O
rH CO
CD O.
3 03
£, C
<2 °
Suitable
dependen
CD
CO
CO
CD
tfl
3
O
rH
CD
JD
1
1
CD
CO
CO
t,
CO
"*
j_i
C
CD
CD
0
to
rH
O,
CO ^
13 10
O
^ C
.H CO
-rf O
CO CD
O S
'CU **-•'
or well purging if used in .conven-
ons; malfunctions are difficult to
significantly lower recoveries of
compounds and gases may occur
d conditions and operator experience
4-1 -rH »»
4-> I_
CD tO 'rH
rH rH tO
.Q rH Q.
fO CO CD
4J 4-> E,
-r-4 CO
3 C £.
CO "H O
CO rH 03
.O CO CO
C CO
>> O CO
CO 'rl CO
S J-> CO
CD
CO
tO
t,
CD
to
g
rH
CO
1
1
CD
CO
tO
CD
^_^
JJ
4H
rH
03
CO
to
s -s
CO O
E C
CO
o —
to CD
rH !>
a. ^
CO L
•r< 13
13
to co
O — '
o
tH
c
tfl
CO
t.
o
CD
CD
4-1
c
o
CO
purgeabl
dependin
purging at depths to approximately
antly lower recoveries of purgeable
and gases will result from sampling
sm.
rH O
rH "H
CD 4-1
3 "3
i. CO
O "H
4-i CO
Suitable
20 feet;
2J,
to
CO
*
S
^
O
rH
0)
CQ
^™*
O
•H
jj
rH
(0
C •*-
O . W
•H -H
o a
3 0
CO **-•
OT -H
T3 C
§S
ao
CD
O CO
•H
O ^
iH 4J
C
tfl JS
to J->
t. -H
O 3
101
-------
ground-water data. The principal processes by which materials can
effect chemical data are:
chemical attack: corrosion/deterioration
microbial colonization, attack
sorption effects: adsorption/absorption
leaching effects: matrix/sorbed component release
These processes may lead to the observation of false trends in
analyte concentrations, highly variable water chemistry and the identi-
fication of artefacts resultant from surface release or sorptive inter-
actions. As with the errors which sampling mechanisms can introduce
into the chemical data,'materials' related errors can be quite signifi-
cant and difficult to predict (23,52). Appropriate materials' choice
for each application must be made on the basis of long-term durability,
cleanability, and the minimization of the secondary effects of sorption
or leaching. Structural integrity is, therefore, the primary criterion
for making reliable material choices. The materials must neither be
attacked nor degraded during the course of the monitoring program. Then
the severity of the effects of loss or contamination resulting from
sorption or leaching of the components of the sampling train must be
considered. In general, it is wise to base materials' choices on the
most error prone constituents of interest.
To evaluate the magnitude of materials' effects, it is instructive
to consider the relative surface area contact which aquifer solids,
well casing and sampling tubing will have with the water samples. Table
2.5 contains a comparison of these materials and their relative surface
area contact under monitoring well sampling conditions. Assuming rela-
tively high linear ground-water velocity (50 cm«d~1) and pumping rates
of -100 mL/min, it follows that aquifer solids are a potentially
102
-------
co
rH
a)
0>
CO
IE
•O
0)
O
0)
CO
S
c
o
co
•H
f.
cd
0
•^^
o
c«
^5
c
o
o
0)
rH
O.
CD
CO
0)
^
•H
(0
«
OS
•
in
CM
0)
1
H
CD
*
(D 4-5
•H O O
4-5
CO <<
rH
0)
OS
4-5
O
rH CO
CO 4-5
CD C
c o
< o
S
rH
>
<
cO CO
M 4J
t- C
3 O
CO 0
^^
T
CD J3
P •
CO CM
^3
CM r- ^>
CT>
CM
t~- 0
**O CO -3"
VO
ro
i
0
Q)T
4^ SL,
co si
t« *
X *£> **O
=f
o
il-
X— %
^_
1
> J
v^ •
^^
o
CM t—
>4 r- \&
• •
on o o
V£3
fc^
0) 4-5
5 -H
3 CO
•H O
0 t.
> O
a
4-5 O 4J
LO "in 0 .4-1
^ • V. • "V
O ** O «J O J
un
v»x
'
CD
o
cO cO
CM CD
t- L
3 CO
CO
4J VO 4-5
\ 0 >v 0 \
CM .CM 'CM
in E OS OS
CM
•a
•H
rH
O
co
^-N
1
•a
S
° 0
CO 0 2!
a in *-•
IH 03
J II O O •
O Z E- Q
CO 3 IH • •
^ CO Q O 0
OS •* • SS
W T3 O O IH S
ry . j^ t ^ j^*
|— | CO J S Ba ^^
O CO J CM S T—
at u <
Q)
4-^
CO
L.
4-5
o
cO
c
o
o
rH
cO
Q)
CO
P, t
O
CD
3
rH
CO
60
CO
(0
O
iH
rH
-------
greater source of material surface effects on water quality than either
well casing or sampling tubing. Since we cannot exert control over the
native geology, the effects of gravel packs and grouting materials may
be expected to be more important than those due to well casing or
sampling train materials. Relative to conditions in the geologic forma-
tion of interest, sampling tubing is in much more intimate contact with
the water sample collected after proper purging than would be the well
casing. One should not assume that materials' effects will cancel out
in comparisons of upgradient or downgradient monitoring well data.
Materials' related bias will be present in all samples though perhaps
not to the same extent. The effects of materials in comparisons of
sample results are most pronounced under differing chemical conditions
where some materials may be attacked or leached to varying degrees.
Purging may minimize the effects of potential well casing interferences,
however this is difficult to substantiate under field conditions.
Sampling train components, particularly sampling tubing, are the
most critical selections which must be made to avoid materials' related
error. A recent study has demonstrated that serious bias of dissolved
organic compound results occurs quite rapidly (within 5-10 minutes) due
to absorption on flexible tubing exposures (76). In this study all
commonly used tubing materials (Teflon(R), polypropylene, polyethylene,
etc.) sorbed organic compounds to some extent. The sorptive error was
most serious for polyvinyl chloride and silicone rubber tubing.
Recommendations for Selecting Sampling Materials
The primary criteria for the selection of materials for all compo-
nents of the sampling point and sample collection train should be mech-
104
-------
anical performance and chemical inertness. Since the actual subsurface
geologic and chemical conditions which may be encountered are very
difficult to predict, the choice of materials must be made carefully.
It is recommended that sampling components be chosen of the most inert
and error-free materials available. The costs of analysis (or repeat
analyses) and the labor involved in sample collection are generally much
u.
higher than the cost of appropriate materials for sampling ground water.
Sampling materials may be categorized as either rigid or flexible.
Rigid components include well casing, pump bodies and fittings, while
tubing, bladders and gaskets are generally flexible materials.
Tables 2.6 and 2.7 detail general recommendations for rigid and flexible
materials, respectively. Teflon(R) components are superior to all other
materials' combinations for ground-water sampling. The mechanical per-
formance of this material may require that it be used in combination
with stainless steel. The available literature on materials evaluation
for sampling ground water substantiates these recommendations (23,52,
72,76).
SAMPLE COLLECTION PROTOCOL
A well conceived sampling protocol consists of a written descrip-
tion of the actual sampling and analytical procedures involved in
obtaining representative ground-water data. The protocol must reflect
special attention to the need to collect high quality hydrologic data
(e.g. water level, hydraulic conductivity, etc.) and to record any
unusual occurrences or departures from written procedures. The value of
water quality measurements has been emphasized repeatedly in the- litera-
ture. However, it is very difficult to fully interpret the water
105
-------
Table 2.6. Recommendations for Rigid Materials in Sampling Applications
•(In decreasing order of preference)
Material
Teflon(R)
(flush threaded)
Recommendations
Recommended for most monitoring situations with
detailed organic analytical needs, particularly
for aggressive, organic leachate impacted
hydrogeologic conditions. Virtually an ideal
material for corrosive situations where
inorganic contaminants are of interest.
Recommended for most monitoring situations with
detailed organic analytical needs, particularly
for aggressive, organic leachate impacted
hydrogeologic conditions.
May be prone to slow pitting corrosion in
contact with acidic high total dissolved solids
aqueous solutions. Corrosion products limited
mainly to Fe and possibly Cr and Ni.
Recommended for limited monitoring situations
where inorganic contaminants are of interest and
it is known that aggressive organic leachate
mixtures will not be contacted. ' Cemented
installations have caused documented inter-
ferences. The potential for interaction and
interferences from PVC well casing in contact
with aggressive aqueous organic mixtures is
difficult to predict. PVC is not recommended
for detailed organic analytical schemes.
Recommended for monitoring inorganic
contaminants in corrosive, acidic inorganic
situations. May release Sn or Sb compounds from ,
the original heat stabilizers in the formulation
after long exposures.
(R) Trademark of DuPont, Inc.
* National Sanitation Foundation approved materials carry the NSF logo
indicative of the product's certification of meeting industry standards
for performance and formulation purity.
Stainless Steel 316
(flush threaded)
Stainless Steel 304
(flush threaded)
PVC
(flush threaded)
other noncemented
connections, only NSF*
approved materials for
well casing or potable
water applications
106
-------
Material
Low-Carbon Steel
Galvanized Steel
Carbon Steel
Table 2.6. (Continued)
Recommendations
May be superior to PVC for exposures to
aggressive aqueous organic mixtures. These
materials must be very carefully cleaned to
remove oily manufacturing residues. Corrosion
is likely in high dissolved solids acidic
environments, particularly when sulfides are
present. Products of corrosion are mainly Fe and
Mn, except for galvanized steel which may
release Zn and Cd. Weathered steel surfaces
present very active adsorption sites for trace
organic and inorganic chemical species.
107
-------
Table 2.7. Recomendations for Flexible Materials in Sampling Applications
(In decreasing, order of preference)
Materials
Teflon(R)
Polypropylene
Polyethylene (linear)
PVC (flexible)
Silicone
(medical grade only)
Neoprene
Recommendations
Recommended for most monitoring work,
particularly for detailed organic analytical
schemes. The material least likely to introduce
significant sampling bias or imprecision. The
easiest material to clean in order to prevent
cross-contamination.
Strongly recommended for corrosive high
dissolved solids solutions. Less likely to
introduce significant bias into analytical
results than polymer formulations (PVC) or other
flexible materials with the exception of
Teflon(R).
Not recommended for detailed organic analytical
schemes. Plasticizers and stabilizers make up a
sizable percentage of the material by weight as
long as it remains flexible. Documented
interferences are likely .with several priority
pollutant classes.
Flexible elastomeric materials for gaskets,
0-rings, bladder and tubing applications.
Performance expected to be a function of
exposure type and the order of chemical
resistance as shown. Recommended only when a
more suitable material is not available for the
specific use. Actual controlled exposure trials
may be useful in assessing the potential for
analytical bias.
(R) Trademark of DuPont, Inc.
108
-------
chemistry or the actual extent of contamination unless high quality
hydrologic data is collected and interpreted properly. Indeed, it may
be advisable to collect the hydrologic data at more frequent intervals
and at finer spatial scale than that used for the chemical data.
The principal steps in the sampling protocol are listed in
Figure 2.16. The goal for each step is also provided with a general
recommendation for achieving it. These general elements are common to
all ground-water sampling efforts. It should be the responsibility of a
designated member of the sampling staff to record progress through the
protocol at each sampling point.
To insure maximum utility of the sampling effort and resulting
data, documentation of the sampling protocol as performed in the field
is essential. In addition to noting the obvious information (i.e.,
persons conducting the sampling, equipment used, weather conditions, and
documentation of adherence to the protocol and unusual observations)
three basic elements of the sampling protocol should be recorded:
1) water level measurements made prior to sampling, 2) the volume and
rate at which water is removed from the well prior to sample collection
(well purging), and 3) the actual sample collection including measure-
ment of well-purging parameters, sample preservation, sample handling
and chain of custody.
Water Level Measurement
Prior to the purging of a well or sample collection, it is
extremely important to measure and record the water level in the well to
be sampled. Water level measurements are needed to estimate the amount
of water to be pumped from the well prior to sample collection. In
109
-------
Step
Goal
Recommendations
Hydrologic
Measurements
Well Purging
Sample Collection
Filtration/
Preservation
Field Determinations
Field Blanks/
Standards
Sampling Storage/
Transport
Establish nonpumping water
level.
Removal or isolation of
stagnant H20 which
would otherwise bias
representative sample.
Collection of samples
at land surface or in
well-bore with minimal
disturbance of sample
chemistry.
Filtration permits
determination of soluble
constituents and is a
form of preservation. It
should be done in the
field as soon as possible
after collection.
Field analyses of samples
will effectively avoid
bias in determinations of
parameters/constituents
which do not store well:
e.g., gases, alkalinity,
pH.
These blanks and standards
will permit the correction
of analytical results for
changes which may occur
after sample collection:
preservation, storage, and
transport.
Refrigeration and protec-
tion of samples should
minimize the chemical
alteration of samples
prior to analysis.
Measure the water level to
±0.3 cm (±0.01 ft).
Pump water until well
purging parameters (e.g.,
pH, T, JT1, Eh) stabilize
to ±1055 over at least two
successive well volumes
pumped.
Pumping rates should be
limited to -100 mL/min
for volatile organics and
gas-sensitive parameters.
Filter; Trace metals,
inorganic anions/cations,
alkalinity.
Do not Filter; TOC, TOX,
volatile organic compound
samples; other organic
compound samples only
when required.
Samples for determina-
tions of gases, alkalinity
and pH should be analyzed
in the field if at all
at all possible.
At least one blank and
one standard for each
sensitive parameter
should be made up in the
field on each day of
sampling. Spiked samples
are also recommended for
good QA/QC.
Observe maximum sample
holding or storage
periods recommended by
the Agency. Documentation
of actual holding periods
should be carefully per-
formed.
Figure 2:16. Generalized ground-water sampling protocol
110
-------
addition, this information can be useful when interpreting monitoring
results. Low water levels may reflect the influence of a nearby produc-
tion well. High water levels compared to measurements made at other
times of the year are indicative of recent recharge events. In rela-
tively shallow monitoring settings high water levels from recent natural
recharge events may result in dilution of the total dissolved solids in
the collected sample. Conversely, if contaminants are temporarily held
in an unsaturated zone above the geologic zone being monitored, recharge
events may "flush" these contaminants in the shallow ground water system
and result in higher levels of some constituents.
Documenting the nonpumping water levels for all wells at a site
will provide historical information on the hydraulic conditions at the
site. Analysis of this information will reveal changes in flow paths
and serve as a check on the effectiveness of the wells to monitor
changing hydrologic conditions. This information is also essential to
develop an understanding of the seasonal changes in water levels and
associated chemical concentration variability at the monitored site.
Purging
The volume of stagnant water which should be removed from the moni-
toring well should be calculated from the analysis of field hydraulic
conductivity measurements. Rule of thumb guidelines for the volume of
water which should be removed from a monitoring well prior to sample
collection ignore the actual hydraulic performance of the sampling
point. These 3~, 5- or 10-well volume purging guidelines are a
liability in time, expense and information return from the sampling
activities.
Ill
-------
The calculated well purging requirement should also be monitored in
the field by the in-line monitoring of the well purging parameters (e.g.
Eh, pH, T, and n~1). In-line measurements provide the most representa-
tive data for these constituents and verify the reliability of the
hydraulic evaluation of the sampling point or well (2,77). These
chemical constituents further aid in the interpretation of water quality
changes as they are affected by hydrologic conditions. Modifications to
the electrode cell in these flow-through instruments have resulted in
their improved performance in the field (78). A photograph of this
instrument is provided in Figure 2.17.
For example, the calculated well purging requirement (e.g. >90%
aquifer water) calls for the removal of five well volumes prior to
sample collection for a particular well. Field measurements of the well
purging parameters have historically confirmed this recommended proce-
dure. During a subsequent sampling effort, twelve well volumes were
pumped before stabilized well purging parameter readings were obtained.
Several possible causes could be explored: 1) A limited plume of
contaminants may have been present at the well at the beginning of
sampling and inadvertently discarded while pumping in an attempt to
obtain stabilized indicator parameter readings; 2) The hydraulic prop-
erties of the well have changed due to silting or encrustation of the
screen indicating the need for well rehabilitation or maintenance;
3) The flow through device used for measuring the indicator parameters
was malfunctioning; or H) The well may have been tampered with by the
introduction of a contaminant or relatively clean water source in an
attempt to bias the sample results.
Documentation of the actual well purging process employed should be
a part of a standard field sampling protocol. Figure 2.18 presents a
one-page form which may be used for documenting field sampling
operations at each sampling point.
Sample Collection
The initial hydrologic and well purging measurements necessary for
reliable ground-water sampling should be entered into the same field
112
-------
Figure 2.17. A well-head instrumentation package for Eh, pH,
conductivity and temperature measurements
113
-------
GROUND WATER SAMPLING RECORD
Facility name
Hell number
Date
Well depth
Well diameter
Casing Matl
Sampling crew
Type of purap
tubing
Weather conditions
Time
Water Pump Volume Pumping Sample Temp
level on pumped rate start/end (°C)
Cond
Eh £H (uS)
Sample delivered to
By
Figure 2.18. Suggested recording format for well purging
and sample collection
114
-------
notebook as that of the discrete samples for field or laboratory deter-
minations. Regardless of the level of analytical detail in the moni-
toring program, it is essential that all samples be collected properly
and that the actual conditions during each sample collection are com-
pletely documented. One member of the sampling staff should be desig-
nated as responsible for this documentation.
The format for documentation should be clear and constant during
the overall program. A set of useful forms for field collection and
measurement are presented in Tables 2.8 and 2.9. They are largely self-
explanatory. It is useful to standardize the format, particularly where
field personnel are responsible for splitting samples for field spikes
or blind control samples. It is recommended to inscribe the bottles
with an identifying marking which, when combined with the date of
sampling, will uniquely identify it in a sampled set.
Water samples should be collected when the solution chemistry of
the ground water being pumped has stabilized as indicated by pH, Eh, JT1
and T readings. In practice, stable sample chemistry is indicated when
the purging parameter measurements have stabilized with ±1056 over 2
successive well volumes. First, samples for -volatile constituents, TOG,
TOX and those constituents which require field filtration or field
determination should be collected. Then large volume samples for extrac-
tible organic compounds, total metals or nutrient anion determinations
should be collected.
All samples should be collected as close as possible to the well
head. A "tee" fitting placed ahead of .the in-line device for measuring
the well purging parameters makes this more convenient. Regardless of
115
-------
01
0>
1— 1
o.
3
CO
•o
4->
H
•^
fa
bO
O
J
Q)
r-l
a
§
CO
o
•r*l
c
«
to
o
c
M
•
CO
CM
0)
i— {
ft
n
c
o
^>
~4 Ci]
n a.
u c
Q) ^
0)
a
m
^
g
o
*4
o a.
a
8
3
•H
O
S*
^
n
n
o
C Cd
is a.
a
o £
1?
s:
_:
0)
E
3
^
O
>
*l
>>
jj
c -a
*H »-1
M CO
a -H
.* [24
•i
*~^
CO
E
3
i-H
O
>
§
o
o
o o
o m
i- CM
8§
•- CM
§
o
o
oo o
in o z
i- X
e
a.
a.
c
o
o
t
o.
o
§
o
i
•a
0
a>
n
o
CO
o o
o in
— CM
O O
o o
*- CM
-------
^—N
>>
k
0}
n
Q)
o
0)
2
OH
M
^
•o
0)
s.
Q)
-P
rH
•H
En
1
S^X
to
o
J
0)
, rH
O.
s
CO
o
iH
§
60
I,
o
•
Oi
OJ
Q)
rH
«
H
^
o
*j
o>
m
n
co
rH
«
jj
o
n)
t.
-p
3
<
n
a
£
n
4-
C
n
!_
>
u
n
0)
rH
• iH
4J
n)
rH
O
>
X
o
E-
&
E-
j
C
^"3
<0
B
3
rH
0
~" ^_^
0] o
n a
3 5
M 00
i. 0
9> c
2 o
g _ 0
£0 ^s
^3 *-s
03 J
CO
CO CO
n e
3
Id rH
K O
t- >
*-^
a.
Cu
o o
Q
c
0
. o
3
rH
0
>
3
O«
0.
^^ o o
o
c
o
•**-! u
CO
n 3
rH rH
nJ o
^ ^-N
« -Q
OT Q.
8 S
M ° °
i i
1*1 "
d>
J S
S3
0 0
^
8
0.
a.
o o
' 0
c
o
£_j
i, §§8S §SS8 §|S§ §||S ||
M ^
co *« 2
r^ ^ J «
2SE §g §£ 5£ §S S
t
in o in o
cvj in CM in
jj
(TJ
i_
Q) O
M o
«-( -=r
C-
CM
e
•a rH O
c a. -^
oJ S *r
4^ CO <«
CO CO >
-O jJ 0)
,H >H n
a> ^ co
•iH CX t.
Cu W O-
117
-------
the sampling mechanism in use or the components of the sampling train,
upgradient wells should be sampled first followed by the downgradient
wells to minimize the potential for cross-contamination. Laboratory
detergent solutions and distilled water should be used to clean the
sampling train between samples. An acid rinse (0.1 N HC1) or solvent
rinse (i.e. hexane or methanol) should be used to supplement these
cleaning steps if necessary. All cleaning should be followed by dis-
tilled water rinses which may be saved to check cleaning efficiency.
Adhesive labels or "indelible" markers can present sample identifi-
cation problems, particularly when a variety of samples, split-samples,
standards and blanks are transported in ice chests. The markings can be
floated off or abraded into an illegible condition during transit.
Serious problems in sample handling and storage can result if
)•
extreme care is not taken during transport and storage. All ice, ice
packs, and ice-chests should be prepared in areas that are remote from
reagent and solvent storage of any kind! Further, the interim storage
of these materials should also be remote from reagent or solvent storage
areas. These precautions will minimize the effect of contamination
errors on the results (79).
Filtration
There are1 instances which arise, even with properly developed
monitoring wells, that call for the filtration of water samples. It
should be evident, however, that well development procedures which
require two to three hours of bailing, swabbing, pumping or air purging
at each well will save many hours of time in sample filtration. Well
development may have to be repeated at periodic intervals to minimize
118
-------
the collection of turbid samples. In this respect, it is important to
minimize the disturbance of fines which accumulate in the well bore.
This can be achieved by careful placement of the sampling pump intake at
the top of the screened interval, low pumping rates, and by avoiding the
use of bailers (60).
It is advisable to refrain from filtering TOG, TOX or other organic
compound samples as the increased handling required may result in the
loss of chemical constituents of interest. Allowing the samples to
settle prior to analysis followed by decanting the sample is preferable
to filtration in these instances. If filtration is necessary for the
determination of extractable organic compounds, the filtration should be
performed in the laboratory by the application of N2 pressure. It may
be necessary to run parallel sets of filtered and unfiltered samples
with standards to establish the recovery of hydrophobia compounds when
samples must be.filered. All of the materials' precautions used in the
construction of the sampling train should be observed for filtration
apparatus. Vacuum filtration of ground-water samples is not recommended.
Water samples for dissolved inorganic chemical constituents (e.g.
metals, alkalinity and anionic species) should be filtered in the field.
The preferred arrangement is an in-line filtration module which utilizes
sampling pump pressure for its operation. These modules have tubing
connectors on the inlet and outlet parts and range in diameter from
2.5-15 cm. Large diameter filter holders, which can be rapidly dis-
sembled for filter pad replacement are the most convenient and efficient
designs (80,81).
119
-------
The choice of filter media must be made on the basis of its
exposure to the water samples and the degree of analytical detail
required for those samples. Clearly, water samples which may be
contaminated with organic solvents limit the use of organic filter
media, such as cellulose nitrate, cellulose acetate or polycarbonate
filters. In these cases glass fiber or Teflon(R) filter media should be
used. Glass fiber filters should be acid rinsed followed by distilled
water rinsed prior to their use for filtering trace metal or nutrient
samples. Once an appropriate filter media has been selected, it is
advisable to choose a 0.45 yM nominal sized filter. The final selection
of the material and type of filter pad should be made carefully, as
there are considerable differences between "screen" or "depth" filtra-
tion media (82). Screen filters are typically less than 50 pM thick
(e.g. polycarbonate filters) which tend to load up and clog more rapidly
than the depth-type filters. Sampling staff should be trained in proper
procedures for filter pad replacement, since fine particles can easily
be transferred to the outlet side of a dissembled filter module. Sloppy
technique may result in solids breakthrough and biased samples. After a
filter pad is charged, the initial 50-100 mL should be discarded as a
rinse. Even if very careful procedures are followed, clogging and
small particle breakthrough are real problems which must be addressed on
a case by case basis (82,83).
Field Versus Laboratory Determinations
Representative sampling results from the execution of a carefully
planned sampling protocol which establishes necessary hydrologic and
120
-------
chemical data for each sample collection effort. An important consider-
ation for maintaining'sample integrity, after collection, is to minimize
sample handling which may bias subsequent determinations of chemical
constituents. Since opportunities to collect high quality data for the
characterization of site conditions in time may be limited, it is
prudent to conduct sample collection as carefully as possible from the
outset. It is preferable to bias data on the conservative side when
doubt exists as to the sensitivity of specific chemical constituents to
sampling or handling errors. Repeat sampling or analysis cannot make
up for lost data collection opportunities.
Samples collected for specific chemical constituents may require
modifications of recommended sample handling and analysis procedures.
Matrix effects and extended storage periods can cause significant
problems in this regard. It is frequently more effective to perform a
rapid field determination of specific inorganic constituents (e.g.
alkalinity, pH, ferrous iron, sulfide, nitrite or ammonium) than to
attempt sample preservation followed by laboratory analysis of these
samples. There are several good references to guide the development of
field analytical procedures (1,2,31). Korte and Ealey (84) have pre-
pared a useful field analytical guide. However, their recommendation
not to filter alkalinity samples is not supported by the literature.
pressure filtration is necessary to insure that the alkalinity results
are reliable for subsequent calculations of solution chemistry equi-
libria (85).
Criteria for the selection of appropriate analytical metho-ds vary
somewhat and the degree of analytical detail required for ground-water
121
-------
monitoring programs is increasing.. It is advisable to select field and
laboratory analytical methods carefully after consultation with the
proper authorities. One should keep in mind that methods for drinking
water or wastewater may encounter significant interferences when applied
to contaminated ground-water samples.
Blanks, Standards and Quality Assurance
The use of field blanks, standards and spiked samples for field
QA/QC performance is analogous to the use of laboratory blanks,
standards and procedural or validation standards. The fundamental goal
of field QC is to insure that the sampling protocol is being executed
faithfully and that situations leading to error are recognized before
they seriously impact the data. The use of field blanks and standards
and spiked samples can account for changes in samples which occur after
sample collection.
Field blanks and standards enable quantitative correction for bias
(i.e., systematic errors), which arise due to handling, storage, trans-
port and laboratory procedures. Spiked samples and blind controls pro-
vide the means to correct combined sampling and analytical accuracy or
recoveries for the actual conditions to which the samples have been
exposed. All QC measures should be performed for at least the most
sensitive chemical constituents for each sampling date. Examples of
sensitive constituents would be: benzene or trichloroethylene as
volatile organic compounds and lead or iron as metals. It is difficult
to use laboratory blanks alone for the determination of the limits of
detection or quantitation. Laboratory distilled water may contain higher
levels of volatile organic compounds (e.g. methylene chloride) than
122
-------
those of uncontaminated ground-water samples. The field blanks and
spiked samples should be used for this purpose, conserving the results
of lab blanks as checks on elevated laboratory background levels. The
usefulness of spiked samples should be obvious. Whether the ground-
water is contaminated with interfering compounds or not, these samples
provide a basis for both the identification of the constituents of
interest and the correction of their recovery (or accuracy) based on the
recovery of the spiked standard compounds. For example, if trichloro-
ethylene in a spiked sample is recovered at a mean level of 80$ (-20%
bias), the concentrations of trichloroethylene determined in the samples
for this sampling date may be corrected by a factor of 1.2 for low
recovery. Similarly, if 50% recovery (-50* bias) is reported for the
spiked standard, it is likely that sample handling or analytical proce-
dures are out of control and corrective measures should be taken at
once. It is important to know if the laboratory has performed these
corrections or taken corrective action when they report the results of
analyses. It should be noted that many regulatory agencies require
evidence of QC and analytical performance but do not generally accept
data which has been corrected.
Field blanks, standards and blind control samples provide inde-
pendent checks on handling and storage as well as the performance of the
analytical laboratory. It should be noted that ground-water analytical
data is incomplete unless the analytical performance data (e.g. accu-
racy, precision, detection, and quantitation limits) are reported -along
with each set of results. Discussions of whether significant changes in
123
-------
ground-water quality have indeed occurred must be tempered by the accu-
racy and precision performance for specific chemical constituents.
Table 2.10 is a useful guide to the preparation of field standards,
and spiking solutions for split samples. It is important that the field
blanks and standards be made on the day of sampling and are subjected
to all conditions to which the samples are exposed. Field spiked
samples or blind controls should be prepared in the field by spiking
with concentrated stock standards in an appropriate background solution.
The choice of spiking solution is particularly critical where volatile
organic compounds are of concern (e.g. TOG, TOX and purgeables). In
this case, pure poly(ethylene glycol) or water:poly(ethylene glycol)
mixtures are very useful (86). The use of methylene chloride as a
standard compound should be avoided. Additional precautions should be
taken against the depressurization of samples during air transport and
the effects of undue exposure to light during sample handling and
storage. All of the QC measures noted above will provide both a basis
for high quality data reporting and a known degree of confidence in data
interpretation. Well planned quality control programs will also mini-
mize the uncertainty in long-term trends when different personnel have
been involved in sample collection and analysis.
Sample Storage and Transport
The storage and transport of ground-water samples are often the
most neglected elements of the sampling protocol. Due care must taken
in sample collection, field determinations and handling. If proper
planning of transport is neglected the samples may be stored for long
periods before laboratory analysis. Every effort should be made to
124
-------
8
rH
O.
O 01
[ -S
C -rH
0 H
4^ CO
3
rH CM
0 0
CO
cu
O T-
0 D.
40 tj
co
•o
>
X CU
CO S
T3 rH
rH O
CU >
0
c to
o cu
cd c
L. O
J D.
C S
CD O
0 0
C
O
O
^
0
in
1
o.
o
o
o
in
CM
0
o
o
o
•ii
,_
1
o.
o
0
0
o
in
o
0
0
in
CM
•i
,—
B
o.
o.
o
0
0
o
r—
o
0
0
in
a1-
•—
1
o.
^rf»
o
o
o
in
CM
o
o
o
o
r—
3.
O
O B B
I Q. O.
~ §. s A
V 5 o o
a oo
o. o o o
D. O
•-• in
3.
O
in in
CM CM
B B B B
O. Q. O. O.
O. O, D. O,
in
..00
O O O
.. o in in
o co • -
o - CM CM
CM T- T- V-
o
in
o o
in in
in in in
CM CM CM
rH
o
o
in
in
CM
o o
CM CM
O
cd
I
o
CO
+
X
fl
o
CM
X
o
CM
X
to
* cu
rH §
O rH
£ Z
o -^
CM CU
* CU
>•> e
rH CU
O rH
o. >.
O 40
CM ^U
o
c
1
rH
§
cd
JC
cu
o o.
in *-*
CM
o
— in
o
o
o
•o
in
S
o.
o.
in
CM
o
o
o o
s e ~2 S
ex a. a. a.
o. o. o. a.
in
CD
in in in
• CM CM
— in in
CO
. CM CM
o r- •— «-
CM
O X> £>
D. D. D.
O. 0. O.
000
in in in
in in in
o
D..
a.
o
in
40
x
O
i-H
•
CN
Composition
\
CO
1
»
z
CD
rH
0
•i
0
in
cu
a
•^
cu
H
V.
c
•rH
rH
a
rH
1
II HH +
J J
,_ ,-
CO
CO C
c o
0 -H
•H 40
C cd
< 0
+ +
+ 00
O «5
D-.
+ " CM
+ -H
O
+." •)•
^ *f,
O CJ
J
r-
co
rH
cd
40
CD
B
CU
O
CD
t,
H
+ B
B t,
s o
CU HH
. C 0
+ O t,
+ 40 O
+ CD O. rH
CU CJ X Xl
fa < 2 0
"i 's
0 O
-5" "^"
0 X
8 B
1
t
X
cu
I
o
fs.
o
I
•rH
Q
Phenol Standards
•J
»-
•=s
CO
cu
cd
40
cd
i.
Cd
0
•H
40
Polynuclear Aroma
Standards
•J
*-
m
rH
S
•g
cd
c.
Standards
As Required
J
•"
o
ja
IB
o
cd
L
3
B
cd
hylene Glycol (100
thanol
40 CD
CD S
rH -O
o cu
O. rH
\ rH
i* "H
CU 40
40 Cf)
cd — i
s o
in co
CM CO
.. cd
in rH
t— o
n n
*
125
-------
inform the laboratory staff of the approximate time of arrival so that
. the most critical analytical determinations can be made within recom-
mended storage periods. This may require that sampling schedules be
adjusted so that the samples arrive at the laboratory during working
hours.
The documentation of actual sample storage and treatment may be
handled by the use of chain of custody procedures. An example of a
chain of custody form is shown in Figure 2.19. Briefly, the chain of
custody record should contain the dates and times of collection, receipt
and completion of all the analyses on a particular set of samples. It
frequently is the only record of the actual storage period prior to the
reporting of analytical results that exists. The sampling staff members
who initiate the chain of custody should require that a copy of the form
be returned to them with the analytical report. Otherwise, verification
of sample storage and handling will be incomplete.
Sample shipment arrangements should be planned to insure that
samples are neither lost nor damaged enroute to the laboratory. There
are several commercial suppliers of sampling kits which permit refrig-
eration by freezer packs and include proper packing. It may be useful
to include special labels or distinctive storage vessels for acid-
preserved samples to accommodate shipping restrictions.
126
-------
CHAIN OF CUSTODY RECORD
Sampling Date
Site Name
Well or Sampling Points:
Sample Sets for Each; Inorganic, Organic, Both
Inclusive Sample Numbers;
Company's Name Telephone (
Address
number street
Collector's Name
Date Sampled
city
state
Telephone (_
zip
Time Started
Time Completed
Field Information (Precautions, Number of Samples, Number of Sample
Boxes, Etc.):
1
name
organization
location
name
organization
location
Chain of Possession (After samples are transported off-site or to
laboratory):
___ (IN)
; (OUT)
(IN)
(OUT)
1.
2.
signature
title
name (printed)
signature
name (printed)
date/time of receipt
title
date/time of receipt
Analysis Information;
Aliquot
1 .
2.'
3.
Analysis Begun Analysis Completed
(date/time) Initials (date/time) Initials
5.
Figure 2.19. Sample chain of custody form
127
-------
SECTION 3
RECOMMENDED SAMPLING PROTOCOLS
The selection of methods and materials for drilling and well con-
struction, sampling and sample handling should be based on a complete
evaluation of site conditions, the analytes of interest and the informa-
tion needs of the program. Integrating all of these elements into a
reliable sampling protocol must be done in phases as information on the
/
actual conditions at a site is collected. Sampling mechanisms and mate-
rials are central to effective monitoring efforts. However, mechanisms
and materials' selections are only the basis for the development of the
sampling protocol. The preliminary protocol must be documented and all
personnel involved in the effort should be well acquainted with it.
Then the sampling protocol can be refined and targeted in development to
meet the critical information needs of the overall program.
In this section, specific recommendations are made for preliminary
sampling protocols applicable to both contaminant detection and assess-
ment programs. General guidelines are presented with a step by step
description of the procedures to develop of specific sampling protocols
for a variety of monitoring applications.
THE BASIS FOR SAMPLING PROTOCOL DEVELOPMENT
The individual elements of effective sampling protocols have been
reviewed in Section 2 of this guide. The generalized sampling protocol
presented in Figure 2.16 provides a review of the procedures undertaken
at each step. Figure 3.1 provides a prioritized schematic for the
execution of steps within the overall protocol which should guide the
128
-------
Step
Well Inspection
Well Purging
Sample Collection
Filtration*
Field
.Determinations**
Procedure
Hydrologic Measurements
+
Removal or Isolation of
Stagnant Water
*
Determination of Well-Purging
Parameters (pH, Eh, T, a"1)**
Unfiltered
Field Filtered*
Preservation
Field Blanks
Standards
Volatile Organ!cs, TOX
4-
Dissolved Gases, TOC
4-
Large Volume Samples for
Organic Compound
Determinations
Essential elements
Water-level Measurements
Representative Water Access
Verification of Representa-
tive Water Sample Access
Sample Collection by
Appropriate Mechanism
Minimal Sample Handling
Head-Space Free Samples
Minimal Aeration or
Depressurization
Assorted Sensitive
Inorganic Species
N02~, NHi, + , Fe(II)
(as needed for good
QA/QC)
Alkalinity/Acidity**
Trace Metal Samples
S=, Sensitive
Inorganics
4
Major Cation and
Anions
Storage
Transport
Minimal Air Contact,
Field: Determination
Adequate Rinsing Against
Contamination
Minimal Air Contact,
Preservation
Minimal Loss of Sample
Integrity Prior to Analysis
*Denotes samples which should be filtered in order to determine dissolved constituents.
Filtration should be accomplished preferably with in-line filters and pump pressure or by
N? pressure methods. Samples for dissolved gases or volatile organics should not be
filtered. In instances where well development procedures do not allow for turbidity-free
samples and may bias analytical results, split samples should be spiked with standards
before filtration. Both spiked samples and regular samples should be analyzed to
determine recoveries from both types of handling.
**Denotes analytical determinations which should be made in the field.
Figure 3.1. Generalized flow diagram of ground-water sampling steps
129
-------
planning of sampling efforts. Essential considerations for the relia-
bility of each step are also provided in the figure to aid planning
specific efforts. The planning should be coordinated with supervisory,
field, and laboratory staff.
Since the sampling mechanism provides the sample for further pro-
cessing, it is useful to consider the degree of analytical detail and
the reliability of specific sampling mechanisms before the remainder of
the protocol is developed. Figure 3.2 provides a matrix which allows
the comparison -of sampling mechanism reliability with the sensitivity of
various classes of constituents to sampling error. This matrix summa-
rizes the detailed recommendations provided in Section 2. Its use
should enable the initial choice of sampling mechanism which will serve
the planning needs for a preliminary sampling protocol. Once the choice
of sampling mechanism has been made, step-by-step sampling procedures
for specific monitoring applications may be designed.
Appropriate ground-water sampling procedures should be selected on
the basis of collecting the most reliable samples possible for the
specific analytes of interest. For purposes of discussion, one may
categorize monitoring efforts into two broad classes (i.e., detection
and assessment) according to the level of analytical detail sufficient
for the information needs of the program.
SAMPLING PROTOCOLS FOR DETECTION MONITORING
In detection monitoring efforts, the information needs are mainly
to detect ground-water contamination and to establish a set of useful
ground-water quality data in the event that contamination is detected.
130
-------
01
C 01
Suctio
mechani
01
> 0}
•H 01
S, 0
Q -H
1 >
01 01
to •a
CO
S
cr
y»
01 «J
> m i
••H Q, C.
•p s £
•*H 3 2
o ac
O, -P 2
rH g!l
Mechanical
displacem<
TTV niT QAMP
h
0
«
t. r
O 01 »
> 6
3 rH 0
J-> CO
•rt !> C.
m • m *
?£ J_ h
C 0 O C
•H 0> r-l «
X! ••-< C
•• ci to c
s
' 0) 01
o o
rH 2
Q. 0.
•H S,
•O 0)
•o
01 -a
> CO
~H rH
•U XI
01
O
O,
«-. C
O 01
3
Q) *J
rH *«H
0. 4->
B 01
ta o
Type of
constituent
•o
O
O 01
SS i.
•o
•a
c
s
£3
jj O
O 01
z: t-
c
0
Vl -H
0! 2 -
4J Ql *O
CO D, 01
3 O rH
0- rH
01 T3 O
•act.
c
CO 01 t.
S T3 CO
May be adequate, i
well purging is
assured
01
o
c
OS
£ o .
01 4*
a i.
3 01
co a.
i-H ss o oe u < co
B
fi
o
2 M
o a:
rH X 00
x: o ac
O E-" O
01
o
rH
rH
01 T3 01
rH 0 C B
-H "H 3 O .
4J C 0 C
td CO Q* • to
> o o o
•o
0)
•a
C
•!-> O
o 01
Si t~
•a
01
•o
c
CO
X
4-> O
O 01
c
o
lu
to
>'oi 01
CO 01 t.
£ -O CO
Cw
'fH
01 oi
CO '
•o to
CO L
3 -0
01 O. 01
•Q 5.
rH 3
XrH 0)
CO 01 01
E 3 CO
L
S
M a o co < E
tM r-
O 1
o c:
c\i 32 x:
O O.U
01
Dissolved Case
Hell-Purging
Parameters
i
01 O.
•a 01 o
ffl tl rt o.
xi 01 t, cd 01
4J 0) -P
>, CO 4J 01 tfl
to 3 CO t. "H
S CP E CO L
01
01 td
Xt 3
o-
>. 01
CO T3
s
tO
. Cl
•«H
i-H
Pi
•a;
M EI I-H > M .t-1 >H
•r
+C0
H
+ " O
itf CO
M-f ••
+ + 1
CO CO rH
as S o
01
g
CO 01
*• "e
z -e
n
0}
131
-------
Analyte Selection and Sampling Recommendations
A list of the corresponding regulated parameters for a detective
monitoring effort is provided in Table 3.1. The listing includes param-
eters of the following types: well purging, contamination indicator,
water quality and those that establish drinking water suitability. The
well-purging parameters provide both a measure of the efficiency of the
well-evacuation procedures prior to the collection of samples and valu-
able data (e.g. Eh, pH, Q-1, T) for the evaluation or interpretation of
water chemistry results. The contamination indicator parameters (e.g.
pH, 8~1, TOC, TOX) may indicate whether or not gross changes in ground-
water solution composition have occurred due to a contaminant release.
The sensitivity of these indicator parameters is somewhat limited with
the exception of TOX which can be determined reliably at sub-ppm
(yg'lT1) levels.
Water quality parameters provide useful information for description
of the ground-water system, particularly when the regulated constituents
(e.g. Cl~, Fe, Mn, Na+, SGij= and phenols) are supplemented with the
major cations and ions which usually comprise the bulk of the dissolved
solids in natural water samples. The water quality parameters may be
used as a basis for comparison in the event that the monitoring program
is triggered into an assessment phase. More importantly, the character-
ization of the inorganic chemical composition of ground water enables
both the quantitative interpretation of the consistency of the analyti-
cal results and the potential to calculate the chemical speciation of
specific dissolved chemical constituents. It is the speciation of
chemical constituents which enables the prediction of their reactivity,
132
-------
Table 3.1. Recommended Analytical Parameters for Detective Monitoring
Type of parameter
Well-purging
Contamination
indicators
Water quality
Drinking water
suitability**
Type of determination
Lab. (L), Field (F)
F
,L
L
L
Analytes
Required by regulation
pH, conductivity (ST1)
pH, ST1
Total organic carbon (TOC)
Total organic halogen (TOX)
Cl~, Fe, Mn, Na+, SOjj"
Phenols
As, Ba.Cd, Cr, F-, Pb, Hg,
N03~, Se, Ag
Endrln, lindane, methoxychlor, toxaphene
2,4-D, 2,4,5-TP (Silvex)
Radium, gross alpha/beta
coliform bacteria
Suggested for
completeness
Temperature (T)
Redox potential (Eh)
Alkalinity (F) or
acidity (F)
Ca++, Mg++, K+, N
POip, silicate,
ammonium
* All parameters are required to be determined quarterly for the first year of network operation.
** These parameters are excluded from the annual reporting requirements of RCRA after the first year.
133
-------
solubility and mobility under the actual conditions at the site. [It
should be noted that mass and charge balance consistency of the ana-
lytical results is a pre-condition for the reliable application of
equilibrium speciation models.] The drinking water suitability param-
eters are often required to be determined in the first year of network
operation on a quarterly basis by monitoring regulations. They may be
excluded from annual reporting requirement in succeeding years. One
year of quarterly data for these parameters may not be sufficient, since
potable water wells may often be used as background (upgradient) compo-
nents of a monitoring network if the original upgradient wells of the
network prove to be contaminated. These requirements may vary somewhat
based on current monitoring regulations. Field blanks, standards and
spiked samples should be done at the same degree of replication for all
parameters on each sampling date.
In summary, recommended parameters for detective monitoring
programs provide a minimum capability to detect contamination and to
serve as a basis for comparison and planning, should the program enter
the assessment phase. Depending on the hydrologic conditions at the
site, -higher sampling frequency (e.g. monthly) will provide a better set
of baseline data for future trend analysis or upgradient-downgradient
comparisons.
Many ground-water monitoring programs will entail the determination
of the sensitive chemical parameters noted in Table 3.1. These param-
eters demand the careful selection of both field and laboratory sample
handling (e.g. pumping, transfer, collection and storage) and analytical
procedures. For example, levels of pH, Eh, TOG, TOX, alkalinity,
134
-------
ammonium, Fe and other trace metals are prone to serious bias (i.e. loss
or inaccuracy) and imprecision (i.e. inconsistent duplicates, high ana-
lytical variance) if volatilization, aeration or degassing occurs during
sample handling or analysis. The severity of these problems will be a
function of solution composition, field conditions and the complexity of
the actual procedures employed. It should be recognized that the
simplest procedures which minimize sample handling and exposure to the
atmosphere or agitation will provide the most reliable results. There-
fore, the use of a sampling mechanism which provides flow sufficient for
well purging and a steady stream of ground-water for the in-line deter-
mination of well-purging parameters (and in-line filtration)is pre-
ferred. This type of mechanism will enable the controlled transfer and
collection of discrete samples for both field and laboratory determina-
tions of specific chemical parameters. Where ground-water availability
is a problem, discrete samples must be collected with every effort to
preserve sample integrity. A schematic diagram of recommended sample
collection, and handling methods for detection monitoring programs, is
shown in Figure 3.3. Specifics on sample handling and preservation are
provided in Table 3.2. These recommendations 'have been based on the
available information from the literature.
ASSESSMENT MONITORING
The information needs of assessment monitoring efforts ai-e more
detailed than those involved in detection monitoring. In detection
monitoring, the indication of contamination and the establishment of a
basis for ground-water quality comparisons are the principal goals. In
the assessment phase, the nature, extent and dynamics of a contaminated
135
-------
Parameters
(type)
Well-purging
(pH, Eh, T, 0-')
Contaalnatlon
Indicators
(pH, D 1)
(TOO, TOX)
Mechanism
Pump
(T.S.P.O)
Flow rates:
0.1-1.0 L/mln
Crab
(T,S,G,P,0)
Pump
(T.S.P.O)
Flow rates:
0.1-1.0 L/mln
Crab
(T,S,G,P,0)
Pump
(T,S preferred;
0,P only where
supporting data
exists)
Grab
(T,S,G preferred;
0,P only where
supporting data
exists)
Hydrogeologic Cc
>100 mL/min yield
Flowing samples
Positive displacement
bladder pump
(air, N2)
Positive displacement
bladder pump
(air, H2)
(Mechanisms as above
operated at f!6w
rates not to exceed
100 raL/min)
W mL vials filled
gently from bottom
up and allowed to
overflow + Teflon
capped H/O headspace
iditions (yield capability)
<100 mL/min yield
Discrete samples
Dual check valve bailers
"thief" samplers
Dual check valve bailers
"thief" samplers
(Volatile fractions of TOC
and TOX may be lost
depending on conditions
and operator skill)
'to mL vials filled from bottom up
and allowed to overflow or gently
poured down the aide of the vial,
Teflon capped w/o headspace
Materials In order of preference include: Teflon (T); stainless steel (s); PVC, polypropylene
polyethylene (P)j boroslllcate glass (G)j other materials: silicone, polycarbonate, mild steel
OtO. (0)
(continued on next page)
Figure 3.3 Recommended sample collection methods for
detective monitoring programs
136
.
-------
Parameters
( type)
Water Quality
Dissolved Gases
(02. CHi,, C02)
Alky/Aedy
(Fe, Mn, P0j|=, Cl",
Na+, 3014-, Ca++,
Hg++, K+, N03-.
Silicate)
(Ammonium, Phenols)
Mechanism
(material)*
Pump
(T,S,P,0)
Grab
(T.S.G.P.O)
Pump
(T,S,P,0)
Grab
(T,S,G,P.O)
Pump
(T,S preferred;
0,P only where
supporting data
exists)
Grab
(T,S,G preferred;
0,P only where
supporting data
exists)
Hydrogeologio Conditions (yield capability)
>100 mL/min yield
Flowing samples
(Mechanisms as above
operated at flow
rates not to exceed
100 mL/min)
Glass containers
filled gently from
bottom up and allowed
to overflow -» Teflon
capped w/o headspace
Positive displacement
bladder pump
(air, N2)
(Mechanisms as above
operated at flow
rates not to exceed
1000 mL/min)
Glass containers
filled from bottom
up
<100 mL/min yield
Discrete samples
(Not recommended)
Fe values sensitive to most
grab mechanisms
Large volumes required may have
to be sequentially collected and
filtered
(Volatile species may be lost
depending on 'conditions)
Glass containers filled from
bottom up
* Materials in order of preference Include: Teflon (T); stainless steel (S); PVC, polypropylene,
polyethylene (P); borosilicate glass (G); other materials: silicone, polycarbonate, mild steel,
(concluded on next page)
Figure 3.3. (continued)
137
-------
Parameters
It.'Ps)
Drinking Hater
Suitability
(A3, Ba, Cd, Cr, Pb,
Hg, Se, Ag, N03-,
F-)
(Reaalning
Piraneters)
Mechanism
Pump
(T.S.P.O)
Grab
(T.S.G.P.O)
Pump
(T.S.P.O)
Grab
(T,S,G,P,0)
(both with
precautions if
radlologio hazards
exist)
Hydrogeologio Co
>100 mL/min yield
Flowing samples
Positive displacement
bladder pump
(air, ND)
Positive displacement
bladder pump
(air, N2)
Flow rates should
not exceed
1,000 mL/min
id It ions (yield capability)
<100 mL/min y,.eld
Discrete samples
Dual check valve bailers
"thief" samplers
(Volatile compounds may be
lost depending on conditions)
Materials In order of preference include: Teflon (T); stainless steel (S); PVC, polypropylene
polyothylena (P), borosilicate glass (a), other materials: silicone, polycarbonate,^? steel!
Figure 3.3. (concluded)
138
-------
Table 3.2. Recommended Sample Handling and Preservation Procedures
for a Detective Monitoring Program*
Parameters
(type)
Well purging
pH (grab)
IT1 (grab)
T (grab)
Eh (grab)
Contamination
indicators.
pH, a"1 (grab)
TOG'
TOX
Water quality
Dissolved gases
(02, CHi), C02)
Alkalinity/
Acidity
(Fe, Mn, Na+,
K+, Ca++,
Mg++)
(P04=, Cl",
Silicate)
N03"
S0i4 =
NHi) +
Phenols
Drinking water
suitability
As,Ba,Cd,Cr,
Pb,Hg,Se,Ag
F~
Volume
required
1 sample**
50
100
1000
1000
As above
to
100
10 mL minimum
100
Filtered
under
pressure
with
appropriate
media
All filtered
1000 mL
§ 50
100
50
1)00
500
Same as above
for water
quality
cations
(Fe,Mn,etc.)
Same as
chloride
above
Container
(material)
T.S.P.O
T.S.P.B
T,S,P,G
T.S.P.G
As above
G,T
G,T
G,S
T.G.P
T,P
(T,P,G
glass only)
T.P.G
T,P,G
T,P,G
T,G
Same as
above
Same as
above
Preservation
method
None-Field Det.
None-Field Det.
None-Field Det.
None-Field Det.
As above
Dark, 1°C
Dark, 1°C
Dark, l»8C
1°C/None
Field acidified
to pH <2 with
HN03
1°C
1°C
1°C
1°C/H2SOn to
pH <2
1°C/H3POi) to
pH <1
Same as above
Same as above
Maximum holding period
<1 hr.***
<1 hr.***
None
None
As above
21 hrs.
5 days
<21 hrs.
<6 hrs.***/<2l hrs.
6 months^
21 hrs./7 days;
7 days
21 hrs.
7 days
' 21 hrs./7 days
21 hrs.
6 months
7 days
Remaining
organic
parameters
As for TOX/TOC, except where analytical method
calls for acidification of sample
21 hrs.
* Modified after Scalf et al. (3)
** It is assumed that at each site, for each sampling date, replicates, a "field blank
and standards must be taken at equal volume to those of the samples.
*** Temperature correction must be made for reliable reporting. Variations greater than
±10? may result from longer holding period.
A In the event that HN03 cannot be used because of shipping restriction, the sample
should be refrigerated to 1°C, shipped immediately, and acidified on receipt at
the laboratory. Container should be rinsed with 1:1 HNOg and included with
sample.
139
-------
ground-water situation must be characterized sufficiently to plan
further investigative or remedial action activities* The level of
detail required in assessment efforts may be an order of magnitude more
complex than those in the detective phase. Therefore, the reliability
of the data in space and time must increase proportionately. Incomplete
characterization of a ground-water sample's solution composition could
lead to the incorrect assessment of the mobility or reactivity of poten-
tial contaminants. The three-dimensional extent of a contaminant pulse
or plume might be lost if the bias introduced into the determination of
the principal contaminants is high relative to background concentra-
tions. Remedial action or mitigative action decisions should be based
on a high quality data set which meets the information needs of the
program. Clearly the experience that operators gain during the detec-
tion phase of monitoring will prepare them for reliable assessment
activities.
The well-purging and contamination indicator parameters are gener-
ally less sensitive to gross sampling and analytical errors than chemi-
cal constituents which may be specific components of a waste from a
landfill, impoundment, waste-pile, spill or storage area. Predictions
of the major contaminants involved and the subset of stable, mobile
constituents that may be expected to be found downgradient must be made.
For exampie,assume that a well-executed detection monitoring effort
at a solvent waste transport station disclosed that TOX values down-
gradient Sre significantly different from those collected during the
past three quarters at upgradient wells. The mean upgradient value
differs from that downgradient by 100 ppb which is of the order of five
140
-------
times the mean precision of the TOX determinations at these levels. The
TOG data, on the other hand, show no statistically significant differ-
ence' between the upgradient and downgradient wells. Since the precision
of the TOG values are ±0.1 rng-L"^ at best, it is quite possible that the
present contamination is the result of halogenated'solvent releases. In
this case it may be that hydrocarbon solvents or petroleum derived com-
pounds are the likely constituents of interest in the assessment phase.
Reliable sampling of the TOX in the ground-water at the site may
permit the scope of the initial assessment to be limited to halogenated
compounds. Additional data would be helpful if the analytical results
clearly reported both volatile and nonvolatile TOG and TOX. If, in the
example above, the observed TOX increase was represented in a propor-
tional increase only in the volatile TOX, the purgable organic compounds
should be investigated in the initial assessment activity.
If the detective monitoring results disclose only secondary,
nonvolatile contaminants (because the volatile fractions of TOG or TOX
were lost during sample collection, handling or analysis), the conse-
quences of relying on a poorly designed sampling protocol could be far
more serious. Precision and bias for determinations of the detective
monitoring parameters can be controlled in the ±10 to 50% range.
However, order of magnitude levels of variance or loss may enter into
sampling and analytical results for trace constituents at the ppb
(pig-LT^) level. Poor precision and accuracy directly reduce the power
of statistical tests for cxomparison of background and potentially
affected downgradient conditions.
141
-------
As the information needs of a monitoring program become more
detailed it is essential to establish control over errors'. Sample
collection and handling problems for TOG and TOX which do not introduce
additional bias or imprecision above those of the analytical methods may
be expected to perform adequately for specific inorganic or organic
chemical constituents of a contaminant release. This will be true if
the chemical constituents of the product/waste release are known and
their mobility or reactivity in the subsurface can be reasonably pre-
dicted. The actual selection of "facility-specific" constituents also
may be very difficult to make if ground-water quality has not been well
characterized in the detection monitoring phase.
Given the wide range of potential contaminants (e.g. potentially
thousands of waste components in RCRA, Appendix VIII, etc. listings) and
those which may be sensitive to sample collection or handling errors, it
is difficult to make a priori evaluations of the adequacy of monitoring
procedures or protocols. However, it is clear that proven sampling and
sample handling procedures which control bias and precision at compa-
rable levels of analytical method performance are most reliable. In
this respect, Fe, pH, TOX and TOG results are parameters which may be
used to gauge the utility of sampling protocols used in detection moni-
toring for application in contamination assessment work. One may gener-
alize reliable sample collection and handling protocols on this basis.
Dissolved iron may be accepted as being representative of inorganic
metallic species which are prone to oxidation and the formation of solid
oxide or oxyhydroxide products. The oxide products have very active
surfaces for the sorption of other metallic ions or organic compounds.
142
-------
If water samples are not carefully collected (i.e. to exclude 0£ or gas
exchange), handled (i.e. filtered under N2 or pump presure prior to
acidification), the reduced iron in many samples would oxidize prior to
preservation and this reaction, as well as the inevitable sorptive
interactions, could seriously bias the analytically determined composi-
tion of' the ground water (87). By analogy, the target chemical con-
stituents in an assessment program for metallic contamination (e.g. Cu,
Cr, Ni from an acidic alloy treating process waste) should be sampled
and handled reliably using the same procedures which permit reliable
dissolved iron samples to be taken. It should be noted that although
many RCRA Appendix VIII parameters are metallic and may require only
metal determinations in the lab, the actual elemented speciation will
impact the reliability of sampling procedures. This may be illustrated
by inspection of Table 3.3. Analysis procedures should be streamlined
to facilitate screening of water samples since the speciation of the
metal may impact on sample preparations and all the steps which precede
them (i.e. sample collection, transfer, filtration, preservation and
storage). It is difficult to specify the optimum sampling procedures
for water samples potentially contaminated with a variety of uneharac-
terized waste mixtures. However, a sampling protocol which is proven
reliable for difficult or sensitive chemical constituents should perform
adequately for most other parameters. Figure 3.2 contains a matrix of
chemical constituents .and appropriate sampling mechanisms. An increase
in the degree of sampling difficulty or sensitivity to bias of a con-
stituent requires that a more robust, fool-proof sampling mechanism be
used. If alternative sampling methods are utilized which are not well
143
-------
Quantity
1
2
1
2
3
2
2
Table 3.3. Equipment for Field Sampling
Item
Compressed N2 cylinder (301 ft3) for bladder pump
sampling oxidation sensitive constituents if needed
Scuba tanks (compressed air) (80 ft3 + 50 ft3) for
bladder pump
Alkalinity box- (battery operated pH meter with
temperature compensation, electrode, battery operated
magnetic stirrer, buret, titrant, beakers)
Flow through cell in box with 3-way valve system to route
pump output to cell (e.g. pH, 2 redox, temp, electrodes +
conductivity cell) or to sample/waste (ref. Figure 2.18)
Meter box (3- pH meters (as above) + 1 battery operated
temperature compensated conductivity bridge) (ref. Figure
2.18)
Regulators for gas cylinder + scuba tanks
Buckets (15 L) and graduated cylinder (5 L) to measure
purge volume and sample waste
Dissolved oxygen field kit (Modified Winkler Method (46)
200 mL titration volume)
5 gallon (LDPE) water jugs for deionized water
Sampling pumps (primary plus a backup and an extra
bladder assembly) Teflon/Teflon bladder and Stainless
Steel/Teflon bladder
Pump tubing sets (Teflon) (1 air, 1 water, @ 50' + tubing
holder, primary plus backup). Tubing diameter should be
no less than 1/4" o.d. and the larger diameter sizes will
minimize tubing material effects if they are
anticipated)
Pump control box with tubing
Gas manifold (to operate multiple pumps from same
compressed gas supply)
Steel measuring tape
Grass whip
Shovel
(concluded on next page)
144
-------
Table 3-3.
Quantity
1
1
3
4
3
(concluded)
Item
Miscellaneous box with (6 boxes Kimwipes, 3 boxes
disposable gloves, aluminum foil, duct tape)
Miscellaneous box with (pH buffers, deionized wash
bottle, Erlenmeyer flasks, beakers, graduate cylinders,
pasteur pipettes, bulbs, cone. HNOg acid, cone. HC1 acid,
filter membranes, filter holders)
Shock cords
Coolers (insulated, 64 qt., 54 qfr, 44 qt - one each)
Toolboxes
* Sample bottles for samples, spiked samples and e;xtras
** Prepared bottles for field blanks and standards
solutions
*** Sampling log, field notebooks, chain of custody forms
with spiking
145
-------
referenced, supporting accuracy and precision data should be provided
for the specific constituents of interest. Regardless of the sampling
mechanism used, the elements of the generalized sampling protocol should
be documented completely.
Field Sampling Procedures
This section of the guide is presented as an example of "how-to"
collect samples as drawn from the authors' experiences. Refinement and
modification will be necessary for application to specific sampling and
analytical needs. In large measure, the degree of preparedness and
skill which these individuals take into the field will determine the
actual number of samples which can be collected. A well prepared team
of three individuals can usually sample between 4-6 monitoring wells
(0-75 ft. deep) in a full 8-hour day, exclusive of travel time. Given
the range of field or hydrogeologic conditions, network complexities and
the analytical detail which ground-water monitoring investigations
demand, no single example can provide all of the elements needed in the
sampling protocol. The following discussion should provide the basis
for the application of effective sampling procedures for either detec-
tion or assessment monitoring investigations.
The following steps in a sampling protocol are covered in detail
below:
Sampling Equipment Setup, Well Inspection and
Water Level Measurement
Verification of Well Purging Requirements
Sample Collection/Filtration/Field Blanks and Standards
146
-------
Field Determinations
Sample Storage/Transport
The importance of careful integration of the efforts of sampling
staff at each point should not be underestimated. Mistakes, lost data
or biased results may exact a heavy price if sampling efforts are not
well planned. The same care taken in the laboratory to prevent mishaps
or contamination should be followed in the field. It should be obvious
that smoking or eating in the vicinity of the well head, pump output or
field analytical setups is strongly discouraged.
Sampling Equipment Setup, Well Inspection and Water Level Measurement
It is a good practice to have a detailed list of all sampling mate-
rials and supplies. The list should be reviewed before the sampling
staff leaves for the field site. This somewhat tedious procedure will
cut down on the frustration or anxiety which may arise later because of
missing equipment, reagents or bottles. An example of a sampling equip-
ment list is shown in Table 3.4 which includes the basic gear needed to
conduct routine sampling and field activities. The list is reasonably
complete for a protocol based on the use of a positive displacement
bladder pump which is sufficient for the well-purging and sample collec-
tion requirements of many monitoring situations.
On arrival at the well-head, the condition of the surface seal and
well protector should be examined to see if any evidence of frost-
heaving, cracks or vandalism are observed, they should be recorded in
the field notebook. The area around the well may have to be cleared of
weeds or other materials prior to beginning the sampling activity. A
drop cloth should then be placed on the ground around the well head,
147
-------
Table 3.4. Metallic Species in RCRA Appendix VIII
Which Require Only Metal Determinations
* Antimony NOS
Arsenic acid
* Arsenic and compounds, NOS
Arsenic pentoxide
Arsenic trioxide
* Barium and compounds, NOS
Barium cyanide
* Benzenearsonic acid
* Beryllium and compounds, NOS
* Cadmium and compounds, NOS
Calcium chromate
* Chromium and compounds, NOS
Copper cyanide
* Dichlorophenylarsine
* Diethylarsine
* Hydroxydimethylarsine oxide
Lead acetate
* Lead and compounds, NOS
Lead phosphate
Lead subacetate
* Tetraethyl lead
* Mercury and compounds, NOS
* Mercury fulminate
* Nickel and compounds, NOS
NOS: Not otherwise specified; signifies those members of the general
class not specifically listed by name in Appendix VIII.
* Metallic species which may exhibit markedly different properties
(e.g. solubility, volatility, reactivity) from inorganic ions or
complexes in ground water
*.Nickel carbonyl
Nickel cyanide
Osmium tetroxide
* Phenylmercury acetate
Potassium silver cyanide
* Selenium and compounds, NOS
Selenious acid
Selenium sulfide
* Selenourea
* Silver and compounds, NOS
Silver cyanide
Strontium sulfide
Thallic oxide
Thallium acetate
* Thallium and compounds, NOS
Thallium carbonate
Thallium chloride
Thallium nitrate
Thallium selenite
Thallium sulfate
Vanadic acid, ammonium salt
Vanadium pentoxide
Zinc cyanide
148
-------
particularly if the land surface is disturbed or potentially contami-
nated. This precaution will save time and the work of cleaning equip-
ment or tubing should they fall on the ground during preparation or
operation. The well protector should then be unlocked and the cap
removed from the top of the well. The previous record of water levels
for the well should be consulted prior to chalking the steel tape and
making three successive measurements of the static water level. The
readings should be recorded to the nearest ±0.01 ft. If the well has a
history of contamination, the water level measurements should be made
with surgical gloves on and the tape should be rinsed with distilled
water and wiped dry with lint-free towels as it is wound on the reel.
While the water level is being measured, the other sampling personnel
should prepare to set up the pumping and flow-through measurement equip-
i
ment and the instrumentation for analytical field determinations.
Blanks and standards should be titrated for alkalinity and dissolved
!
oxygen determinations at this time. Also, the pH meters, Eh electrode
combinations and the conductivity bridge should be calibrated (78). The
assembly of the Teflon and stainless steel bladder pump and the tubing
bundles should be performed as well. Gloves should be worn at all times
during pump assembly. These activities should take approximately 35-45
minutes and may be completed at a location central to all the wells
which will be sampled during the day. At this point, the sample bottles
should be checked for proper labelling. Then the field and sampling
logs should be readied for the next steps. It is important to record
the stagnant water volume in the well from the water level reading and
149
-------
compare it to that calculated for the well from the evaluation of
pimping requirements.
Verification of the Well Purging Requirement
. Well purging requirements should be calculated from the hydraulic
performance of the well and verified each day by measurement of the
well-purging parameters. Let us presume that the example well has been
properly evaluated as to its hydraulic performance by the methods
described in the examples in Section 2. In this case, the calculated
purging requiement is approximately 80 L (-4 well volumes) which should
be purged prior to the collection of representative samples. Since the
well was developed at a flow rate of approximately 6 L/min, a conser-
vative pumping rate of 3 L/min has been chosen for purging the well
pumping rate of 1 L/min has been chosen. The pump is lowered to the
point where the pump intake is at the top of the screened interval. It
is useful to use a "keeper" which consists of a wooden or plastic
rectangle with holes drilled in it to allow the gas and water tubes to
slide through and be held in place with a knotted cord or wire tie. At
this time the pump should be started and adjusted to produce a steady
output- through the flow-through cell and into a collection bucket or
drum. At intervals equal to (10$ of the calculated purging requirement
(-8 L), the readings of Eh, pH, T, and n~1 are then recorded and the
cumulative volume pumped (including that in the cell) should be measured
and recorded. When the calculated purge volume is approached the read-
ings should be made at more frequent volume intervals and the pump may
be slowed to -1,000 mL/min. When the readings of the well purging
parameters have stabilized to within ±10$ over two successive volume
150
-------
increments (i.e., no less than 20% of the required purge volume; -16 L),
the pump output may be considered equilibrated and sampling may begin.
The data in Table 3.5 show the gradual stabilization of the pH and JT1
values at -49 L which was verified by pumping through 16 more liters.
In this example, about 80$ of the calculated well purging requirement
was pumped prior to equilibration..
Sample Collection/Filtration
Samples should be taken in a prearranged priority so that all
sample handling and preservation takes place as rapidly as possible.
Although no significant error has been reported for gas sensitive con-
stituents pumped with a positive displacement bladder device when air is
used as the drive gas, it may be prudent to switch the drive gas from
air to N2 at this point. Samples for dissolved gases are then taken,
in-line ahead of the flow-through electrode cell at a flow rate of
-100 mL/min.
The samples for dissolved gases, volatile organic constituents, TOG
and TOX are taken by carefully slowing the delivery rate to 100 mL/min
or less and directing the flow to the bottom of the sample vessel (e.g.
or by flowing into a syringe of appropriate volume)' and allowing the
vessel .to, overflow at least 1.5 volumes. The samples should be rapidly
capped, excluding any heads pace and preserved or put in the sample
cooler as soon as possible. At this point, the time (and volume) of
initial sample collection should be recorded. An effort should be made
to keep track of the cumulative volume pumped during sampling and all
subsequent steps. Samples for 'extract-able organic compounds and total
metals can then be collected. In filling the large volume bottles the
151
-------
Table 3.5.
Quantity
pumped
(liters)
8
16
24
32
40
49
57
65
Sample Purging Parameter Readings
Conductivity
pH
8.01
7.67
7.54
7.19
7.22
7.16
7.17
7.16
580
625
623
622
619
620
621
620
152
-------
flow rate can be increased but should not exceed the pumping rate during
purging.
At this point, the pump discharge is connected to an on-line filter
apparatus and the samples for alkalinity, dissolved metals and other
inorganic constituents can be collected in priority order. When the
filtered samples have been collected, the time and cumulative volume
pumped are recorded. One member of the sampling team should oversee the
operation, insure proper preservation of the samples, and make the
entries into the field and sampling logs of the time of sample collec-
tion, double-checking the labels on the storage vessels. Another member
of the team should begin titrating alkalinity samples, at least in
duplicate. The titrations should not be delayed more than two hours
from the initial sampling time. The other member of the sampling team
should be in charge of sample collection, time and volume measurements
to insure that the samples and replicates are properly taken. Then the
full flow is redirected through the electrode cell. Values of the well
purging parameters should be recorded after the cell has been flushed
at least once, if volume permits. These values should later be compared
to those taken just prior to the collection of the initial samples to
check on the stability of the water during the time of sampling.
Now the samples and field blanks should be properly preserved and
stored. At least one replicate of each sample (excluding dissolved
oxygen and alkalinity) should be spiked with an appropriate stock solu-
tion to provide a blind control standard for sensitive analytical deter-
minations. To insure good quality control, these blind samples are
153
-------
labelled as an extra well and placed in the normal sample handling
scheme enroute to the laboratory.
The gas supply to, the pump is then turned off, and the pump with
the tubing bundle can be retrieved. Before proceeding to the next well,
the pump should be placed in a graduated cylinder of rinse or cleaning
solution. The pump should be operated to detect any leakage and to
clean the pump and the interior surfaces of the sampling train. Any
waste water that may be contaminated with hazardous constituents should
be managed in a responsible manner. Under no circumstances should it
be returned to the well.
Field Determinations
The determination of alkalinity dissolved oxygenand other field
constituents (e.g. pH, Eh, T and a~1) should be completed at this point.
Dissolved oxygen samples should be kept out of light, preserved and cold
until the precipitate has formed and settled to the bottom of the
bottle. After an hour or so, they should be shaken again and allowed to
settle. So long as they are kept in the dark, they can be held for 4-8
hours prior to acidification and titration.
All other field parameters can be determined after method cali-
bration has been performed, as conditions permit. At this time, the
field and sampling logs should be checked for completeness and the
initial chain of custody documentation has been completed.-
Sample Storage and Transport
The procedures described in Section 2 should be followed explicitly
from this point until delivery to the laboratory. Any unique circum-
stances (e.g. extreme heat or cold, delays in sample handling, etc.)
154
-------
should be recorded in the field notebook. It is essential that the
laboratory receive all information which may affect analytical pro-
cessing. Notice of any extreme turbidity, reactivity with the preserva-
tion reagents, etc. should be provided _in writing to the laboratory
personnel.
These sampling procedures are sufficient to the needs of most
ground-water sampling programs. If unusual conditions exist, they
should be reported to the person in charge of the monitoring effort at
once. This will help prevent undue exposure of sampling staff or water
samples to conditions that may jeopardize health or the collection of
high quality data.
155
-------
SECTION 4
CONCLUSIONS
The development of reliable sampling protocols for ground-water
quality monitoring is a complex, programmatic process that must be
designed to meet the specific goals of the monitoring effort in
question. The long-term goals and information needs of the monitoring
program must first be thoroughly understood. Once these considerations
have been identified, the many factors that can effect the results of
chemical analyses from the monitoring program can be addressed.
In formulating the sampling protocol, the emphasis should be to
collect hydrologic and chemical data that accurately,represent in situ
hydrologic and chemical conditions. With good quality assurance guide-
lines and quality control measures, the protocol should provide the
needed data for successful management of the monitoring program at a
high level of confidence. Straightforward techniques that minimize the
disturbance of the subsurface and the samples at each step in the
sampling effort should be given priority.
The planning of a monitoring program should be a staged effort
designed to collect information during the exploratory or initial stages
of the program. Information gained throughout the development of the
program should be used for refining the preliminary program design.
During all phases of protocol development, the long-term costs of
producing the required hydrologic and chemical data should be kept in
mind. These long-term costs are several orders of magnitude larger than
the combined costs of planning, well construction, purchase of sampling
and field equipment, and data collection start-up. it also should be
156
-------
remembered that high quality data cannot be obtained from a poorly con-
ceived and implemented monitoring program, regardless of the added care
and costs of sophisticated sampling and analytical procedures.
Finally, the ultimate costs of defending poor quality data in the
legal arena or in compliance with regulatory requirements should not be
overlooked. The damage to the credibility of the program can be sub-
stantial.
Due to the lack of documented standard techniques for developing
monitoring programs, constructing monitoring wells, and collecting
samples, quality control measures must be tailored for each individual
site to be monitored. They should be designed to insure that distur-
bances to both the hydrogeologic system and the sample are minimized.
The care exercised in well placement and construction, and sample col-
lection and analysis can pay real dividends in the control of systematic
errors. Repeated sampling and field measurements will further define
the magnitude of random errors induced by field conditions and human
error. Still the burden of assuring the success of a program relies on
careful documentation and the performance of quality assurance audit
procedures.
The hydrogeologic conditions at each site must be evaluated in
terms of the potential impacts the setting will have on the design and
effectiveness of the developed program. Documentation of the hydrology
of the site is essential at the planning stage, as well as during the
operational life of the program. Too little attention has been given to
fully understanding the environment that is the source of water col-
lected from monitoring wells. Only after the source'of water is known
157
-------
(delineation of the vertical and horizontal components of ground-water
movement) can the effectiveness of the program be assured.
The placement and construction of monitoring wells can be among the
most difficult tasks involved in developing an effective monitoring
program. The positioning of a monitoring point in a contaminant flow
path must be determined on the basis of hydrologic data to insure that
the well is capable of monitoring the contaminant plume or release. The
monitoring wells also should be constructed using drilling techniques
that avoid the disturbance of subsurface conditions due to the intro-
duction of fluids or muds. Monitoring wells should be sized both to
provide depth discrete hydrologic and chemical data and to maximize the
usefulness of the collected data. The materials selected for monitoring
well construction should be durable for the intended installation and
minimize interference with the samples to be collected. The wells also
should be properly developed to maximize their hydraulic efficiency and
minimize the need to filter water samples.
Sampling mechanisms for the collection of ground-water samples are
among the most error prone elements of monitoring programs. Documenta-
tion of the field performance for most devices and materials is lacking.
Many of the sampling designs may be expected to provide adequate perfor-
mance for conservative chemical constituents which are not affected by
aeration, gas-exchange and degassing. Testimonials of sampling perfor-
mance based on the recovery of conservative, unreactive chemical consti-
tuents are not reliable for planning effective monitoring efforts. It
should be recognized that the purchase of a suitable sampler for most
158
-------
ground-water investigations is usually a very small portion of the over-
all program cost. It is further obvious that the choice of the right
sampler made of appropriate materials will determine the ultimate use- .
fulness of the chemical data. The recommended approach is to make the
choice of both samplers and materials on the basis of the most sensitive
chemical constituents of interest. Typically, reliable samples for
dissolved gases, ferrous iron and volatile organic compounds are the
most difficult to collect and handle.
The information needs of a ground-water monitoring program are
t
determined by the stated goals of the program. They should be deter-
mined by the program manager, and field and laboratory personnel during
the planning phase of the project. The long-term goals or anticipated
needs of the program also should be addressed at the outset of the
program to insure data consistency and quality throughout the life of
the program.
The definition of a representative ground-water sample will vary
from site to site and perhaps from sampling point to sampling point,
depending on the situation under investigation. Performance criteria
for the achievement of representative sampling should include the
accuracy, precision, sensitivity and completeness necessary to provide a
minimum level of confidence in the data. The criteria should be based
on both knowledge of the system to be measured and the experience of the
project planning staff. Close attention must be paid to the preliminary
investigation, well placement and construction, hydrologic data,
sampling frequency, and mobility and persistence of likely chemical
contaminants. Natural or man-induced variability in the hydrogeology
159
-------
and geocheraical characteristics of the site can only be distinguished
from each other by the interpretation of high quality sampling results.
It is hoped that by the careful implementation of the recommendations
for sampling in this guide, that a level of confidence in ground-water
data can be established. Our understanding of subsurface processes
should improve in great measure as reliable investigations proceed.
160
-------
SECTION 5
RECOMMENDATIONS
Well drilling/completion, purging, sampling and analysis steps all
contribute to error in ground-water monitoring results. The procedures
must be, better understood as they effect particular classes of contami-
nants. This information is necessary in order to facilitate the
development of efficient protocols and QA/QC programs. Specific problem
areas which require further research include:
Drilling mud composition and effects on subsurface geochemistry.
Grouting materials and procedures which effectively seal screened
intervals from leakage or cross-contamination, especially adverse
effects of contaminants on grout set-up and integrity.
Well development procedures which are effective in reducing par-
ti culate matter in water samples.
Efficient methods for establishing monitoring points and sampling
free-product or non-aqueous contaminant phases in the subsurface.
The effects of inadequate well-purging protocols prior to sampling
for chemical analysis, emphasizing long term and short term well-
casing material effects on sample integrity.
Once the most critical sources of error involved in specific con-
taminant monitoring situations have been identified, more basic studies
of subsurface hydrogeology and sample handling must be done to minimize
sources of systematic error and imprecision. Research is needed on:
Filtration effects on ground-water samples used for transport or
contaminant flux investigations. The significance of total-
recoverable (i.e., non-filtered)'water sample analytical results
for assessment work and colloidal transport effects require further
investigation.
161
-------
Methods for the interpretation of observed contaminant distribu-
tions in creviced or fractured geological materials and the unsatu-
rated zone need improvement.
Improvements in geophysical monitoring methods and their relation
to more traditional contaminant detection methods are needed.
One area that needs particular attention is the training of field
and laboratory personnel in reliable monitoring techniques. The scien-
tific literature on ground-water monitoring is developing rapidly. All
monitoring personnel should make an effort to acquaint themselves with
published materials and maintain a current understanding of advances in
the field.
162
-------
SECTION 5
REFERENCES
1. U.S. Geological Survey. 1977. National Handbook of Recommended
Methods for Water-Data Acquisition. USGS Office of Water Data
Coordination, Reston, Virginia.
2. Wood, W. W. 1976. Guidelines for Collection and Field Analysis of
Groundwater Samples for Selected Unstable Constituents. In: U.S.
Geological Survey Techniques for Water Resources Investigations,
Book 1, Chapter D-2.
3. Scalf, M. R., J. F. McNabb, W. J. Dunlap, R. L. Cosby, and
J. Fryberger. 1981. Manual of Ground-Water Quality Sampling
Procedures. National Water Well Association, Worthington, Ohio.
4. Brass, H. J., M. A. Feige, T. Halloran, J. W. Mellow, D. Munch, and
R. F. Thomas. 1977. The National Organic -Monitoring Survey:
Samplings and Analyses for Purgeable Organic Compounds. In;
Drinking Water Quality Enhancement through Source Protection (R. B.
Pojasek, ed.), Ann Arbor Science Publishers, Ann Arbor, Michigan.
5. Dunlap, W. J., J. F. McNabb, M. R. Scalf, and R. L. Cosby. 1977.
Sampling for Organic Chemicals and Microorganisms in the
Subsurface. Office of Research and Development, USEPA,
Robert S. Kerr Environmental Research Laboratory, Ada, Oklahoma.
6. Sisk, S. W. 1981. NEIC Manual for Groundwater/Subsurface Investi-
gations at Hazardous Waste Sites. USEPA Office of Enforcement,
National Enforcement Investigations Center, Denver, Colorado.
7. Fenn, D., E. Cocozza, J. Isbister, 0. Braids, B. Yare, and P. Roux.
1977. Procedures Manual for Ground Water Monitoring at Solid Waste
Disposal Facilities. EPA/530/SW611 , USEPA, Cincinnati, Ohio.
8. Tinlin, R. M., ed. 1976. Monitoring Groundwater Quality:
Illustrative Examples. EPA 600/4-76-036, USEPA, Environmental
Monitoring and Support Laboratory, Office of Research and Develop-
ment, Las Vegas, Nevada.
9. National Council of the Paper Industry for Air and Stream
Improvement. 1982. A -Guide to Groundwater Sampling. Technical
Bulletin 362, NCASI, 260 Madison Avenue, New York, New York.
10. Todd, D. K., R. M. Tinlin, K. D. Schmidt, and L. G. Everett. 1976.
Monitoring Ground-Water Quality: Monitoring Methodology.
EPA-600/4-76-026, USEPA, Las Vegas, Nevada.
11. Gibb, J. P., R. M. Schuller, and R. A. Griffin. 1981. Procedures
for the Collection of Representative Water Quality Data from
Monitoring Wells. Cooperative Groundwater Report 7, Illinois State
Water Survey and Illinois State Geological Survey, Champaign,
Illinois.
163
-------
12. Grisak, G. E., R. E. Jackson, and J. F. Pickens. 1978. Monitoring
Groundwater Quality: The Technical Difficulties. Water Resources
Bulletin, June 12-14, 1978, San Francisco, Calif., p. 210-232.
13. Gillham, R. W., M. J. L. Robin, J. F. Barker and J. A. Cherry.
1983. Ground-Water Monitoring and Sample Bias. Department of Earth
Sciences, University of Waterloo. Prepared for the American Petro-
leum Institute. API Pub. 4367, June 1983, 206 pp.
14. Brown, K. W. and S. L. Black. 1983. Quality Assurance and Quality
Control Data Validation Procedures Used for the Love Canal and
Dallas Lead Soil Monitoring Programs. Environmental Monitoring and
Assessment 3, p. 113-122.
15. Nacht, S. J. 1983. Monitoring Sampling Protocol Considerations.
Ground Water Monitoring Review, Summer, 1983, p. 23-29.
16. Keith, S. J., M. T. Frank, G. McCarty and G. Massmah. 1983.
Dealing With the Problem of Obtaining Accurate'Ground-Water Quality
Analytical Results. In; Proceedings of the 3rd National Symposium
on Aquifer Restoration and Ground Water Monitoring. May 25-27,
1983, Columbus, Ohio. D. M. Nielsen, ed., National Water Well
Association, Water Well Journal Publishing Company, Worthington,
Ohio, 1983, 461 pp.
17. Kirchmer, C. J. 1983. Quality Control . in" Water Analyses.
Environmental Science and Technology 17, 4, p.r174A-181A.
18. Kirchmer, C. J., M. C. Winter and B. A. Kelly. 1983, Factors
Affecting the Accuracy of Quantitative Analyses of Priority
Pollutants Using GC/MS. Environmental Science and Technology 17,
396-401.
19. Dressman, R. C. 1982. Elements of a Laboratory Quality Assurance
Program. Proc. of AWWA Water Quality Technology Conference,
Nashville, TN, December 5-8. American Water Works Association,
1982, p. 69-75.
20. Dux, J. P. ,1983. Quality Assurance in the Analytical Laboratory.
American Laboratory, July, 1983, p. 54-63.
21. Kingsley, B. A. 1982. Quality Assurance in a Control Laboratory.
Proc. of the AWWA Water Quality Technology Conference, Nashville,
TN, December 5-8. Journal American Water Works Association, 1982,
p. 69-75.
22. Kingsley, B. A., C. Gin, W. R. Peifer, D. F. Stivers, S. H. Allen,
H. J. Brass, E. M. Glick and M. J. Weisner. 1981. Cooperative
Quality Assurance Program for Monitoring Contract Laboratory Per-
formance. In: Advances in the Identification and Analysis of
Organic Pollutants in Water, Chapter 45, Vol. 2. L. H. Keith, ed.,
Ann Arbor Science, Ann Arbor, MI, 1981.
164
-------
23. Barcelona, M. J. 1983. Chemical Problems in Ground-Water
Monitoring Programs. in: Proceedings of the 3rd National
Symposium on Aquifer Restoration and Ground-Water Monitoring,
Columbus, OH, May 25-27, 1983, p. 263-271. D. M. Nielsen, ed.,
National Water ' Well Association, Water Well Journal Publishing
Company, Worthington, OH, 461 pp.
24. Taylor, J. K. 1983. Quality Assurance of Chemical Measurements.
Analytical Chemistry 53, 14, p. 1588A-1593A.
25. Taylor, J. K. 1981. .Validation of Analytical Methods. Analytical
Chemistry 55, 6, p. 600A-608A.
26. ACS. 1980. Guidelines for Data Acquisition and Data Quality
Evaluation in Environmental Chemistry. American Chemical Society
Committee on Environmental improvement, Analytical Chemistry 52,
2242-2249.
27. USEPA. 1982. Test Methods for Evaluating Solid Waste, SW-846, 2nd
edition. Office of Solid Waste and Emergency Response, Washington,
D.C. 20460, July.
28. USEPA. 1979a. Methods for Chemical Analysis of Water and Wastes.
EPA-600/4-79-020, USEPA-EMSL, Cincinnati, OH 45269, March.
29. USEPA. 1979b. Handbook for Analytical Quality Control in Water
and Wastewater Laboratories. EPA-600/4-79-19, USEPA-EMSL,
Cincinnati, OH, 1979.
30. Kratochvil, B. and J. K. Taylor. 1981,. Sampling for Chemical
Analysis. Analytical Chemistry 53, 8, 924A-938A.
31. Claassen, H. C. 1982. Guidelines and Techniques for Obtaining
Water Samples that Accurately Represent the Water Chemistry of an
Aquifer. U. S. Geological Survey, Open-file Report 82-1024, Lake,
CO, 49 pp. '
32. Ingamells, C. 0. 1974. New Approaches to Geochemical Analysis and
Sampling. Talanta 21, 141-155.
33. Ingamells, C. 0. and P. Switzer. 1973- A Proposed Sampling
Constant for Use in Geochemical Analysis. Talanta 20, 547-568.
34. Keely, J. F. 1982. Chemical Time-Series Sampling. Ground Water
Monitoring Review, 29~37.
35. Keely, J. F. and F. Wolf. 1983- Field Applications of Chemical
Time-Series Sampling. Ground Water Monitoring Review, 26-33.1
36. Hansen, E. A. and A. R. Harris. 1980. An Improved Technique for
Spatial Sampling of Solutes in Shallow Ground Water Systems. Water
Resources Research 16, 4, 827-829.
165
-------
37. Gillham, R. W. 1982. Syringe Devices for Ground-Water Sampling.
Ground Water Monitoring Review, Spring 1982, 36-39.
38. Barvenik, M. J. and R. M. Cadwgan. 1983. Multilevel Gas-Drive
Sampling of Deep Fractured'Rock Aquifers in Virginia. Ground Water
Monitoring Review, Fall, 1983, 34-40.
39. McLaren, F. R., R. Armstrong, G. M. Carlton. 1982. Investigation
and Characterization of Large-scale Ground Water Contamination in
Alluvial Aquifers. Presented at Water Pollution Control Federation
Conference, Oct. 3~8, 1982, St. Louis, MO, 20 pp.
HO. Pankow, J. F., L. M. Isabelle, J. P. Hewetson, and J. A. Cherry.
1984. A Syringe and Cartridge Method for Down Hole Sampling for
Trace Organics in Ground Water. Ground Water 22, 3, 330-339.
41. Eccles, L. A. and R. R. Nicklen. 1978. Factors Influencing the
Design of a Ground Water Quality Monitoring Network. Water
Resources Bulletin, Establishment of Water Quality Monitoring
Programs. American Water Resources Association, June 1978,
196-209. ;
42. Todd, D. K. 1980. Ground Water Hydrology. John Wiley and Sons,
N.Y., 534 PP.
43. Nelson, J. D. and R. C. Ward. 1981. Statistical Considerations
and Sampling Techniques for Ground-Water Quality Monitoring. Ground
Water 19, 6, 617-625.
44. Casey, D., P. N. Nemetz and D. H. Uyeno. 1983. Sampling Frequency
for Water Quality Monitoring: Measures of Effectiveness. Water
Resources Research 19, 5, 1107-1110. •
45.
46.
47.
48.
49.
Cook, J. M. and D. L. Miles. 1980. Methods for the Chemical
Analysis of Ground Water. Report 80/5, Institute of Geological
Sciences, Natural Environment Research Council, U. K. London,
55 pp.
APHA, AWWA, WPCF. 1980. Standard Methods for the Examination of
Water and Wastewater, 15th Edition, American Public Health
Association, American Water Works Association, and the Water
Pollution Control Federation.
Barcelona, M. J. 1984.
Water 22, 1, 18-24.
TOG Determinations in Ground Water. Ground
Baker, E. L., P. J. Landrigan, P. E. Bertozzi, P. H. Field,
B. J. Basteyns and H. G. Skinner. 1978. Phenol Poisoning Due to
Contaminated Drinking Water. Arch. Env. Health, March/April, 1978,
89-94.
Elder, V. A., B. L. Proctor and R. A. Kites. 1981. Organic
Compounds Found Near Dump Sites in Niagara Falls, N.Y. Environ.
Sci. and Techn. 15, 10, 1237-1243.
166
-------
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63-
Yare, B. S. 1975. The Use of a Specialized Drilling and
Ground-Water Sampling Technique for Delineation of Hexavalent
Chromium Contamination in an Unconfined Aquifer, Southern
New Jersey Coastal Plain. Ground Water 13, 2, 151-15**.
Seanor, A. M. and Brannaka. 1983. Efficient Sampling Techniques.
Ground Water Age, April, 41-46.
Barcelona,' M. J., J. P. Gibb and R. A. Miller. 1983. A Guide to
the Selection of Materials for Monitoring Well Construction and
Ground-Water Sampling. Illinois State Water Survey Contract Report
#327; USEPA-RSKERL, EPA-600/52-84-024, 78 pp.
Barcelona, M. J., J. A. Helfrich, E. E. Garske and J. P. Gibb.
1984. A Laboratory Evaluation of Ground Water Sampling Mechanisms.
Ground Water Monitoring Review 4, 2, 32-41.
Johnson, E. E;, Inc. 1966. Ground Water and Wells, St. Paul, MN.
pp. 440.
Richard, M. R. 1979. The Organic Drilling Fluid Controversy,
Part I. Water'Well Journal, April, pp. 66-74.
Richard, M. R. 1979. The Organic Drilling Fluid Controversy,
Part II. Water Well Journal, May, pp. 50-58.
Brobst, R. B." 1984. • Effects of Two Selected Drilling Fluids on
Ground Water•Sample Chemistry. Monitoring Wells, Their Place in
the Water Well Industry Educational Session, NWWA National Meeting
and Exposition, Las Vegas, NV, September 1984.
Villaume, J. F. 1985. Investigations at Sites Contaminated with
Dense, Non-Aqueous Phase Liquids (NAPLS). Ground Water Monitoring
Review 5(2), 60-74.
Walker, S. E, 1983. Background Ground-Water Quality Monitoring:
Well Installation Trauma. In: Proceedings of the Third National
Symposium on Aquifer Restoration and Ground-Water Monitoring,
May 25-27, NWWA Fawcett Center, Columbus, OH 1983, p. 235-246.
Strausberg, S. 1983. Turbidity Interferences with Accuracy in
Heavy Metals Concentration. Industrial Wastes, March/April,
pp. 20-21.
Saines, M. 1981. Errors in Interpretation of Ground-Water Level
Data. Ground Water Monitoring Review 1(1), 56-61.
Freeze, A. R. and J. A. Cherry. 1979. Groundwater. Prentice-Hall,
Inc., Englewood Cliffs, New Jersey.
Prosser, D. W. 1981. A Method of Performing Response Tests on
Highly Permeable Aquifers. Ground Water, Vol. 19, No. 6.
167
-------
64. Hvorslev, M. J. 1951. Time Lag and Soil Permeability in
Groundwater Observations. U.S. Army Corps of Engineers Waterways
Experiment Station Bulletin 36, Vicksburg, Virginia.
65. Cooper, H. H., J. D. Broedehoeft, and I. S. Papadopulos. 1967.
Response of a Finite-diameter Well to an Instantaneous Charge of
Water. Water Resources Research, No. 3, pp. 263-269.
66. Theis, C. V. 1955. The Relation Between the Lowering of the
Piezometric Surface and the Rate and Duration of Discharge of a
Well Using Groundwater Storage. Trans. American Geophysical Union,
16,518-524.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
Jacob, D. E. 1950. Flow of Ground Water. Engineering Hydraulics,
edited by H. Rouse, John Wiley and Sons, New York.
Papadopulos, I. S., and H. Cooper. 1967. Drawdown in a Well of
Large Diameter. Water Resources Research, 3( 1):24l-244.
Faust, C. R. and J. W. Mercer. Evaluation of Slug Tests in Wells
Containing a Finite Thickness Skin. Water Resources Research 20.
4, 504-506.
Stallman, R. W. 1956. Numerical Analysis of Regional Water Levels
to Define Aquifer Hydrology Transactions. American Geophysical
Union 37, 4, 451-460. ~~ ~
Holden, P. W. 1984. Primer on Well Water Sampling for Volatile
Organic Compounds. University of Arizona, Water Resources Research
Center, Tucson, AZ, 44 pp.
Ho, J. S-Y. 1983. Effect of Sampling Variables on Recovery of
Volatile Organics in Water. Journal American Water Works
Association, December 1983, 583-586.
Korte, N. and P. Kearl. 1984. Procedures for the Collection and
Preservation of Ground Water and Surface Water Samples and for the
Installation of Monitoring Wells. Bendix Field Engineering
Corporation. Prepared for U.S. Department of Energy
#DE84007264-GJ/TMC-8. January, 1984, 58pp.
Armstrong, R. and F. R. McLaren. 1984.
Catcher in Ground Water Quality Sampling.
Review, Fall 1984, p. 48-53.
The Suction Side Sample
Ground Water Monitoring
Cherry, J. A., R. W. Gillham, E. G. Anderson, and P. E. Johnson.
1983. Migration of Contaminants at a Landfill: A Case Study, 2.
Ground Water Monitoring Devices. Journal Hydrology 63, 31-49.
Barcelona, M. J., J. A. Helfrich, and E. E. Garske. Sampling
Tubing Effects on Ground Water Samples. Analytical Chemistry 57,
2, 460-464. ' ~
168
-------
77.
78.
79.
80.
81
82.
83.
84,
85.
86.
87.
Gdrvis, D. G. and D. H. Stuermer. 1980. A Weil-Head Instrument
for Multi-Parameter Measurement During Well Water Sampling. Water
Research 14, 1525-1527.
Garske, E. E. and M. R. Schock. 1985. An Inexpensive Flow-Through
Cell and Measurement System for Monitoring Selected Chemical
Parameters in Ground Water (In Press). Ground Water Monitoring
Review.
Levine, S. P., M. A. Puskar, P. P. Dymerski, B. J. Warner and
C.S.Friedman. 1983. Cross-Contamination of Water Samples Taken for
Analysis-of Purgeable Organic Compounds. Environmental Science and
Technology 17, 2", 125-127.
Skougstad, M.; W. and G. F. Scarbo, Jr. 1968. Water Sample
Filtration Unit. Environmental Science and Technology 2, 4,
298-301 .
Kennedy, V. C., E. A. Jenne, and J. M. Burchard. 1976.
Backflushing Filters for Field Processing 'of Water Samples Prior to
Trace-Element Analysis. U.S.G.S. Water Resources Investigations,
Open-File Report 76-126.
Kennedy, V. C., G. W. Zellweger and B. F. Jones. 1974. Filter
Pore Size Effects on the Analysis of Al, Fe, Mn, and Ti in Water.
Water Resources Research 10, 4, 785-790.
Silva, R. J. and A. W. Yee. 1982. Geochemical Assessment of
Nuclear Waste Isolation: Topical Report - Testing of Methods for
the Separation of Solid and Aqueous Phases. Lawrence Berkeley
Laboratory Report #LBL-14696, UC-70.
Korte, N. and D. Ealey. 1983. Procedures for Field Chemical
Analyses of Water Samples. Technical Measurements Center, U.S.
Department of Energy, Grand Junction Area Office, GJ/TMC-07(83),
UC-70A, 48 pp.
Schock, M. R. and S. C. Schock. 1982. Effect of Container Type on
pH and Alkalinity Stability. Water Research 16, 10,•1455-1464.
Ligon, W. V. and H. Grade. 1981. PolyCethylene glycol) as a
DiluenT" for Preparation of Standards for Volatile Organics in
Water. Analytical Chemistry 53, 920-921.
Stolzenburg, T. R. and D. G. Nichols. 1985. Preliminary Results
on Chemical Changes in Ground . WAter Samples Due to Sampling
Devices. report to Electric Power Research Institute, Palo Alto,
CA, EA-4118 by Residuals Management Technology, Inc., Madison, WI,
June.
169
«U.S.GOVERNMENTPRDmNCOmCE:1992-6'»8. 003^.07. 3
-------
------- |