United States Office of Research and EPA/600/R-98/058
Environmental Proection Development August 1998
Agency Washington, DC 20460
vvEPA Application of the
Electromagnetic
Borehole Flowmeter
-------
EPA/600/R-98/058
August 1998
by
Steven C. Young, Hank E. Julian, and Hubert S. Pearson
Tennessee Valley Authority
Engineering Laboratory
Norris, TN
and
Fred J. Molz and Gerald K. Boman
Auburn University
Auburn, AL 36849
Interagency Agreement
DW64934812
Project Officer
Steven D. Acree
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, OK 74820
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development, partially funded and collaborated
in the research described here under Interagency Agreement DW64934812 to the Tennessee Valley Authority. It has been subjected
to the Agency's peer and administrative review and has been approved for publication as an EPA document. Mention of trade names
or commercial products does not constitute endorsement or recommendation for use.
All research proj ects funded by the U. S. Environmental Protection Agency that make conclusions or recommendations based on
environmentally related measurements are required to participate in the Agency Quality Assurance Program. This project was
conducted under an approved Quality Assurance Project Plan and the procedures therein specified were used. Information on the
plan and documentation of the quality assurance activities and results are available from the Principal Investigator.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land, air, and water resources.
Under a mandate of national environmental laws, the Agency strives to formulate and implement actions leading to a compatible
balance between human activities and the ability of natural systems to support and nurture life. To meet these mandates, EPA's
research program is providing data and technical support for solving environmental problems today and building a science knowledge
base necessary to manage our ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for investigation of technological and management
approaches for reducing risks from threats to human health and the environment. The focus of the Laboratory's research program is
on methods for the prevention and control of pollution to air, land, water, and subsurface resources; protection of water quality in
public water systems; remediation of contaminated sites and ground water; and prevention and control of indoor air pollution. The
goal of this research effort is to catalyze development and implementation of innovative, cost-effective environmental technologies;
develop scientific and engineering information needed by EPA to support regulator}7 and policy decisions; and provide technical
support and information transfer to ensure effective implementation of environmental regulations and strategies.
This report presents a discussion of the applications of vertical-component borehole flowmctcrs to site characterization with an
emphasis on the design and use of a sensitive electromagnetic flowmeter partially developed under this project. The methodologies
discussed in this document provide cost-effective means for obtaining detailed definitions of hydrogeologic controls on ground-
water flow and contaminant transport. Such information is often essential for evaluation of contaminant fate in the environment and
design of effective monitoring and remediation systems. It is published and made available by EPA's Office of Research and
Development to assist the user community.
Clinton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
111
-------
Abstract
A prerequisite for useful monitoring, modeling, and remedial design strategies is knowledge of the network of hydraulicaHy
active fractures in bedrock aquifers and three-dimensional hydraulic conductivity fields in granular aquifers. Due to a relative lack
of practicable characterization technologies, many ground-water remediation strategies have been designed based on very little, if
any, detailed information regarding fracture or hydraulic conductivity distribution. As a result, the underestimation of aquifer
heterogeneity may contribute to inadequate conceptual models of contaminant transport/fate and inadequate design/performance of
many remediation systems.
Borehole flowmeters are effective tools for measuring vertical variations in the flow field of an aquifer. Although borehole
flowmeters have been used in the petroleum industry for decades, they are not common in ground-water studies due, in part, to the
lack of suitable meters. A vertical component electromagnetic (EM) borehole flowmeter with the versatility and sensitivity required
for application at most sites has been developed by the Tennessee Valley Authority (TVA). The prototype TVA EM flowmeters have
outer diameters of less than 5.25 cm. and inner diameters of either 2.54 cm or 1.27cm. The detection limits for the 1.27-cm and 2.54-
cm flowmeters are near 0.005 L/min and 0.03 L/rnin. respectively. In addition to a low detection limit, attractive features of the
prototype and commercial versions of this EM flowmeter include a wide measurement range, durable construction, no moving parts,
and a design that facilitates use with packer assemblies.
This report describes the operation and application of the TVA prototype EM borehole flowmeters, including theory, design,
calibration, basic field applications, data analysis, and potential effects of various well construction and development procedures on
data. The majority of these results are also applicable to the commercial version of this meter and othervertical component borehole
flowmeters, including heat pulse and impeller tools. Several case studies illustrating specific uses of these tools are also discussed.
-------
Contents
Section Page
Notice ii
Foreword iii
Abstract iv
List of Figures vii
List of Tables ix
I. Theory* and Application of the Electromagnetic Borehole Flowmeter 1
1-1. Overview 1
1-2. Principles of Operation 2
1-3. Collection of Flowmeter Logs 2
1-4. Analysis of Flowmeter Logs 2
a. Flow Distribution Logs 2
b. Hydraulic Conductivity Logs in Granular Aquifers 6
c. Data Analysis 7
1-5. Conclusions 8
II. Electromagnetic Borehole Flowmeter System 9
II-1. Design and Construction 9
II-2. Laboratory Calibration 10
III. Hydrogeologic Characterization Test Design 13
Ili-1. Overview 13
III-2. Measurement Interval Selection 13
III-3. Flowmeter Measurements 13
III-4. Pumping Rate and Drawdown Measurements 14
III-5. Selection of Pumps 14
III-6. Packer Selection 14
III-7. Method of Inflation 15
III-8. Issues Regarding Flowmeter Measurements in the Field 15
a. Well or Borehole Storage 15
b. Selection of Pumping Rate 15
c. Background Electromagnetic Currents 16
d. Rcproducibility of Flowmeter Data 16
III-9. Flowmeter Investigations in Consolidated Materials 16
a. Introduction 16
b. Fractured Rock Applications 17
c. Conclusions 18
IV. Well Construction and Development 19
IV-1. Background 19
a. In-Situ Hydraulic Conductivity Estimates Using Wells 19
b. Formation Damages and Skin Effects 19
IV-2. Well Design 19
IV-3. Well Installation Methods 20
a. Overview 20
b. Drilling Techniques and Hydraulic Conductivity Estimates 20
-------
Contents
(continued)
Section
IV-4. Well Development Methods 21
a. Overview 21
b. Development Methods and Hydraulic Conductivity Estimates 21
V. Field Studies of Well Construction and Development at Columbus, Mississippi 23
V-l. Description of Test Site 23
a. Site Location 23
b. Aquifer Characteristics 23
c. Previous Pumping Tests 23
d. Monitoring Well Installation 24
V-2. Test-Descriptions 25
V-3. Test Analyses and Results 26
a. Ambient Flow Distributions 26
b. Induced Flow Distributions 28
c. Specific Capacity Values 28
d. Transmissivity Values 28
V-4. Summary 31
VI. Field Studies of Well Construction and Development at Mobile, Alabama 33
VI-1. Description of Test Site 3"
a. Site Location 3
b. Aquifer Characteristics 3
c. Well Installation 3
VT-2. Test Description 34
VI-3. Test Results 35
a. Shallow Well A 35
b. Shallow Well B 36
c. Deep Well C 38
VI-4. Summary and Discussion 40
VII. Case Studies 43
VII-1. Field Applications 43
VII-2. Columbus AFB, Mississippi 43
VII-3. Oak Ridge National Laboratory, Tennessee 45
VTI-4. The Oklahoma Refining Company Superfund Site. Oklahoma 46
VII-5. Gilson Road Superfund Site. New Hampshire 46
VTi-6. Mirror Lake, New Hampshire 46
VII-7. Logan Martin Dam, Alabama 48
VII-8. Cape Cod, Massachusetts 50
VII-9. Summary 51
VIII. References 53
APPENDIX A. Field Data Sheets for Borehole Flowmctcr Tests 57
APPENDIX B. Equipment Checklist 59
VI
-------
Figures
Number Page
1-1 Schematic diagram of the TVA electromagnetic borehole flowmeter 3
1-2 Apparatus and geometry associated with a borehole flowmeter test 3
1-3 Graphical illustration of the hypothetical case presented in Table 1-1 4
1-4 Assumed layered geometry within which flowmeter data are collected and analyzed 5
II-1 Mechanical packer (a) assembled on probe and (b) in schematic view 9
II-2 Inflatable packer assembled on probe 10
II-3 Conceptual system for flowmeter calibration 10
11-4 Calibration data for the 1.27-cmlD flowmeter in 5.1 -cm PVC pipe 11
II-5 Frictional losses associated with the 1.27-cmand2.54-cmIDflowmeters 12
111-1 Acoustic-televiewer, caliper, single-point resistance, and flowmeter logs for borehole
DH-14 in northeastern Illinois (Paillet and Keys, 1984; Molz et al., 1990) 17
V-l Ox bow meander at the Columbus AFB site drawn from a 1956 aerial photograph 23
V-2 Well network at the 1-Ha test site 24
V-3 Design of wells used to evaluate the effect of well development on flowmeter tests 25
V-4 Ambient flow profiles after successive well development 27
V-5 Induced flow profiles after successive well development 29
V-6 Effect of well development on the drawdown values for pumping tests at the 38-39-40 well cluster 30
V-7 Comparison of specific capacities at different wells 31
V-8 Calculated transmissivities using early time data 32
VI-1 Vertical cross-sectional illustration of the subsurface hydrologic system at the Mobile site 33
VI-2 Schematic diagram providing the details of the shallow and deep wells constructed at the Mobile site.
Well casing and well screen are the same dimension and schedule. The only difference is the depth
dimensions of the casing and screen 34
VI-3 (Well A) (a) Ambient flow as measured by the EM flowmeter and (b) net flow (total flow -
ambient flow). Dalawere obtained after the 2nd and 3rd developments 36
VI-4 (Well A) Differential net flow obtained (a) after 2nd development, (b) after 2nd
development and an overnight waiting period, and (c) after 3rd development 37
VII
-------
Figures
(continued)
Page
VI-5 (Well B) (a) Ambient flow as measured by the EM flowmeter and (b) net flow (total flow - ambient flow).
Data sets were obtained prior to development and after 1st, 2nd, and 3rd developments 38
VI-6 (Well B) Differential net flow obtained (a) prior to development, (b) after 1st development, (c) after 2nd
development and (d) after 3rd development 39
VI-7 (Well B) Ambient flow after the first development and repeated after an overnight period 40
Vl-8 (Well C) (a) Ambient flow as measured by the EM flowmeter and (b) net flow (total flow -
ambient flow). Data were obtained prior to development and after 1 st, 2nd, and
3rd developments for the ambient flow and after 1st, 2nd, and 3rd developments for net flow 40
VI-9 (Well C) Differential net flow obtained (a) after 1st development, (b) after 1st development and overnight
waiting period, (c) after 2nd development, and (d) after 3rd development 41
Vll-1 Vertical profile of hydraulic conductivity values along the longitudinal axis of the MADE tracer plume 44
VII-2 Depth-averaged hydraulic conductivity values for the lowermost 2 m (left) and the uppermost 2 m (right)
of the unconfined aquifer below the 1-Ha test site at Columbus AFB 44
V1T-3 (a) Ambient flow distribution in a well and (b) the flow distribution caused by constant-rate pumping 45
₯11-4 Ambient and induced flow distributions for wells NE-2, SBB-36, and IBB-4 at the ORC Supcrfund Site 47
VTI-5 Ambient and induced flow distributions for wells I, J, and K at the Gilson Road site 47
Vll-6 Ambient and induced flow distributions for wells FSE-06. FSE-09, and FSE-10 at Mirror Lake,
New Hampshire 49
VII-7 Substantial ambient flow moving from one stratum to another as detected by an impeller flowmeter. The
flow is moving under a dam, the base of which is at an elevation of about 140 m AMSL. Flowmeter
data were used to help select a geologic model for explaining the large amount of leakage of water low
in dissolved oxygen that was observed below the dam 50
Vlll
-------
Tables
Number Page
1-1 Hypothetical data and data analysis related to the application of a borehole flowmeter. 6
1-2 Results of the analysis of data from a borehole flowmeter test to estimate hydraulic conductivity distributions 8
11-1 Sample calibration factors (liter/min/volt) from a linear regression of discharge versus voltage for
electromagnetic flowmeter data 12
III-l Generalized borehole flowmeter field test procedures 13
TIT-2 Problems that: produce errors in EM flowmeter measurements 14
V-l Testing sequence 26
V-2 Early-time CJSL transmissivity values 32
VI-1 Pump-induced flow and ambient flow tests performed on wells A, B, and C 35
VII-1 Characterization objectives for borehole flowmeter studies 43
VII-2 Summary statistics of hydraulic conductivity data obtained using a borehole flowmeter method
and a permeameter analysis method (Hess etal. (1992)) 51
IX
-------
-------
Chapter 1
Theory and Application of the
Electromagnetic Borehole Flowmeter
I-l Overview
The underestimation of aquifer heterogeneity has significantly
contributed to the inefficient design and, consequently,
inadequate performance of many types of remediation systems
(Mercer et al, 1990; Haley et al, 1991). This is especially
true of pump-and-treat remediation systems. In many cases,
interpretation of hydraulic properties is limited by field
measurement capability. Characterization of a hydrogeologic
system requires an effective method for measuring vertical
variations in hydraulic properties. At least three research
teams (Boggs et al., 1989; Rehfeldt et al., 1989b; Molz et al.,
1989, 1990) have evaluated alternative methods for measuring
the vertical variation of hydraulic conductivity (K). These
methods included: small-scale tracer tests, multilevel slug
tests, laboratory permeameter tests, empirical equations based
on grain-size distributions, and borehole flowmeter tests.
These comparisons indicated that the borehole flowmeter test
is one of the most promising methods for measuring the
spatial variability of the hydraulic conductivity fields.
Various types of flowmctcrs based on impeller, thermal-pulse,
tracer-release or electromagnetic technologies have been
devised for measuring flow distribution along a borehole or
well screen. Impeller meters (also known as spinners) have
been used for several decades in the petroleum industry, and
such instruments suitable for some ground-water applications
are now available. A meter of this type was applied in the
field by Hufschmied (1983), Molz et al. (1990), and others.
Most impeller meters cannot operate at the lower flow ranges
often required for ground-water applications. Based on the
need to measure low flow velocities, the United States
Geological Survey (USGS) developed a thermal-pulse
flowmeter (Hess,'l982, 1986; Morin et al., 1988a; Paillet et
al., 1987). Such tools are quite sensitive to low flow
velocities. Electromagnetic flowmeters can also operate
within a relatively large range of flow rates suitable for most
ground-water investigations. Although this document focuses
on the design and application of the electromagnetic
flowmeter developed under this project for the measurement
of the vertical component of flow within a well, the data
collection and methods of analysis apply to other types of
vertical component flowmeters.
A sensitive, vertical-component borehole flowmeter enables
one to accomplish two basic tasks:
1. It allows one to measure the natural (ambient)
vertical flow that exists in manv wells: and
2. If the well is pumped at a steady rate, it enables one
to determine the flow distribution that is entering (lie
well from the surrounding medium.
Ambient flow distributions provide information on the
direction of the vertical component of the hydraulic gradient,
and on the location of hydraulically active fractures in the case
of a fractured formation. If certain conditions are met, flow-
distributions during pumping provide information on the
relative differences in the permeability of selected aquifer
zones or additional information on fracture hydraulic
characteristics.
Flowmeter measurements depend on factors such as skin
effects due to well construction and development and ambient
hydrologic conditions. Such factors may be variable within a
well and some may change with time. Also, it cannot be
overemphasized that a borehole flowmeter should be viewed
only as another tool available to ground-water hydrologists.
As expected, the best hydrogeologic characterizations are
achieved when this tool is used in combination with other
methods (e.g., geophysical measurements and traditional
aquifer tests). Application of several technologies that are
suitable for characterization of various aspects of site
hydrology on different scales are necessary for definition of
hydrogeologic controls on ground-water flow.
A sensitive borehole flowmeter, such as tools based on
thermal-pulse technology or the electromagnetic system
discussed in this report, is most applicable for characterization
objectives that require few assumptions regarding aquifer
properties or other complex variables. These tools are directly
applicable for such objectives as identifying transmissive
intervals in wells, evaluation of Hie effects and state of well
development, and identification of natural flow patterns within
conventional monitoring wells. Information regarding zones
where ground water enters and, in the case of ambient flow,
exits a well is often essential for detailed evaluation of
ground-water monitoring efforts. Studies (e.g., Church and
Granato, 1996; Collar and Mock, 1997; Martin-Hayden and
Robbins, 1997) have demonstrated the potential effects of the
mixing of waters with different chemistries within a
conventional monitoring well on contaminant transport
evaluations. Borehole flowmeters provide direct information
concerning the mix of water that enters a well under either
ambient or pumping conditions. In fractured rock settings,
where water may enter the well from only a few discrete
intervals, the flowmeter allows identification of those
intervals. Such intervals may then be individually targeted for
-------
characterization of water chemistry. Direction of natural flow
within a well under ambient conditions is a direct means of
assessing vertical components of the hydraulic gradient at the
time of the study. This information is useful in conceptu-
alization of the hydrologic setting in various parts of a site.
Such direct uses of these tools are readily apparent and not
discussed in detail in this report.
These tools may also provide data used in the estimation of
detailed profiles of the differences in hydraulic conductivity of
granular aquifer materials. However, many simplifying
assumptions are required in such analyses increasing
uncertainty in the results as is true for more traditional aquifer
tests. This indirect use of data obtained using these tools is
not readily apparent and is discussed in detail in this
document. As with any site characterization tool, study
objectives and site conditions must be critically evaluated
during test selection and design. Results obtained from
borehole flowmeter studies, such as investigations described
in this document, should be viewed as only one piece of
information used in the overall characterization of site
conditions.
1-2 Principles of Operation
An electromagnetic (EM) prototype flowmeter (Figure 1-1)
was developed at the Tennessee Valley Authority (TVA)
Engineering Laboratory in Morris. Tennessee. It consists of an
electromagnet and two electrodes that are cast in a durable
epoxy. The epoxy is molded in a cylindrical shape to
minimize turbulence associated with channeling water past the
electrodes and electromagnet. Having no moving parts, the
flowmeter operates according to Faraday's Law of Induction,
which states that the voltage induced across a conductor
moving at right angles through a magnetic field is directly
proportional to the translational velocity of the conductor.
Flowing water is the conductor, the electromagnet generates a
magnetic field, and the electrodes are used to measure the
induced voltage. Electronics connected to the electrodes
transmit a voltage that is directly proportional to the velocity
of the water.
1-3 Collection of Flowmeter Logs
The concept of borehole flowmeter measurements using the
simplest test design is illustrated in Figure 1-2. A flowmeter
log is recorded before pumping to measure any ambient flow
in the well. This step is very important, particularly in the
case of highly sensitive flowmeters and low permeability7
formations, where the ambient flow may be a significant
fraction of the flow induced by pumping. Following the
ambient test, a pump is placed in the well and operated at a
constant flow rate, QP. After a steady flow field toward the
well is obtained, Hie flowmeter is positioned near the bottom
of the well and a measurement of discharge rate is obtained.
The meter is then raised a distance Az and another reading is
taken. As illustrated in Figure 1-2, the result is a series of
measurements of cumulative vertical discharge, Q, within the
well screen as a function of vertical position, z. Just above the
top of the screen the meter reading should be equal to QP. the
steady pumping rate that is measured independently at the
surface. This procedure may be repeated several times to
ascertain that the readings are stable and flow to the well has
reached a steady-state condition. EM flowmeters arc capable
of measuring upward or downward flow. Therefore, if the
selected pumping rate, QP, causes excessive drawdown or
there are concerns associated with disposal of contaminated
ground water, one can employ an injection procedure as an
alternative.
1-4 Analysis of Flowmeter Logs
a. Flow Distribution Logs
The basic data obtained in the field (Table 1-1) are:
Column (a) - the elevations where readings are
taken.
Column (b) - the ambient flow log, and
Column (d) - the total flow log, which is measured
flow under pumping conditions.
The following data are then calculated:
Column (c) - differential ambient flow,
Column (e) - the net flow log, and
* Column (I) - the differential net flow.
The data in Table 1-1 are shown graphically in Figure 1-3. The
aquifer base is located at z = 0, and the water table or upper
confining layer is located at z = 20 m.
Measurements and calculations are made based on the
assumed layered geometry illustrated in Figure 1-4. Flow to
the well is assumed to be horizontal. The basic flow logs,
Columns (b) and (d) represent upward (positive) or downward
(negative) vertical flow within the well itself; while the
differential flow logs represent horizontal flow in the assumed
layers to the well (positive) or from the well (negative). The
sign convention introduced herein is not universal. However,
the same sign convention should be followed for any
particular application.
For the example of Table 1-1 and Figure 1-3, the ambient flow-
is upward. The differential ambient flow log, obtained by
taking differences between adjacent values of ambient flow in
the well, indicates that water enters the well at varying rates
from the bottom half of the aquifer and exits the well at
varying rates in the top half of the aquifer.
At this stage in the data analysis, a distinction must be made.
The drawdown that is measured while recording the total flow
log is due only to pumping. However, the measured flow log
is due to a superposition of the ambient flow and the pumping
flow. Therefore, to be consistent one must subtract the
ambient flow log from the total flow log to obtain the net flow
log, which is that portion of the total flow due to the measured
drawdown. If this subtraction is not made, the results could be
ambiguous.
-------
Amplifier .
Electromagnet I
Electrodes -
Iron Core -
\
/
']
\
*
,
Top View
Section AA
Side View
Section BB
Magnet
Wire
Magnet
Wire
(E = Vx B)
B Magnetic Field
E Induced EM Field
v Velocity of Fluid
Figure 1-1. Schematic diagram of the TVA electromagnetic borehole flowmeter.
(Discharge from Pump)
Pump
Screen
or
Borehole
Borehole
Flowmeter
To Flowmeter Logger (Q)
LU
Data
Flow (Q)
Figure I-2. Apparatus and geometry associated with a borehole flowmeter test.
-------
re
Q
3
re
Q
c-
o
o
o
o
o
o
15
re
Q
O
O
i
0)
15
c
u
o.
o
Q.
a
O
3
cn
-------
N
II
N
N
II
N
O
a
a
o
II
a
1
1
1
N <
\
x
\
X
j J
X
, ""
{. '"
\ s
--'c
. a
Ei
o
W
O)
E
O)
a>
W
CN CO
CO CN T- O
a
c
re
u
0)
o
u
E
OS
to
"5
a
IB.
£
13
U
'E
c
'E
"
o
o
en
13
0)
3
tfl
B)
3
en
-------
Table 1-1. Hypothetical Data and Data Analysis Related to the Application of a Borehole Flowmeter
Elevation
(m)
Ambient Flow
(L/min)
A Ambient Flow
(L/min)
Total Flow
(L/min)
Net Flow
(L/min)
ANet Flow
(L/min)
(a)
(b)
(c)
(d)
(e)
(f)
20
18
16
14
12
10
8
6
4
2
0
o.o
0.07
0.17
0.33
0.70
0.75
0.60
0.39
0.18
0.05
0.00
-.07
-0.10
-0.16
-0.37
-0.05
+0.15
+0.21
+0.21
+0.13
+0.05
6.0
5.2
4.6
3.6
2.7
2.5
1.8
1.4
1.0
0.3
0.0
6.0
5.13
4.43
3.27
2.00
1.75
1.20
1.01
0.82
0.25
0.00
0.87
0.70
1.16
1.27
0.25
0.55
0.19
0.19
0.57
0.25
The final column in Table 1-1, the differential net flow, is
obtained by calculating the difference between adjacent values
in the net flow log. This yields the horizontal flow that is
entering the well from each interval due to pumping. Under
conditions of steady, horizontal flow into the well, this flow is
proportional to the vertical distribution of horizontal hydraulic
conductivity in the vicinity of the test well (Molz et al, 1988;
1989).
h. Hydraulic Conductivity Logs in Granular
Aquifers
For purposes of these analyses, the aquifer is assumed to be
composed of a series of n horizontal layers (Figure 1-4). In
practice, hydrogeologic information, such as data available
from cores and geophysical logs, may be used to
conceptualize potential geologic layers and choose appropriate
intervals for measurements during the flowmeter study. The
difference between two successive flowmeter readings
obtained under pumping conditions yields the flow, AQr
entering a well screen between the elevations where the
readings are taken, which are assumed to bound layer i (i =
1.2,...,n). The Aq, from the ambient flow log are computed in
an identical manner. Most, aquifers are not composed of
horizontal, homogeneous layers. However, the n-layer case at
this scale is more realistic than the single-layer case of
classical pumping test analyses.
To calculate a hydraulic conductivity profile, three methods
have been detailed in the literature. The first is based on the
Cooper-Jacob equation relating drawdown to a constant
pumping rate in a fully penetrating well (Cooper and Jacob,
1946). The second is based on the numerical results by
Javandel and Witherspoon (1969). Molz and Young (1993)
provided an overview of these methods. A third method is the
two-step procedure described by Kabala (1994), which yields
an S, as well as a K distribution. Drawdown data collected at
different times and flowmeter data are used to estimate the
vertical distribution of hydraulic parameters. Such a
procedure does not require knowledge or estimation of
specific storage (Ss) and may have advantages in certain
situations where Ss is expected to vary greatly. Only the first
two methods are discussed herein. Xiang (1995). Hanson and
Nishikawa (1996), and Ruud and Kabala (1996) provide
additional discussion on numerical evaluation of flowmeter
tests.
In the Cooper-Jacob method, the assumed horizontal flow in
each layer is treated as if it was from an aquifer of infinite
horizontal extent and thickness, Az,. Then for each layer, i,
one can write
(AQ- Aq
InKAz.
In
(I-D
where:
AH,
AQ,
Az,
rw
S;
drawdown in ith layer
induced flow from ith layer
ambient flow from ith layer
horizontal hydraulic conductivity of the ith
layer
ith layer thickness
effective well radius
time since pumping started
storage coefficient for the ith layer.
If one assumes that head losses associated with flow within
the well are negligible, then all the AH, are equal to AH, the
measured drawdown in the test well. If such an assumption
cannot be made, then one must measure the head loss opposite
each layer associated with pumping the test well. Rehfeldt et
al. (1989b) provide further discussion of various possible head
losses.
-------
Often, a value for the storativity of the aquifer being studied
will be estimated, and the question becomes how to use this
information to obtain S, for each layer. In previous studies,
two assumptions have been made. The most basic assumption
is to assume that Ss, the specific storage, is constant, in which
case S, = SsAzi (Morin et aL 1988a; Molz et aL 1989). An
alternate assumption used by Rehfeldt et al. (1989b) is that Ss
varies in such a way that the hydraulic diflusivity of each
layer. KjAz/Sj, remains constant and equal to the hydraulic
diffusivity (T/S) for the aquifer, where T is transmissivity and
S is the storage coefficient. If the latter assumption is made,
Equation (1-1) may be solved for Kj yielding
_
(4 a-
2nAHA
,
1.5 \Tt_
r 1 S
(1-2)
If the constant Ss assumption is made, one can solve Equation
(1-1) for Kj outside the logarithmic term to yield
K. =
(AQ-
2nAHAi
In
(1-3)
which can be solved iteratively to obtain a value for JQ.
Further details may be found in Morin et al. (1988a), Molz et
al. (1989), or Rehfeldt et al. (1989b).
Javandel and Witherspoon (1969) showed that, in idealized
layered aquifers, flow at the well-bore radius, rw, rapidly
becomes horizontal even for relatively large permeability
contrasts between layers. Under such conditions the radial
gradients along the wellbore are constant and uniform, and
flow into the well from a given layer, due to pumping, is
proportional to the transmissivity of that layer, that is:
(AQ-Aq)=
(1-4)
where a is a constant of proportionality^ This condition occurs
when the dimensionless variable tD = Kt/(ssr£)is > 100.
In this expression, K is the average or bulk horizontal
hydraulic conductivity defined as ZKjAz/b, where b is the
aquifer thickness. Ss is the aquifer specific storage, t is time
since pumping started, and rw is the wellbore radius.
To solve for a, sum the (AQj-Aq;) over the aquifer thickness,
to obtain
X (AQ- Aq)= QP= a X
(1-5)
Multiplying the right-hand side of Equation (1-5) by b/b and
solving for a yields
QP
a= -%=
bK
d-6)
Finally, substituting for a in Equation (1-4) and solving for
K;/"K~ gives
(*Q-^)/^., 72
gP/Z>
d-7)
To obtain Equation (1-7), it was assumed that AQ{ and QP do
not change with time and pseudo-steady-state conditions were
reached. Thus, a plot of K;/ K versus elevation may be
obtained from the basic data. If one then has an estimate of
K from a conventional aquifer test (e.g., fully-penetrating
pumping test), dimensional values for K; can easily be
calculated by taking the product of "K" and the flowmeter test
result.
In situations where the inherent assumptions are valid, the
K;/ K approach has practical appeal because the values for rw
and Sp which are difficult to estimate, are not required. Also,
errors in flowmeter readings involving constant multipliers arc
canceled out, and the meter calibration is not critical as long as
its response is linear. However, a reliable aquifer test must be
performed to estimate K . A detailed example of data analysis,
including head loss measurements for each layer, is presented
in the following section.
c. Data Analysis
The data, presented in this section were obtained from a
borehole flowmeter test conducted in a fully-penetrating well
in a confined aquifer located 40 m to 61 m below ground. The
well screen was 10.2-cm diameter slotted PVC pipe (0.025-cm
slots). Testing began with mild redevelopment and cleaning of
the test well screen using injected air.
Prior to the flowmeter survey, caliper and ambient pressure
logs were run. Data obtained from the caliper logs were used
to verify and compute the cross-sectional area of the well, and
the hydraulic head distribution derived from the pressure logs
served as calibration references for evaluating AH; produced
by pumping. No ambient flow was detected. Subsequently, a
pressure transducer and a flowmeter with centralizer were
lowered into the well, followed by a submersible pump. The
pump was started and allowed to ran for about 50 minutes
prior to taking pressure and impeller meter readings.
Listed in the first three columns of Table 1-2 are corresponding
values of depth, head change, and discharge associated with
pumping. This constitutes the basic data resulting from the
flowmeter test. The discharge measured at the 40-m depth
(0.24 m3/min) was taken as the pumping rate QP.
Additional hydraulic information about the aquifer in the
vicinity of the test well was obtained from small-scale
pumping tests. A standard Cooper-Jacob (1946) analysis
resulted in a transmissivity of 0.83 m2/min, a K of 0.039m/
min, a storage coefficient of 4.5 x 1Q-3, and a specific storage
of 2.1 x K>4 m-1. Data analysis proceeds by subtracting
-------
Table 1-2. Results of the Analysis of Data from a Borehole Flowmeter Test to Estimate Hydraulic Conductivity Distributions
z(m)
AHj(m)
G(m3/min)
AQi(m3/min)
K2i(m/min)* K3j(m/min)*
K7|(m/min)*
'Based on Equations 1-2, 1-3, and 1-7, respectively.
Interval (i)
39.6
41.1
42.7
44.2
45.7
47.2
48.8
50.3
51.8
53.3
54.9
56.4
57.9
61.0
0.371
0.366
0.362
0.359
0.355
0.353
0.350
0.348
0.347
0.347
0.347
0.347
0.347
0.347
0.240
0.234
0.229
0.214
0.207
0.195
0.189
0.178
0.136
0.120
0.110
0.077
0.034
0
0.0062
0.0057
0.0147
0.0071
0.0119
0.0057
0.0113
0.0422
0.0153
0.0099
0.0331
0.0436
0.0337
0.014
0.013
0.034
0.016
0.027
0.013
0.027
0.099
0.036
0.023
0.078
0.103
0.040
0.013
0.012
0.033
0.015
0.027
0.012
0.026
0.105
0.036
0.023
0.082
0.109
0.042
0.014
0.012
0.033
0.016
0.027
0.013
0.026
0.098
0.035
0.023
0.077
0.101
0.039
13
12
11
10
9
8
7
6
5
4
3
2
1
neighboring Q values to get the AQ; for each of the 13 layers
(Column 4, Table 1-2), noting that layer 1 is 3.05-m thick and
succeeding layers are 1.52-m thick. Listed in the last three
columns of Table 1-2 are K distributions denoted by K2, K3,
and K7. These were calculated based on Equations (1-2),
(1-3), and (1-7), respectively. In order to account for the small
change in pressure head between the bottom and top of the
screen, Equation (1-7) was modified to read
K. AH(AQ- Aq}/Az.
K
AH. QP/b
(1-8)
Equation (1-8) is a good approximation when the AHj are close
to AH, the average change in pressure head. When the
aquifer/well hydraulic data, including rw = 0.0634 m and
t = 60 min. are substituted in Equations (1-2), (1-3), and (1-8),
one arrives at (meter, minute units)
K2. =
0.8176
AH.
1.5 Conclusions
Presented in this chapter is an overview of vertical-component
flowmeter operation, application, and data analysis. While the
EM flowmeter was emphasized, the data collection and
analysis procedures apply to many types of flowmeters. A
flowmeter of this design measures only one thing directly:
upward or downward water velocity in a borehole, screen, or
casing due to natural head differences (ambient flow) or to
artificial head differences (pumping-induced flow). By
calibrating the meter for volumetric flow-rate and taking
differences between two vertical flow measurements at known
elevations, assuming that no undetected flow is leaking past
the meter, it is possible to calculate the incremental horizontal
flow in the aquifer to or from the well. It is the horizontal
flow distribution over the vertical extent of the well that
enables one to calculate a horizontal hydraulic conductivity
distribution as a function of position in the vertical, K(z). The
calculated K distribution may be absolute, K(z), or relative,
K(Z)/ K . Alternate formulas_are given for both quantities.
The advantage of the K(Z)/ K calculation [Equation (1-7)] is
that well radius and storativity do not have to be known or
estimated.
K3=
0.1044 4 Q In (12,677
and
K7. =
0.80314Q
~AH.
d-9)
with the exception of the bottom layer where Azj = 3.05 m
rather than 1.52 m. The results show that there are no
significant differences between the three alternate hydraulic
conductivity calculations using data from this flowmeter test.
-------
Chapter II
Electromagnetic Borehole
Flowmeter System
II-1 Design Construction
The electromagnetic borehole flowmeter system discussed in
this report can measure vertical flow in various types of wells
and boreholes. The flowmeter system has three main
components: downhole flowmeter, packer assembly, and
above-ground electronics. The flowmeter provides accurate
flow measurements across a four order-of-magnitude range
and fits snugly in wells with casing diameters as small as
5.1 cm (standard 2-inch schedule-40 PVC pipe). A packer
assembly may be attached to the flowmeter to direct flow
through the meter in larger diameter wells. The above-ground
electronics package provides the magnetic drive and converts
the flowmeter signal to a discharge reading.
Two of the major components of the meter are an
electromagnet and a pair of silver chloride electrodes mounted
at right angles to the pole pieces of the electromagnet (Figure
1-1). During operation, the electromagnet creates a strong
magnetic field across the flow passage. As water (die
conductor) flows through the magnetic field, a voltage
gradient is generated. The voltage is proportional to the
average velocity of the water across the magnetic field and is
detected by the electrodes. The magnitude of the voltage is
unaffected by the electrical conductivity of ground water
under normal conditions. The polarity of the generated
voltage is dependent on the direction of flow. Upward flow is
generally designated as a positive voltage and downward flow-
as a negative voltage.
Flowmctcrs of this design can be built to any outer diameter
greater than 4.8 cm. As flow is channeled through this meter,
the applicable flow range is a function of the inner diameter of
the open probe core. Flowmeters with inner diameters (IDs)
of 1.27 cm or 2.54 cm have been produced. The length of the
current prototype flowmeter is 30 cm, but flowmeters as short
as 10 cm have been used successfully. The 1.27-cm and 2.54-
cm ID flowmeters are typically used to measure low and high
flow rates, respectively. In wells or boreholes with relatively
large diameters, the effectiveness of die flowmeter diminishes
unless a packer assembly is used to direct flow through the
flowmeter. Both mechanical (Figure II-1) and inflatable
(Figure II-2) packers have been developed.
The mechanical packer assembly used with prototype meters
was a collar consisting of a rubber gasket sandwiched between
two plexiglass or stainless steel rings that slipped over the
flowmeter and was held in place with set screws. The rubber
gasket was sized to insure a tight seal between the probe and
the inner surface of the well. The mechanical packer was
easily used, but its application was tedious because the friction
between the collar and well made lowering of the flowmeter
Bolts
Downhole
Probe
Retaining Ring
Rubber Gasket
Retaining Ring
Retaining
Rings (2)
Rubber Gasket
(a)
Bolts (4)
(b)
Figure 11-1. Mechanical packer (a) assembled on probe and (b) in schematic view.
-------
Clamps
Air or
Water
Inflatable
Rubber
Bladder
Downhole
Probe
Figure 11-2. Inflatable packer assembled on probe.
more difficult. In situations where the screen or borehole was
uneven, the collar design may not have provided an adequate
seal and was not used.
The inflatable packer for the prototype system consisted of a
rubber sleeve attached to a stainless steel assembly. The
packer assembly easily slipped onto the flowmeter and sealed
with "O" rings. This assembled unit had a diameter of 8.9 cm.
If the rubber sleeve was slipped directly onto the flowmeter,
the flowmeter and sleeve had a combined diameter of 8.7 cm.
Approximately 103 kPa (15 psi) and 172 kPa (25 psi) of
pressure was needed to inflate the rubber sleeve to diameters
of 10 cm and 20 cm, respectively. Inflation was achieved by
injecting water into the rubber sleeve through a tube in the
packer assembly. Both a pressurized chamber at the ground
surface and a submersible pump have been used to inject
water into the sleeve. For most applications, the submersible
pump was the method of choice. However, at shallow depths
(<20 meters) where only a few flowmeter measurements were
desired, the pressurized chamber had an advantage in its
simplicity and relatively short set-up time. The seal provided
by a properly inflated packer was such that the flowmeter
could not be moved by pulling the cable and rope.
The above-ground electronics includes an electromagnet drive.
power supplies, amplifiers, and synchronous demodulator for
converting the voltage from the probe's electrodes to flow
rate. The probe's signal is in the microvolt range and is
typically several orders-of-magnitude less than background
noise. Synchronous demodulation is used to extract the signal
and lias the effect of canceling noise out of phase with the
electromagnetic drive. With additional amplification and
filtering, a DC signal proportional to water velocity through
the flowmeter is generated. The electronics package collects
and processes signals from the flowmeter every second. At
the end of a pre-set time interval or upon keyboard command.
the signals are averaged and the standard deviation is calcu-
lated. This average and standard deviation are displayed,
stored on disk, and printed to a hardcopy device.
II-2 Laboratory Calibration
Calibration of the prototype flowmeter was performed in a
facility located at the TVA Engineering Laboratory, Norris,
Tennessee. The apparatus included standard 5.1.-cm, 10.2-cm,
and 15.2-cm schedule-40 flush-joint PVC pipes. Before and
after any field measurements, calibration checks were
performed at a range of flow rates using the same flowmeter
system (above-ground electronics, cable, and flowmeter) used
in the field. Calibrations were simple and consisted primarily
of establishing a constant uniform flow through a vertical PVC
pipe and comparing the flowmeter measurements to other flow
measurements at the intake and/or the outlet of the PVC pipe
(Figure II-3). Flow rates were maintained by throttling a
pressure valve on the public watcrlinc and measuring the
generated flow rate near the PVC pipe intake with an in-line
commercial flowmeter. For all flows, the baseline rate was
determined by measuring the time required for the discharge
to fill a calibrated container. At lower flow rates (< 30 ml/
niin), the discharge volume was determined by dividing the
weight of water by the temperature corrected density of water.
Electromagnetic flowmeter calibrations are similar to those for
other flowmeter types and require: (1) the use of proper and
certified measuring devices for volume, time, length, weight.
and temperature; and (2) the documentation of calibration
conditions including personnel, date and time, and equipment
set-up. Instructions and guidelines for performing flow
Well Casing
Electronics
Flowmeter
L Water Supply
Calibrated Vessel
O
Figure 11-3. Conceptual system for flowmeter calibration.
10
-------
calibrations can be found in standard texts such as Person
(1983) and Liptak and Venczel (1982). Besides standard
flowmeter calibrations, similar equipment has been used to
investigate phenomena such as turbulence, factional losses,
and flow around the probe. The three main PVC pipes used in
the calibration facility are transparent to permit visual
monitoring of dye releases so that turbulence and flow around
the probe can be evaluated. One of the PVC pipes is fitted
with manometers to measure head losses within the pipe and
flowmeter. All of the PVC pipes possess side ports to permit
the introduction of flow. This allows monitoring of the
flowmeter response as it approaches and passes a horizontal
inlet source. Horizontal inflows may be large enough to
produce turbulence effects, and gauging the sensitivity of the
flowmeter to these inflows was important. Testing indicates
that the electromagnetic flowmeter is insensitive to the
proximity of horizontal inflows.
Figure II-4 provides a sample calibration data set associated
with field testing of the EM flowmeter. The figure shows the
calibration data on both linear and logarithmic scales. The
linear scale illustrates the linear response between volts
(flowmeter signal) and flow. The logarithmic scale shows the
sensitivity of (lie meter at lower flow ranges. The 2.54-cm ID
flowmeter exhibits good repeatability and linearity from about
100 ml/min to 40 L/min. The 1.27-cm ID flowmeter exhibits
good repeatability and linearity from 30 ml/min to 10 L/min.
Below 30 ml/min, the repeatability and linearity of the
Calibration
1.27-cm ID Flowmeter in 5.1-cm PVC Pipe
10
5
2
1
» 0.5
-§
0.2
0.1
0.05
0.02
nm
: *"
_
7
L
-
o - 5-13-92
- 5-27-92
# - 6-29-92
X - 7-08-92
D - 7-10-92
I -+ 1 Standard
Deviation
.*
.*'
* '
*
*
* i
[ ***
- £*7^
***
0.01
0.1 1
Discharge (L/min)
10
10
> 4
6 8 10
Discharge (L/min)
12
14
Figure II-4. Calibration data for the 1.27-cm ID flowmeter in 5.1-cm PVC pipe.
11
-------
1.27-cm ID flowmeter drops significantly. Deviation from
linearity at these low flows may be a result of shifts in the
flowmeter response over time, or problems/uncertainties
associated with maintaining and measuring low flows during
calibration.
In Figure TI-4. each calibration point is the mean of sixty
readings taken over a one minute period. For the majority of
the flow measurements, the relative standard deviation is less
than 2 percent of the average flow rate. Significantly lower
percentages occur at the higher flow rates and slightly higher
percentages occur at the lower flow rates. Calibration data
show that the sensitivity and low end accuracy (below 50 ml/
min) of the 1.27-cm ID flowmeter is less in a 5.1-cm pipe
without a packer assembly, than in larger diameter pipes with
a packer assembly. For the 5.1-cm pipe, the calibration plot is
essentially linear up to 6 L/min. Above 6 L/inin, the
flowmeter signal becomes slightly nonlinear. As no packer
assembly was used in the 5.1-cm application, the nonlinearity
is probably the result of water flow around the probe.
Flowmeter sensitivity can be expressed in a calibration factor.
Calibration factors are used to convert voltage from the probe
electrodes to flow and have the units of volume per time per
volt (e.g., liter/minute/volt). Table II-1 provides example
calibration factors calculated for different combinations of
Table 11-1. Sample Calibration Factors (Liter/Min/Volt) from a
Linear Regression of Discharge Versus Voltage for
Electromagnetic Flowmeter Data
1.27-cm ID
EM Flowmeter
2.54-cm ID
EM Flowmeter
Flowmeter in a 5.1-cm pipe* 1.46
Flowmeter with a mechanical
collar in a 10.2-cm pipe* 1.38
Flowmeter with an inflatable
packer in a 15.4-cm pipe* 1.32
3.99
3.95
3.94
flowmeter sizes, packer types, and well diameters. Greater
sensitivity is achieved with a packer because less water flows
around the exterior of the flowmeter. The flowmeter should
be calibrated using the same equipment and in the same type
of casing as the wells to be surveyed prior to use.
At flow rates above 10 L/min, two concerns exist with use of
the 1.27-cm ID flowmeter: high frictional losses through the
orifice, and an electrode voltage that will exceed the capacity
of the above-ground electronics. High frictional losses (Figure
II-5) affect the hydraulic head distribution and, consequently,
the distribution of flow in the well and through the meter. In
selecting between electromagnetic flowmeters with different
orifice diameters, the trade-offs associated with decreases in
the detection limit and increases in frictional losses should be
considered. As a general rule, the frictional losses should be a
concern if they exceed 10 percent of the total drawdown in a
well or borehole for cases where the data will be used to
calculate a hydraulic conductivity profile.
25
5 20
1
s
E 15
o
10
re
c
o
2.54-cm ID (10-cm length)
- 2.54-cm ID (25-cm length)
1.27-cm ID (25-cm length)
10 15 20 25
Flow (L/min)
30
35
' pipe is schedule-40 PVC
Figure II-5. Frictional losses associated with the 1.27-cm and
2.54-cm ID flowmeters.
12
-------
Chapter III
Hydrogeologic Characterization Test
III-l Overview
The primary objective of an electromagnetic borehole
flowinder test is to measure the profile of horizontal flow into
or out of designated aquifer intervals during ambient and/or
pumping conditions. These profiles are used to identify the
zones of relatively high and low permeability. Combined with
a transmissivity measurement or the proper drawdown data at
the tested well, the flowmeter data may be used to calculate a
vertical profile of horizontal hydraulic conductivity. In many
situations, the pumping tests performed concurrently with
such flowmeter tests will provide the drawdown data
necessary to calculate a transmissivity value.
Table III-l lists the general procedures which were used in
these studies to perform borehole flowmeter tests. Detailed
descriptions of these procedures and associated equipment are
contained in the following sections. Appendix A provides the
forms used to record the field data and Appendix B provides
an equipment checklist.
III-2 Measurement Interval Selection
Selection of appropriate intervals for obtaining flowmeter
measurements in porous media involves consideration of site
stratigraphy interpreted from geologic and geophysical logs.
However, such logs generally provide only enough
information to determine the basic hydrogeologic framework
and not the detailed variability within geologic units.
Although, these logs may provide constraints for specifying
measurement intervals, the scale on which measurements are
ultimately obtained will usually be based on other factors.
Such factors include the pumping rate to be used during Hie
test. Constraints on pumping rate, such as limitations related
to water storage, treatment, and disposal, may necessitate
increased interval size to provide meaningful data. The
thickness of the disturbed /one surrounding the well screen is
also a factor that affects interval selection. Results of studies
by Ruud and Kabala (1997) indicate that skin effects bias the
distribution of flow to a well. As the ratio of the flowmeter
measurement interval thickness to the thickness of the
disturbed zone decreases, the bias in the test results increases.
In general, local deviations from horizontal flow will occur
during most tests. Errors associated with these deviations will
increase as the size of the measurement interval decreases.
111=3 Flowmeter .Measurements
Prior to testing, the flowmeter is connected to the electronics
and permitted to warm up. This warm-up period was
approximately 30 minfor the prototype electromagnetic
system. Well construction logs, including a caliper log of the
test well, are reviewed if there is any doubt about the screen or
borehole diameter. Field testing has shown that the calibration
of the prototype electromagnetic flowmeter may drift up to
10 millivolts (0.1 percent full scale) over several hours if large
temperature changes (> 10°C) occur. In many instances, this
temperature sensitivity is not a concern because of the
relatively short time needed to obtain a flow log. However,
flowmeter calibration at zero flow is checked prior to initiation
of the test and at the end of each test. This is accomplished by
positioning Hie flowmeter at a location in Hie well where no
Table 111-1. Generalized Borehole Flowmeter Field Test Procedures
1. Measure ground-water elevation in the well relative to a fixed datum (such as top of well casing) and total well depth. Examine well
construction, geologic, and geophysical logs to determine measurement intervals.
2. Zero flowmeter under no-flow conditions. Measure the ambient vertical flow profile. In these studies, this profile typically consisted of a
minimum of ten measurements. Obtain flowmeter reading under no-flow conditions to evaluate baseline drift.
3. Install equipment as for a single-well pumping test, including pressure transducer, data logger, and pump. Place intake hose or pump
above the pressure transducer and within the well casing above the screen, if possible. Measure ground-water elevation to insure static
conditions are attained.
4. Start constant-rate pumping test. Accurately measure pump discharge rate on a routine basis. Record drawdown. Verify that wellbore
storage effects have dissipated and estimate transmissivity based on drawdown data.
5. After pseudo-steady-state conditions have been achieved, obtain flowmeter measurements at the same elevations as the ambient
measurements. Repeat flowmeter measurements to demonstrate a stable flow profile has been achieved. Shut system down and obtain
flowmeter reading under no-flow conditions.
13
-------
ambient flow exists. In a well, no ambient flow should exist
above or below the screened interval. In a borehole, no
ambient flow should occur in the cased interval. If there is
doubt about the location of a no-flow zone, the problem is
resolved by manually plugging the flowmctcr and lowering it
into llie well. Plugging may be accomplished by securely
fastening a cork in the lower orifice of the flowmeter.
Lowering or raising the flowmeter induces water movement in
a wellbore and it may require several minutes to regain
quiescent conditions in low permeability aquifers. After
moving the flowmeter, measurements should not begin until
the flowmeter signals have stabilized. If the readings do not
stabilize, and/or if the standard deviation is approximately ten
times greater than observed in the calibration data for the
measured flow, then reject the reading and attempt to diagnose
the problem. Table III-2 lists problems that have been
encountered in Hie field. Problems 3 and 4 can often be
remedied by rapidly moving the flowmeter up and down the
well or borehole several times.
Table 111-2. Problems that Produce Errors in EM Flowmeter
Measurements
1. Insufficient time to regain quiescent conditions in well after
flowmeter movement.
2. High ground currents (above- or below-ground power lines or
power sources in vicinity of well or boreholes.)
3. Coating on the electrodes such as mud or oil.
4. Blockage of signal path to electrodes by gas bubbles.
5. High flow rate entering well near location of flowmeter.
IH-4 Pumping Rate and Drawdown
Measurements
One of the most important aspects of the pumping test is the
pumping rate, which should be high enough to achieve a
measurable drawdown, if analyses of time/drawdown data are
desired. The rate should also be low enough to support the
assumption of negligible head loss within the well, if head loss
across each interval is not measured. Additional
considerations include methods for maintaining a constant
flow rate throughout the test and constraints due to
containment, treatment, and disposal requirements for
contaminated water. In practice, it is often necessary to set the
pumping rate by trial and error. After each trial start,
sufficient time should be allowed for the well to recover
before beginning a subsequent pumping test.
Constant flow rates arc important for two reasons. First, a
constant flow rate permits calculation of horizontal inflow
from aquifer intervals by subtracting Hie cumulative flow rates
at different elevations. Second, a constant flow rate permits
the use of conventional pumping test analytical solutions to
calculate a transmissivitv value at a well location.
Problems associated with maintaining a constant flow rate are
greatest where large drawdowns occur and the lift require-
ments of a pump are affected. If no adjustments are made to
the pump as the drawdown increases, the pumping rate will
decrease until the drawdown stabilizes. In regions of low
transmissivity, the problems associated with a constant flow
rate can be severe because of large drawdowns. Although data
analysis methods can be adjusted to compensate for changes in
llie pumping rate, these changes are not desirable. Injection
tests may be considered instead of pumping tests if large
drawdowns are expected.
Accurate measurement of drawdown response during the
pumping test generally requires a pressure transducer and data
logger to record data at frequent intervals. As any movement
in the well has die potential to alter the water level or the
position of the transducer, sufficient drawdown data for
calculating transmissivity should be collected before
flowmeter testing begins.
III-5 Selection of Pumps
The type of pump used for the test depends on site conditions,
well construction, and available power. If the depth of the
water from the top of casing will not exceed approximately 7
m. then a surface pump may be used. For flows greater than 4
L/min, centrifugal surface pumps have proven to be reliable.
This type of pump is tolerant of suspended solids and high or
blocked discharge pressures, has good lift, and can provide an
order-of-magnitude flow range by manually adjusting the back
pressure. A disadvantage of this pump is the difficulty of
presetting the pump to a known rate. A peristaltic pump is
often a good choice for flows less than 4 L/min. It is easy to
use, there is no contamination of the pump itself, and it may
be preset within reasonable tolerances.
Dissolved gases may be a problem with surface pumps when
air bubbles accumulate in the intake line. The intake line
pressure will be lowest near the surface pump. In some cases,
the pressure may be sufficiently low to cause the water to
degas. If sufficient gas accumulates in the intake line, then
flow will be restricted to Hie pump and the pumping rate will
decrease. In situations where degassing is a concern, the
pump should be located above the well head to allow gas
movement upward and through the pump.
If the depth to water exceeds approximately 7 m, then
submersible pumps are required. In general, submersible
pumps present less operational problems than surface pumps.
However, their use is often limited to wells with an ID greater
than 5.1 cm due to the size of most pumps and the induced
flow test design. If a submersible pump is used, it should be
small enough to fit into the well and allow passage of
flowmeter and pressure transducer cables.
III-6 Packer Selection
For wells greater than 5.1 cm in diameter, a packer may be
needed to channel water through the flowmeter. If the
velocities are sufficiently high, and the well has a constant
14
-------
diameter, the flowmeter may be operated as a velocity meter
without a packer. When used as a velocity meter, flow occurs
around, as well as through, the EM flowmeter. Hence, a
different set of calibration data is required in lieu of that
obtained under conditions where packer assemblies arc used.
Where low flows are of concern and/or (lie well or borehole is
not of constant diameter, a packer generally is required.
Packer systems of different designs may be appropriate for
these investigations.
III-7 Method of Inflation
Two methods generally were used for inflating/deflating the
prototype inflatable packer assembly. One method
incorporated an above-ground pressure chamber and the other
used a submersible pump positioned above the flowmeter.
Inflation using the pressure chamber was relatively slow and
use was limited to sites where the water table was relatively
shallow. The submersible pump has the advantage of rapid
inflation and operates at wide ranges of depth to water and
pressure heads. At Oak Ridge National Laboratory, Oak
Ridge. Tennessee, submersible pump inflation has been
successful at depths below 300 m, and depths-to-water of
greater than 15 m.
Installation using the above-ground pressure chamber was
relatively simple. The pressure chamber has a fill port, a vent
port, a pressure gauge, and a discharge port attached to a tube
that inflates the packer. The most common setup involved
filling the chamber 75 % full with water. The pressure
chamber was charged to the required pressure with an air
compressor or air tank. When the discharge valve was
opened, the packer assembly inflated. Packer inflation was
monitored by observing the decline in water level in the
chamber. Adequate inflation was checked by tugging on a
rope connected to the .flowmeter. Packer deflation was usually
accomplished by venting only the connection hose.
In situations where the packer was inflated at large depths, or
inflated and deflated relatively quickly, a submersible pump
was used with the packer assembly. Several alternative and
equally viable submersible pump configurations exist. A small
ground-water sampling pump located approximately 0.3m
above the packer was used in these studies. The pump
discharge was connected to the packer via a tube and integral
solenoid valve. The assembly was wired such that the valve
could not be closed when the pump was operating. In
operation, the pump was turned on and the valve was opened
to inflate the packer. Once the packer was inflated, the pump
was stopped and the valve closed simultaneously. For
deflation, the valve was opened, allowing the relatively high-
pressure water inside the packer to flow back through the
pump. Operation was reasonably fail safe as the pump could
not operate against a closed valve, and the valve would open
and deflate the packer upon power failure.
Inflation times ranged from approximately 10 s to 15 s for a
10.2-cm well to greater than 3 min for a 17.8-cm well. At
shallow depths, packer inflation was checked by pulling the
cable to determine if the flowmeter was firmly in place.
However, at depths of 500 ft and greater the weight of the
cable makes it difficult to check for inflation. At these depths,
rapid increase in the electrical current drawn by the pump was
monitored as an indicator of inflation.
111=8 Issues Regarding Flowmeter
Measurements in the Field
Successful performance of investigations using the procedures
described in this document requires knowledge of the
mechanics of single-well pumping tests. This understanding
provides the framework necessary to designate appropriate
pumping rates, ascertain that flow conditions have stabilized,
and select elevations for flowmeter measurements. In
additioa familiarity with local geology, analytical solutions
for pumping tests, and issues such as those listed below is
necessary.
a. Well or Borehole Storage
During a pumping test, a common assumption is that the
pumping rate equals the discharge from the aquifer. This
assumption is acceptable only as long as a small percentage of
discharge originates from within the well or borehole. In any
large diameter well, borehole storage is a potential concern. If
borehole storage dominates the response of a single-well
pumping test, the time/drawdown data will appear as a straight
line on a log-log plot with a slope of one. During a pumping
test, the effects of borehole storage diminish with time and
may be ignored at late times. Until storage effects dissipate.
the drawdown will be less than if storage effects were not
present.
Two problems exist if flowmeter measurements are made
before borehole storage dissipates. One problem is that the
total flow into the well from the aquifer is increasing with
time. Hence, the flow distribution to the well will be changing
with time. Another problem is that calculated transmissivity
values will be too high unless the drawdown values are
corrected for wellbore storage. When performing a flowmeter
test of this desiga borehole storage should be considered in
the analyses, especially in low permeability aquifers or in
boreholes of large diameters.
b. Selection of Pumping Rate
Factors affected by the pumping rate include the magnitude of
the total discharge, the drawdown response, and friction
losses. In situations where ground-water contamination is of
concern and discharged water may require treatment, a lower
pumping rate is beneficial. Minimizing the drawdown also
reduces the importance of vertical flow in unconfined aquifers.
As discussed in Chapter I, horizontal flow is an assumption in
the data analysis. For large drawdowns, horizontal ground-
water flow is not a valid assumption for unconfined aquifers.
During any flowmeter test, frictional losses occur through the
well screen, well casing, and flowmeter. Because they are
15
-------
difficult to measure in the field and can complicate the
analysis of flowmeter data, these losses should be minimized.
Conditions for which frictional losses may be of concern
include heterogeneous highly transmissive aquifers and deep
boreholes. In the former case, frictional losses associated with
high flows entering the well through a narrow interval may be
a problem if the total drawdown is small (e.g.. < 25 cm) as the
well loss becomes a significant fraction of the drawdown. In
deep boreholes, cumulative friction losses associated with
flow through a well pipe or borehole may be significant.
Concerns associated with low pumping rates include:
possible reduced sensitivity of the electromagnetic flowmeter,
difficult}' in monitoring the drawdown response, and
adequately stressing the aquifer. Hence, although low
pumping rates might be desirable, they should be high enough
to ensure that the aquifer and well responses can be
accurately measured with the available equipment.
c. Background Electromagnetic Currents
The flowmeter shell and telemetry cables are shielded
against ground currents, but the inlet and outlet of the
prototype flowmeter arc not shielded. As a result, large
background voltage gradients may influence the circuitry.
Although the flowmeter is designed to reject such noise, its
signal is in the micro-volt range and large voltage
gradients may produce masking effects. Several wells
where instabilities occurred in the flowmeter readings have
been observed during development and use of the
flowmeter. In such cases, the problems were believed to
be caused by background currents. One of these wells.
located near a building with large machinery, was surveyed
twice. The first survey occurred with the machinery
operating. During this survey, unstable flowmeter readings
were recorded. The second survey occurred when the
machinery was turned off. Stable, reproducible flowmeter
readings were recorded.
Due to the uncertainties associated with field testing, some
simple laboratory tests were performed to better identify
the problem. The tests involved inducing a voltage
gradient across a pipe in the calibration facility and
recording flowmeter responses. As long as the induced
gradient was a sine wave synchronized with the power
line, no instabilities were observed. However, once the
voltage gradient varied in both amplitude and frequency,
instabilities occurred. To shield the flowmeter from the
voltage gradient, a metal screen was placed across the
flowmeter's outlet and inlet. This modification reduced the
sensitivity of the flowmeter's response, but appeared to
eliminate the instabilities induced by background currents.
In situations where electromagnetic background currents
arc large, problems may occur with the performance of the
electromagnetic flowmeter. Preliminary investigations
indicate these problems may be overcome by providing
additional shielding for the flowmeter. However,
additional testing is required before appropriate shielding
can be developed.
d. Reproducibility of Flowmeter Data
A concern with any test is the reproducibility of the data.
During a borehole flowmeter test, it is a good practice to
repeat several flowmeter measurements without changing
flowmeter position. If two measurements produce similar
means and standard deviations, then the flowmeter system is
assumed to be functioning properly. As standard practice, the
flowmeter was considered to be performing adequately if two
successive measurements produced mean values that
overlapped in the range of their standard deviations.
In developing a quality assurance plan for field testing, two
types of duplicate measurements should be taken. The first
type involves two successive measurements without moving
the flowmeter. An assumption in comparing these values is
that flow conditions in the aquifer have remained constant
between measurements so that any differences reflect the
uncertainties associated with the measurement technique. The
second type involves measurements at the same location but at
different times. Over periods of minutes and/or hours, the
flow conditions in the aquifer may change. Differences in
flow measurements would be primarily produced by variations
in the saturated thickness of an unconfincd aquifer in response
to a pumping test, variations in the pumping rate caused by
fluctuations in the performance of the pump, and inability to
exactly reoccupy the same elevation in the well. The second
type of duplicate measurement reflects the error associated
with assuming that the flow conditions are at steady-state.
III-9 Flowmeter Investigations in Consolidated
Materials
ft Introduction
Many of the most intractable ground-water contamination
problems occur in consolidated rocks and sediments where
transport is dominated by flow through fractures or solution
features. In systems where flow is dominated by secondary
porosity (fractures, conduits, etc.), the identification of
fractures that are hydraulically active and how they are
interconnected is often important. This task may be
complicated by the general observation that many fractures arc
inactive. Therefore, identifying fracture distributions using
only a caliper log or a televiewer log provides limited
information on the actual flow distribution that exists (Paillet
et al, 1987; Paillet and Hess, 1987; Molz et al., 1990).
Characterization of ground-water flow in fractured media will
often involve measurements of flow to or from individual
fractures or fracture zones. Over the past decade, researchers
at the U.S. Geological Survey (USGS) have been refining
procedures developed using a sensitive heat-pulse flowmeter
(Hess, 1982, 1986; Paillet et al., 1987; Paillet and Hess, 1987;
Hess and Paillet 1990; Paillet and Kapucu, 1989; Paillet,
1991a, 199Ib, 1991c; Paillet et al., 1992). It is mainly their
contributions that will be reviewed in the remainder of this
section. The applications described below illustrate the need to
incorporate a variety of geological, hydrogeological, and
16
-------
geophysical tools in the investigation of site hydrology.
Information from only one tool, such as a borehole flowineter,
will generally be insufficient to evaluate ground-water flow
and contaminant transport and fate issues. This integrated
characterization approach is necessary for investigations in
porous media as well as fractured rock settings.
h. Fractured Rock Applications
Some of the most direct uses of a sensitive borehole flowmeter
are in fractured rock investigations. These applications
demonstrate the advantages in combining flowmeter
information with other types of geophysical data. Perhaps the
best way to describe this process is to present an example
application in fractured dolomite in northeastern Illinois. This
study was perfonned by the USGS (Molz et al, 1990).
Acoustic-televiewer, caliper, single-point-resistance, and
flowmeter logs were obtained in a 64-m deep borehole (DH-
14) in the northeastern Illinois area as part of a study of
ground-water flow and transport in fractured dolomite (Figure
III-l). The acoustic-televiewer log is a magnetically
orientated, television-like image of the borehole wall which is
produced with a short-range sonar probe (Zemanck et al.,
1970). Irregularities in the borehole wall, such as fractures
and vugular openings, absorb or scatter the incident acoustic
energy, and result in dark features on the recorded image.
Such televiewer logs may be used to determine the strike and
dip of observed features. The acoustic-televiewer and caliper
logs for borehole DH-14 (Figure III-l) indicate a number of
nearly horizontal fractures that seem to be associated with
bedding planes. The larger of these fractures are designated A,
B, C, and D, respectively. The caliper log indicates that the
major planar features on the televiewer log are large fractures
or solution openings associated with substantial borehole
diameter enlargements. The large but irregular features on the
televiewer log between fractures B and C also are associated
with borehole enlargements, but these are interpreted as
vugular cavities within the dolomite rather than fractures. The
Caliper Log
ACOUSTIC
TELEVIEWER LOG DIAMETER, IN INCHES
5678 9 10
RELATIVE SINGLE-POINT
RESISTANCE LOG
Flowmeter Log
DOWNFLOW, IN GALLONS
PER MINUTE
6.0 5.8
0.2
0.0
INFLOW
OUTFLOW
INFLOW
OUTFLOW
30
OL
D
40
a:
UJ
50 -
H
Q_
UJ
Q
60
23.0 22.0
0.5 0.0
DOWNFLOW, IN LITERS
PER MINUTE
Figure III-l Acoustic-televiewer, caliper, single-point resistance, and flowmeter logs for borehole DH-14 in northeastern Illinois (Paillet
and Keys, 1984; Molz et al., 1990).
17
-------
single-point-resistance log indicates abrupt shifts in resistance,
at depths of about 40 m and 56 m. These shifts may reflect
differences in the dissolved-solids concentration of the water
in the borehole.
The pattern of vertical flow determined by the flowmeter
measurements indicates the probable origin for the inferred
water quality contrasts in the borehole (Figure III-l). The
flowmeter log indicated downflow, which probably was
associated with naturally occurring hydraulic-head differences,
causing water to enter at the uppermost fracture, A, and exit at
fracture B. A much smaller flow of water with the same
electrical conductivity and dissolved-solids concentration
continued down the borehole to fracture C. At this fracture,
the downflow increased and the water inflow apparently
contained a greater concentration of dissolved solids, which
accounts for the shift to greater electrical conductivity. This
increased downflow exited the borehole at fracture D, where
there was another, somewhat smaller, shift in single-point-
resistancc. Although not rigorously proven from the
geophysical logs, the second shift in resistance appears to be
associated with the dissolved-solids concentration of the water
entering at fracture C.
Subsequent water sampling confirmed that there were
differences in the dissolved-solids concentration of the water
at the different depths. Sample analysis indicated that the
water entering at fracture A had a dissolved-solids
concentration of about 750 mg/L; and the water entering at
fracture C had a dissolved-solids concentration of about 1,800
mg/L. In this instance, the geophysical data, especially the
thermal-pulse flowmeter data, were useful in planning
subsequent packer testing of the aquifer and in interpreting
water quality measurements. Identification of substantial
natural differences in background water quality was useful in
conceptualization prior to modeling of conservative solute
transport. At the same time, measurements of vertical velocity
distributions in the borehole provided useful indications of
hydraulic head differences between different depth intervals.
This information could not be obtained from conventional
water level measurements without the time-consuming
installation of packers at multiple levels in all of the boreholes
at the site.
Paillet et al. (1992) consider the application of borehole
flowmeters to measure flow transients in pumped boreholes
and in adjacent observation boreholes. Their approach is
based on the theoretical observation by Long et al. (1982) that
fracture connections are more important than local fracture
aperture in controlling the rate of flow through random
distributions of finite fractures. Thus, their technique is
designed to identify fracture connections. Paillet et al. (1992)
also consider approaches to the problem at different scales of
measurement which begins to attack the question of how
individual fractures and fracture sets arc integrated into larger-
scale flow systems.
c. Conclusions
The previous study illustrates potential applications and
integration of borehole flowmeters with other geophysical
tools in the interpretation of flow in fractured aquifers. The
relative ease and simplicity' of flowmeter measurements
permits reconnaissance of naturally occurring flows prior to
hydraulic testing and identification of transient pumping
effects. Flowmeter surveys may provide a valuable means by
which to identify fracture interconnections and solute transport
pathways during planning for much more time-consuming
packer and solute studies. The borehole flowmeter is
especially useful at sites where boreholes arc intersected by
permeable horizontal fractures or bedding planes. The simple
and direct measurements of vertical flows provided
information pertaining to the relative magnitude and vertical
extent of naturally occurring hydraulic-head differences in a
few hours of measurement. While the studies described in this
section did not involve contaminated ground water, the
potential application to contaminant migration problems and
monitoring well screen location is obvious.
18
-------
Chapter IV
Well Construction Development
IV-1 Background
a, In-Situ Hydraulic Conductivity Estimates
Using
Well installation results in disturbance of aquifer materials.
This immediately raises Hie question of what effects, if any,
well construction and development techniques have on
hydraulic conductivity measurements? There arc many
considerations that contribute to the answer of this question.
Such considerations include the following.
1. Is the subsurface material consolidated or
unconsolidated?
2. What is the relative importance of primary versus
secondary porosity?
3. What drilling technique is selected?
4. Is the well constructed with an artificial filter pack?
5. What drilling fluid, if any, is used?
6. How is the finished well developed?
7. What scale does the hydraulic conductivity
measurement represent?
It is beyond the scope of this report to provide a compre-
hensive answer to the question of how well construction and
development affects hydraulic conductivity measurements.
Especially for small-scale measurements, this potential
problem is neither well-documented nor understood. The
approach herein will be to discuss the various drilling and
development methodologies, review previous studies, discuss
problems, and present test results concerning the sensitivity of
borehole flowmeter measurements to well development and
construction from two test sites.
b. Formation Damages and Skin Effects
All drilling methods alter the hydraulic characteristics of an
aquifer near the wellbore. The impairment may be caused by
the physical rearrangement of the matrix of aquifer material,
by the smearing of silt and/or clay particles across the
borehole face, or by invasion of drilling fluids or solids into
aquifer formations. The amount of damage that occurs is
related to the drilling method used for well construction and
subsurface geology. The changes in hydraulic conductivity
resulting from formation damage can produce skin effects in
the near-well aquifer formations.
The term "skin effect" was introduced by van Everdingen and
Hurst (1949) when they discovered mismatches between
analytical solutions and field data from well tests. The skin
effect was caused by mud particle invasion into aquifer
formations during well installation producing a decrease in
hydraulic conductivity. Skin effects can be either negative or
positive and represent increases or decreases, respectively, of
hydraulic conductivity near the wellbore. Negative skin
effects can be created at wellbores where drilling causes
fractures in the aquifer and/or development causes excessive
removal of fines from aquifer sediments. Negative skin
effects may also result from well construction using an
artificial filter pack that is significantly coarser than aquifer
material. Positive skin effects can result from compaction,
invasion of drilling fluids, smearing of clays, etc.
Skin effects influence the flow distribution along the well. In
single-well pressure transient tests, wellbore damage has been
recognized (Dudgeon and Huyakorn, 1976; Chu et al, 1980;
Moench and Hsieh, 1985) as adversely affecting the tests.
Analysis of the test data can result in large errors in estimates
of storativity and transmissivity if skin effects and wellbore
storage are not properly delineated. There are solutions
available that account for both wellbore storage and
infinitesimally thin skin in the pumping well (Sandal et al..
1978; Chu et al., 1980) and both the pumping and observation
wells (Tongpenyai and Raghaven, 1981; Ogbe, 1984; Ogbe
and Brigham, 1984). The mathematical treatment used in
these models to account for the skin region, however, is only
an approximation and (he hydraulic head drop across the skin
is presumed to occur under steady flow conditions.
Well development can reduce positive skin effects by repairing
damage to the aquifer formations so that natural hydraulic
properties are partially restored. For production wells,
development comprises the systematic procedures followed to
ensure the maximum discharge rate at the highest specific
capacity with minimum production of particulate matter (Moss
and Moss, 1990). More specific construction and
development considerations are necessary for wells that are to
be used to obtain geohydrologic data for aquifer
characterization.
IV-2 Well
Ideally, wells designed for characterization using techniques
described in this document should be fully screened across the
interval of interest, such as the contaminant plume, be
constructed without an artificial filter pack, and be screened
such that the top of the screen is far enough below the water
table to allow placement of the pump intake in the well casing
19
-------
during the test. Boman et al. (1997) discuss the potential
influence of an artificial filter pack on flowmeter results. A
coarse-grained gravel or sand pack may result in significant
vertical flow through the pack prior to entering the screen.
This phenomena is enhanced by use of high flow rates that
result in resistance to flow through the downhole probe of the
electromagnetic flowmeter. In this situation, the measured
flow distribution is skewed with a high influx of water near
the top of the screen. For this reason, use of a borehole
flowmeter in wells constructed with a filter pack of gravel-
sized material may not provide meaningful data and generally
should be avoided. It is also possible to observe some effects
in wells with filter packs constructed of coarse-grained sand if
the hydraulic conductivity of the pack material is significantly
greater than aquifer materials. However, use of a natural
collapse well construction is not always feasible. For
example, fonnations with a significant fraction of fine-grained
materials or units may not be appropriate for natural collapse
construction and may result in void space within the annulus
that affects test results. Results of field applications of the
borehole flowmeter indicate that it is possible to obtain
representative flowmeter measurements in many wells
constructed using an artificial pack. In general, wells
specifically constructed for these tests should avoid use of
artificial filter packs, where possible. If an artificial pack is
used, the data should be carefully examined for evidence of
bias in the flow distribution.
Wells that are screened across the water table may be subject
to increased head loss near the pump intake. This is
particularly true for wells constructed with an artificial filter
pack that is significantly more conductive than surrounding
aquifer materials (Boman et al., 1997). This situation may
result in a nonuniform head distribution in the well and an
increased influx of water near the pump intake. Such a
situation would also bias flowmeter test results.
Presence of a low permeability skin (i.e., positive skin effect)
adjacent to the well may also significantly bias flowmeter
results by affecting the distribution of flow to the well.
Studies by Ruud and Kabala (1997) indicate that the presence
of a low permeability skin adjacent to the well may have a
much greater effect than (lie presence of a /one of increased
permeability. As previously noted, the bias in the flowmeter
measurements increases with increasing thickness of the
disturbed zone and increasing difference between the
permeability of the disturbed zone and the formation
materials.
These studies imply that wells should be designed and
constructed to minimize the size of disturbed zones and the
degree of disturbance. In general, formation of a higher
permeability disturbed zone would be preferable to the
presence of a lower permeability zone. The results also imply
that the minimum size of measurement intervals deserves
careful consideration in test design. Potential effects of well
design and construction on the representativeness of the data
obtained from these investigations should be carefully
considered prior to well installation. Many of these questions
are still areas for continuing research.
IV-3 Well Installation Methods
a. Overview
Drilling methods chosen for construction of test wells should
be those techniques that result in the least possible disturbance
to the formation surrounding the well and result in a minimum
annular space between the well screen and the borehole wall.
The most appropriate method will depend on site-specific
conditions. A number of references have appeared during the
past decade regarding well drilling techniques, technologies.
and related subject matter (e.g., Driscoll, 1986; Aller et al.,
1989; Harlan ct al., 1989; and Moss and Moss, 1990).
b. Drilling Techniques and Hydraulic
Conductivity Estimates
There are many possible interactions between the various
drilling methods and the types of subsurface media at a
particular site. Optimum methods depend on site conditions.
In addition, the various studies that have dealt with the
interaction of well drilling and hydrogeologic measurements
have focused more on chemical measurements than on
hydraulic measurements. Prominent studies during the past
decade include Minning (1982). Barcelona and Heffrich
(1986), Hackett (1987, 1988), Keely andBoateng (1987a,b),
Paul et al. (1988), and Strauss et al. (1989).
Keely and Boateng (1987a,b) present an illuminating
discussion of various options and trade-offs when constructing
monitoring wells. Much of the following discussion is based
on those two references. Mud rotary has been a relatively
common drilling technique for water supply wells because it is
rapid, economical, and essential to the performance of certain
geophysical logs. However, the technique has the
disadvantage that the potential exists for large volumes of
drilling fluids to enter the formation resulting in formation of a
low permeability skin. Subsequent development sufficient to
remove this skin effect is usually difficult or impossible.
Improvements to this technology include driving a temporary
casing while drilling or use of dual-wall reverse circulation
drilling. In the dual-wall method, mud travels down a
cylindrical annulus that is bounded by the outer drill pipe and
an internal, rotating drill stem connected to the bit. Mud is
ejected just above the bit, picks up cuttings, and is pumped up
the inner drill stem to the surface. This technique limits
exposure of the borehole wall to drilling fluids and decreases
mud invasion into the formation.
Many of the problems associated with introduction of drilling
fluids using mud rotary techniques arc also common to air
rotary methods. These problems can be minimized by driving
a temporary casing flush with the borehole wall. In fact, this
is often necessary to prevent caving of the borehole walls
when drilling in unconsolidated materials. However,
compressed air can still enter the formation, and "air binding"
is a well-known phenomenon that decreases the apparent
hydraulic conductivity of porous materials. Just as with mud
20
-------
invasion, air invasion would be expected to be particle-size
dependent.
According to Keely and Boateng (1987a), the cable tool
method appears to have several advantages when used to
construct monitoring wells, although it is rarely used for this
purpose. In cable tool drilling, no drilling fluids are necessary.
A temporary casing is driven as the drilling proceeds. Water
may be added to the hole to aid bailing, but it is not under
pressure. Driving and removing the temporary casing may not
disturb the borehole wall significantly because there is a sharp
drive shoe at the bottom, the casing is smooth, the annular
spacing between the casing and the borehole wall is
sufficiently small that material is not dragged along with the
moving casing, and the casing is slowly moved up and down.
This method may produce a well that is exceptionally well-
suited for hydraulic conductivity and other measurements.
The hollow-stem augering method is relatively fast, does not
involve addition of drilling fluid, and formation samples may
be obtained easily during the course of drilling. A major
disadvantage, however, especially from the viewpoint of
estimating hydraulic conductivity, is that the augering process
causes wet clay and silt material to be smeared along the
borehole wall. In certain types of soils such as clayey
saprolites, the smearing and destruction of the open pore
structure of the soil can be so severe that the resulting well is
not suitable for hydraulic conductivity tests. Another potential
disadvantage is that the annular space between a well screen
installed through the hollow-stem auger and the borehole wall
can be relatively large. Thus, if one docs not wish to install a
sand pack, there would be a relatively large volume into which
the surrounding formation would collapse, further disrupting
the structure of the natural formation. If collapse of cohesive
soils was not complete, channels for vertical water movement
would be formed.
Much of the available literature contains information
concerning well construction and measurements of various
types in a variety of subsurface environments. This
information is usually in the form of case histories, descrip-
tions of problems that arose on particular projects, and the
manner in which these problems were addressed. There have
been relatively few investigations performed to study the
effects of well construction on hydraulic conductivity and
other measurements in a systematic fashion. One exception to
this is the work of Morin et al. (1988b). In this study,
epithermal neutron and natural gamma logs were used to
measure the formation disturbance caused by three different
well drilling techniques: hollow-stem augering; mud rotary;
and hammer-driven, flush-jointed, temporary casing within
which a permanent casing and screen were installed. Each
type of construction was utilized for a number of wells so that
a statistically significant set of logs could be obtained for each
method.
The study was conducted in a glacial outwash plane deposited
in a braided stream environment. The study aquifer was
composed of coarse sand and gravel in horizontal lenses and
layers with silt and clay comprising less than five percent of
the formation. This resulted in significant vertically-
distributed heterogeneity7 that was reflected in the geophysical
logs. The basic concept of the study was to associate an
increase in formation disturbance due to well construction
with a decrease in the degree of heterogeneity (more
homogeneous appearance) observed in the well logs. Using
this criterion, the order of increasing disturbance was: driven
casing, mud rotary, and augering.
IV-4 Well Development Methods
a. Overview
Well development includes a broad spectrum of techniques,
procedures, and tools for applying some form of energy to the
well screen and adjacent formation. Historically, several
development methods that have been used following well
installation include: overpumping, backwashing, mechanical
surging, air development, and high-velocity jetting. Well
development methods such as chemical treatment, hydraulic
fracturing, and use of explosives to maximize the water yield
in production wells are not applicable to these studies.
Discussion in this text is restricted to aquifer characterization
wells and development methods that assist in restoring the
undisturbed hydraulic characteristics of aquifer formations.
Determining which development methods are most
appropriate for any given well requires an understanding of
available methods, a knowledge of the aquifer formations
present at the well location, and particulars of the well design.
In some cases, a single method might be effective for
development. However, in mam instances, the use of more
than one development method will yield much better results.
Driscoll (1986), Moss and Moss (1990), and Harlan et al.
(1989) explain the mechanics of the different well
development methods.
b. Development Methods and Hydraulic
Conductivity Estimates
The literature offers little guidance on the selection of an
optimum method of well development for any particular site.
Much of the former work related to well development has
concentrated on maximizing water yield in production wells.
In general, these types of studies (National Ground Water
Association, 1989) provide evaluations of different well
drilling and development techniques based upon specific
capacity measurements. This is a measure of the reduction in
positive skin effects produced by well development.
Although the influence of skin effects on pumping tests has
long been recognized in the petroleum industry (e.g., van
Everdingen and Hurst, 1949; Hawkins, 1956) and by ground-
water scientists (e.g., Mocnch and Hsieh, 1985), very few
studies provide comparisons of different development methods
using explicit hydrogeologic measurement techniques.
Rehfeldt et al. (1989b) provide results from experiments that
were undertaken using an impeller flowmeter to investigate
well installation and development methods on flowmeter
21
-------
discharge profiles. The test site was a heterogeneous sand and
gravel aquifer located at the MacroDispersion Experiment
(MADE) site on the Columbus Air Force Base, Columbus,
Mississippi. One of the major objectives of the study was to
determine the most appropriate well installation and
development method that would minimally disturb the aquifer
and allow accurate estimates of hydraulic conductivity at the
test site. Wells evaluated in the experiment included a driven
steel well screen, drivc-and-wash wells using compressed air,
and hollow-stem auger wells, constructed with both natural
and artificial filter packs. The wells were developed in three
stages; first by cyclic overpumping and backwashing, then
with a surge block attached to the drill rig, and finally using a
hand-drawn swab. Hydraulic conductivity profiles were
determined from impeller flowmeter measurements after each
stage of development. When hydraulic conductivity estimates
were reproducible, well development was considered
complete.
The study indicated that a large diameter (30 cm) hollow-stem
auger method of well installation was the least satisfactory of
the test methods because of the degree of disturbance
associated with auger drilling and the fact that more annular
space exists between the wellbore and screen. Although the
driven steel well screen was installed to represent a well with
no open annular space, it was substantially more expensive
than other methods because of material costs. The internal
vertical rods that support the steel screen also presented
problems with well development since the surge block and
swabbing tool could not develop a good seal along the inside
of the screen. Rehfeldt et al. (1989b) provide the following
observations based on their testing.
1. Well development caused changes in the aquifer
material adjacent to the wellbore and altered the
hydraulic conductivity profiles.
2. Overpumping and backwashing development
produced changes in the profiles of both the augcrcd
well and the well installed using drive-and-wash
methods. Order-of-magnitude increases in hydraulic
conductivity were observed at a minimum of five
locations for the angered well, but were never that
large for the drive-and-wash well. Additionally, the
trends in the profiles after overpumping and
backwashing development were generally similar in
the drive-and-wash well but not in the angered well.
3. After swabbing, the augered well again showed
significant differences in the profiles, although less
than for the first two development cycles. The profile
for the drive-and-wash well after swabbing was
relatively consistent.
4. The profile for the driven steel screen generally had
lower values than the profile for both the augered and
drive-and-wash wells. The lower hydraulic
conductivity values may have been partly caused by
compressed aquifer material near the well, the
adherence of fine-grained material to the well screen.
or incomplete development of the well.
22
-------
Chapter V
Field of Well Construction
Development at Columbus, Mississippi
V-l Description of Test Site
a. Site Location
The test site occupies one of twenty-five hectares (Ha) in the
northeastern comer of Columbus Air Force Base (CAFB),
Columbus, Mississippi, leased from the U.S. Air Force for the
MacroDispersion Experiment (MADE). The site is
approximately 6 km cast of the Tombigbce River and 2.5 km
south of the Buttahatchee River, and lies above the 100-yr
floodplain of both rivers. The site is situated on the youngest
terrace deposits associated with the Buttahatchee River.
b. Aquifer Characteristics
Young (1991b) describes the Columbus Aquifer as being
composed of approximately 10 m of Pleistocene and Holocene
fluvial deposits. The aquifer overlies the Eutaw formation that
consists primarily of marine clay and serves as an aquitard. In
addition to numerous samples from boreholes, a geological
investigation of the aquifer included mapped geological facies
at a gravel pit and an aerial photography survey that indicated
an abandoned river meander crossing the northern region of
the 1 -Ha test site and passing through the middle of the
MADE test site (Figure V-l).
The aquifer at the site is composed of poorly-sorted to well-
sorted sandy gravel and gravelly sand with minor amounts of
silt and clay. Sediments are generally unconsolidated and
cohcsionlcss below the water table. The upper portion of the
aquifer is generally composed of coarse-grained sediments and
bar deposits from a meandering river system (Kaye. 1955;
Rehfeldt et al, 1989b). The abrupt changes in vertical
sequences and coarse texture suggest that the pointbar
sediments were deposited during catastrophic flooding events.
Chute bars and channels with clay drapes were also deposited
during flood stage. Such events result in an uneven sand
distribution and a seemingly chaotic occurrence of gravel
lenses and clay drapes (Collinson and Thompson, 1989).
The lower portion of the aquifer is composed of sediments
from a braided river system. This model implies an irregular
pattern of coarse gravelly lenses deposited as longitudinal and
transverse bars at high flow stage, alternating laterally and
vertically with finer sand and silt deposited in channels at low-
flow stage. As these depositional bodies were formed by
short-lived branding and rejoining shallow channels, and
subsequent partial truncation by secondary stream channels,
Figure V-1. Ox bow meander at the Columbus AFB site drawn
from a 1956 aerial photograph.
a large range of shapes and sizes exist for the depositional
bodies. The shapes might best be described as irregular
tongues, shoestrings, wedges, and pods.
c. Previous Pumping Tests
A series of single-well pumping tests, slug tests, and
electromagnetic borehole flowmeter tests were conducted at
the original thirty-seven wells located across the 1-Ha test site.
The results of these tests indicate that a zone of interconnected
high-K deposits exists at elevations of 59 m to 62 in above
mean sea level (AMSL) within boundaries mapped by the
former meander shown in Figure V-l. These results also
indicated that positive (low-K) skin effects exist at most, if not
all, of the wells. Young (1991a,b) provides an analysis of the
pumping test data using the Cooper-Jacob equation and the
Cooper-Jacob Straight-Line (CJSL) method (Cooper and
Jacob, 1946). The analysis indicates that the two approaches
provide significantly different transmissivity values.
Considering that poorly-sorted coarse-grain sediments often
lie adjacent to well-sorted fine-grain sediments, and that
highly permeable lenses can dominate ground-water flow at
Columbus, significant skin effects were considered probable.
The following are potential causes of positive skin effects:
23
-------
(1) smearing of silt and clay particles into and across high-
permeability zones; (2) compaction of aquifer material by
methods that include advancement of a protective casing; and
(3) alignment of blank (i.e.. nonslotted) sections of well casing
with aquifer zones of high permeability.
(L Monitoring Well Installation
For the purposes of this study, two clusters of three wells each
were added to an existing network of thirty-seven wells at the
1-Ha test site (Figure V-2). The drilling and development
methods used to install the first thirty-seven wells at the 1-Ha
test site are described by Young (199 la). The six new wells
are numbered 38 through 43. The well clusters were
configured as equilateral triangles with individual wells being
separated by distances of two meters so that aquifer
conditions near each well would be similar. The two well
clusters are separated by a distance of about 60 m. Well
cluster 38-39-40 resides on the southern side of the 1-Ha test
site within possible pointbar deposits in the upper aquifer and
possible braided river deposits in the lower aquifer. Well
cluster 41-42-43 is located in the northern portion of the test
site, within possible channel sediments from a meandering
river system and with possible channel sediments at depth
from a braided river system.
Each well is approximately llm deep and consists of 5.1-cm
(inside diameter), schedule-40 PVC casing and screen (Figure
V-3). The screen for each well is 9.14 m in length and
machine-slotted with 0.25-mm slots spaced at 3.18-mm
intervals. The following three drilling methods were used for
wells at each well cluster:
1. 19.4-cm hollow-stem auger - natural filter pack
2. 27.0-cm hollow-stem auger - artificial filter pack
3. 11.4-cm drive-and-wash - natural filter pack
Wells 39 and 43 were installed using a 19.4-cm (outside
diameter) hollow-stem auger. The aquifer material was
allowed to collapse around the screen and casing, and auger
cuttings were used to backfill approximately the top two
meters of borehole. Wells 38 and 41 were installed using a 27-
cm (outside diameter) hollow-stem auger with artificial filter
packs. The filter pack was comprised of 1.6 mm to 3.2 mm,
subangular to angular, quartz sand. A disadvantage of the
larger auger is the size of the annulus created.
Wells 40 and 42 were installed using a modified rotary wash
method. The method involved driving an 11.4-cm steel outer-
casing ahead of the rotary bit and then washing the cuttings
out of the casing with water. The well screen and casing were
lowered into place, and the outer steel casing was removed
allowing formation material to collapse around the well. The
modification minimizes both the wash into undisturbed aquifer
sediments and the amount of formation material removed.
In unconsolidated aquifer materials, the driving of casing
generally creates a region of compacted aquifer material
around the casing. Pulling the outside casing and allowing
aquifer material to collapse on the inner well screen and
casing may relieve this compression. Based upon studies by
110
100
80
70
B 60
CD
50
40
30
20
10
37
*
27
41
6
*
26
*
1416
31
*
23
9
*
36
20 11
* **15
30
*
17 21
* *
4
*
32
*
7
*
33
*
V'
,--' 38
29
*
22*
35*
10 20 30 40 50 60 70 80 90 100 110
Meters
Figure V-2. Well network at the 1-Ha test site.
24
-------
Cap *[
Borehole by
15.9-cm ID,
27-cm OD -
Hollow Stem
Auger
5.1-cm
Schedule-40
PVC Slotted
Screen, -
.25mm Slots
Artificial
Filter
Pack
V/f
>
>
->
=
=
E
E
-
~
E
E
E
E
~
E
E
-i Non-Shrink
./ Grout /
(£
^
^,
t
o
Solid Casing W
>
i
Threaded c
. . 0
Joints 0
i
Bottom Cap
Cap J
MI
Borehole by
10.8-cm ID,
19.4-cmOD *
Hollow Stem
Auger
5.1-cm
Schedule-40
PVC Slotted
Screen, >
.25mm Slots
!!
-
E
E
-
~
E
E
E
E
~
E
E
L.
-, Non-Shrink
f Grout
1
* Solid Casing
i
i
Threaded £=
* Joints £
Bottom Cap
Cap J
sss
III
f
t
Borehole by
11.4-cmOD
Casing and >
Rotary Wash
5.1-cm
Schedule-40
PVC Slotted
Screen, >
.25mm Slots
&
~
E
£_
-
~
E
E
~
~
E
*_
n Non-Shrink
J^ Grout
* Solid Casing
Threaded
* Joints
Bottom Cap
(3) (b) (C) (NOT TO SCALE)
WELLS 38 & 41 WELLS 39 & 43 WELLS 40 & 42
ARTIFICIAL FILTER PACK ROTARY WASH
Figure V-3. Design of wells used to evaluate the effect of well development on flowmeter tests.
Relifeldi et al. (1989a) at CAFB, this was expected to be the
preferred method of installation as the extent of aquifer
disturbance is much less than in auger drilling. The annular
space surrounding the well into which the aquifer material
must collapse is also effectively reduced from about 6.7 cm
for small (19.4-cm) diameter auger wells to 2.7 cm for the
driven wells. Hence, the in-situ hydraulic conductivity
estimated from a well installed by drive-and-wash methods is
more likely to be representative of the true aquifer hydraulic
conductivity.
V-2 Test Descriptions
This study consisted of a series of single-well pumping tests
conducted at six wells. Between pumping tests, additional
well development was performed. During each pumping test,
drawdown and electromagnetic -flowmeter data were collected.
Testing at each of the two well clusters was conducted
concurrently using two teams outfitted with flowmcters,
pressure transducers, data loggers, lap-top computers, and
other equipment required for monitoring and testing. The
flowmeters were calibrated against known flows prior to, and
after, their use in the field. Both 1.27-cm and 2.54-cm (orifice
diameters) electromagnetic flowmeters were used during
testing.
The procedure for sequential development and testing of the
monitoring wells consisted of the following steps:
1. Ambient flowmeter tests were conducted at each
well.
2. Pumping tests and flowmeter tests were
performed.
3. Wells were developed according to prescribed
methods.
4. Retesting and development sequence was completed
for three development cycles.
5. Injection tests and high-rate pumping tests were
performed.
Ambient flowmeter tests were conducted to measure natural
(background) ground-water flow within each well. The type
of well installed, the aquifer formations present, and the
degree of formation damage from drilling have differing
effects on ambient flows. Ambient measurements were taken
at 30.5-cm increments beginning at the bottom of the well and
advancing upwards until the water table was reached. The
ambient measurements were taken after the test wells had
sufficient time to return to a quiescent state from development
activities and pumping tests, which was usually overnight.
Single-well pumping tests and borehole flowmeter tests were
completed at each well following ambient flowmeter testing.
Discharge measurements were made several times during each
test. The flowmeters were used to obtain a profile of the flow-
distribution after a steady flow field was achieved in the
vicinity of the well, which generally occurred after 20 min to
-------
30 niin of pumping. Flowmeter readings were obtained at
30.5-cm increments as in ambient testing. Flow distribution
profiles were compared in the field after progressive well
development. Test wells were allowed to fully recover before
proceeding with each stage of development, usually overnight
A total of six pumping and flowmeter tests were completed for
each well (Table V-l). The first four tests were conducted
prior to and after each development cycle. The last two tests
were injection and high-rate pumping tests. The first method
used for well development was overpumping and
backwashing. Initially, the well was pumped until the water
began to clear. Development was then completed by pumping
water through a discharge hose several feet above the top of
the well, and then allowing it to flow back through the pump
and out through the well screen. Water was occasionally
discharged to remove the fine-grained materials. Several
pumping cycles were completed until the discharge was clear.
The second development consisted of a modified swabbing
method that was alternated with overpumping and
backwashing. This method is similar to the procedure used
by Rehfeldt et al. (1989b). A section of galvanized pipe was
added to weight the swab and increase its rate of fall. The
swab and galvanized pipe had a combined weight of over 2.9
kg. Beginning at the bottom of the well, the swab was
manually raised with force over a 1-meter increment of the
well screen and then allowed to fall back to its original
position to provide a surging action. This swabbing
movement was repeated for a total of four repetitions, until
the water table was reached. Water was then pumped to
remove fines liberated during development and overpumping/
backwashing was conducted. The entire cycle was repeated
using three swabbing repetitions. The third and final
development was a reiteration of the alternating modified
swabbing and overpump/backwash method that was identical
to that used in the second development.
V-3 Test Analyses and Results
a. Ambient Flow Distributions
Ambient flow measurements collected after each successive
phase of well development (Figure V-4) varied from -0.34 L/
min to 0.04 L/min for well cluster 38-39-40, and from -0.06 LI
min to 0.29 L/min for well cluster 41-42-43. The sign
convention is positive for upward flow and negative for
downward flow. The initial development, if properly
conducted, should displace extraneous fines from the
wellborc. From observation of the flowmeter plots, it is
apparent that the initial development (overpumping and
backwashing in this case) is important. A relatively large
change in ambient flow is displayed after the first
development by all wells except wells 42 and 43. The
magnitude of the changes in ambient flows decrease with
ensuing stages of development. Following the first stage of
development, the changes that occur in the ambient flows also
vary according to well type and differences in aquifer
materials. The modified rotary wash wells (40 and 42) exhibit
ambient flow profiles that were essentially reproducible after
llie first stage of development.
The ambient flow profiles for the wells constructed with an
artificial filter pack (38 and 41) show differing responses after
die second and third development (Figure V-4). The ambient
Table V-1. Testing Sequence
Test
No.
Test Type
Development Cycle
Development Method
Ambient Borehole Flowmeter (BHFM) Tests
Constant Discharge Pumping Tests
BHFM Tests During Pumping
Ambient BHFM Tests
Constant Discharge Pumping Tests
BHFM Tests During Pumping
Ambient BHFM Tests
Constant Discharge Pumping Tests
BHFM Tests During Pumping
Ambient BHFM Tests
Constant Discharge Pumping Tests
BHFM Tests During Pumping
Ambient BHFM Tests
Constant-Rate Injection Tests
BHFM Tests During Injection
Ambient BHFM Tests
Constant Discharge Pumping Tests
BHFM Tests During High-Rate Pumping
Pre-development
None
After 1st development Overpump/Backwash (15 minutes)
After 2nd development
After 3rd development
Alternating Modified Swabbing and
Overpump/Backwash (60 minutes)
Alternating Modified Swabbing and
Overpump/Backwash (60 minutes)
Development complete None
Development complete None
26
-------
WELL 38 -ARTIFICIAL FILTER PACK
WELL 39 - AUGER HOLE
WELL 40 - ROTARY WASH
^ 60
E
E
ill
Qj
58
56
54
52
downward flow
-0.4 -0.3 -0.2 -0.1 0 0.1 -0.4 -0.3 -0.2 -0.1 0 0.1 -0.4 -0.3 -0.2 -0.1 0 0.1
WELL 41-ARTIFICIAL FILTER PACK
WELL 43-AUGER HOLE
WELL 42 - ROTARY WASH
DZ
^ 60
E
- 58
g
^ 56
111
111
54
52
-0
r
- /
/
:v
^_
«.' T^-*1^"
__j SN;\
^J
upward flow ^
.100
FLC
i i i i
.1 0.2 0.3 0
4 -0
- \ i
- Jx
"^ V s
i U
i i i i i i i i
.1 0 0.1 0.2 0.3
3W (L/min) FLOW (L/min)
PRE-DEVELOPMENT
AFTER 2nd DEVELOPMENT
AFTER 1st DEVELOPMENT
AFTER 3rd DEVELOPMENT
I
1
0
.4 -0
- i 1
r ; '-
(\ -
/ I
i i i i i i i i i
.1 0 0.1 0.2 0.3 0.
FLOW (L/min)
§ -^ WELL SCREEN
I -^ SOLID CASING
P AT JOINT
Figure V-4. Ambient flow profiles after successive well development.
-------
flows at well 41 increase incrementally with diminishing
differences between development cycles. Well 38. on the
other hand, exhibits some irregularities in ambient flow
profiles. Below about 56 in. the first development of well 38
produces the highest ambient flows with succeeding
development resulting in reductions of ambient flow. Above
about 56 m, the second development of well 38 produces the
highest ambient flows with ambient flows being reduced
following the third development. These responses may simply
be due to stabilization of the filter pack or to effects of minor
temporal changes in local hydrology.
The angered wells with natural filter packs (39 and 43) display
ambient flow profiles that are quite variable (Figure V-4). For
both wells, the ambient flows measured following the second
stage of development are greater than those of the third stage of
development. The two wells also show the largest differences in
their ambient profiles between the second and third development
cycles. These results suggest that the material around the
angered wells was the most sensitive to well development.
Several shifts and peaks in the ambient flow profiles were
observed through the first two stages of well development for
both well clusters. These phenomena are attributed to
inadequate collapse and stabilization of aquifer material
around the screens for the auger hole wells, and to incomplete
grading and stabilization of aquifer and filter pack materials
for the wells with artificial filter packs. Very little deviation in
the ambient flow profiles are shown by the modified rotary
wash wells after initial development. By the third stage of
development, the shifts and peaks in all the ambient flow
profiles had decreased to the point that wellbore material was
stabilized and development was deemed complete.
Well cluster 38-39-40 shows good correlation between wells
for (lie final profiles of ambient flow. The profiles indicate
that most of the water is entering between about 57 m and 58
m AMSL and leaving between 54 m and 55 m AMSL, with
little ambient flow above 58 m AMSL. These two high
ambient flow zones suggest that sequences of high hydraulic
conductivity sediments may intersect the wells at these
horizons. For well cluster 41-42-43, there is less similarity
among the ambient flow plots. All three wells have upward
ambient flow but there is considerable variability in the source
and magnitude of the flow.
b. Induced Flow Distributions
The flow distributions measured under pumping conditions
(Figure V-5) represent the results from flowmeter testing during
steady-state flow conditions. The mean flow rates used for
pumping tests were 8.0 L/min and 13.4 L/min for well clusters
38-39-40 and 41-42-43, respectively. The pumping rates used
for the tests were fairly constant and had variances of only 0.015
L/min and 0.076 L/min for well clusters 38-39-40 and 41-42-43,
respectively. The cumulative flows were adjusted by taking the
net difference between the measured flows under pumping and
ambient conditions, and then normalizing the result to the
pumping rate.
The induced flow distributions for wells 41 and 43 show
changes after the initial development that are relatively large
in comparison to succeeding development cycles. The
magnitude of the changes in flow decreases incrementally
with ensuing stages of development for the well with the
artificial filter pack (well 41). Observation of the induced
flow distributions for well cluster 38-39-40 does not indicate a
large change in the profiles of wells 38 and 40 between pre-
and initial development. The greatest change in the flow
profiles for these two wells occurs after the second
development, indicating that the modified swabbing method
was necessary to effectively develop the wells. The induced
flow profiles for well 39 indicate that it is experiencing a
moderate degree of formation stabilization during later stages
of development.
c. Specific Capacity Values
Drawdown data, (Figure V-6) were used to calculate specific
capacity (Figure V-7) for each well based on the drawdown
responses at 1500 s after initiation of extraction. In general,
the greatest increases in specific capacity occur after the initial
development. Except for well 41, specific capacity values
stabilized after the second development. Well 41 has an
artificial filter pack and is located in a region where highly
transmissive aquifer zones exist near the top of the well
screen. The different response at well 41 may be caused by
spatial variability within the aquifer and the possible
intersection of a highly permeable lens that is not
interconnected with the neighboring wells.
(I Transmissivity Values
The single-well test drawdown curves (Figure V-6) were
analyzed using the Cooper-Jacob straight line (CJSL) method
(Cooper and Jacob, 1946) to calculate transmissivity. The
semilog plots of drawdown data are characterized by an early
slope that is about five to twenty-five times steeper than the
late slope (after 100 s). For comparison purposes,
transmissivities were calculated for both early and late times.
Typically, the late-time portion of the curve used for the CJSL
analysis began between 100 s and 300 s and ended between
900 s and 1,000 s. Late-time average transmissivities were
calculated using the data from all four pumping tests at each
cluster. The average transmissivities for well clusters 38-39-
40 and 41-42-43 are 1.9.8 cm2/s and 34.0 cm2/s, respectively.
At evety well, the ratio of the transmissivity after each well
development to the average transmissivity were within a factor
of two. The results indicate increases and decreases in
transmissivity with additional well development. However, no
consistent trends are evident in the data. It is possible that the
changes in the calculated transmissivity values were caused
more by the differences in the testing and analyses procedures
than the aquifer conditions. In order to minimize recovery
time between pumping tests, low pumping rates were used.
As a result, trends in the drawdown curves are difficult to
define because the aquifer was not adequately stressed.
The CJSL analysis for the early-time portion of the semilog
drawdown curves typically began near 20 s and ended at about
28
-------
WELL 38 - ARTIFICIAL FILTER PACK
62
WELL 39 - AUGER HOLE
WELL 40 - ROTARY WASH
E
LU
_l
LU
60
58
56
54
WELL 41 - ARTIFICIAL FILTER PACK
62
W 60
E
^T 58
o
I-
LU
LU
54
52
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
NORMALIZED FLOW
i i i i i
WELL 43 - AUGER HOLE
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
NORMALIZED FLOW
WELL 42 - ROTARY WASH
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
NORMALIZED FLOW
PRE-DEVELOPMENT
AFTER 2nd DEVELOPMENT
AFTER 1st DEVELOPMENT
AFTER 3rd DEVELOPMENT
WELL SCREEN
SOLID CASING
AT JOINT
Figure V-5. Induced flow profiles after successive well development.
-------
WELL 38 - ARTIFICIAL FILTER PACK
PRE-DEVELOPMENT
AFTER FIRST
DEVELOPMENT
Q = 7.8 L/min//
AFTER SECOND
DEVELOPMENT
Q = 7.8 L/min //
AFTER THIRD
DEVELOPMENT
Q = 8.3 L/min
12 1020 100200 1,000 12 1020 100200 1,000 12 1020 100200 1,000 12 1020 100200 1,000
TIME (seconds) TIME (seconds) TIME (seconds) TIME (seconds)
PUMPED WELL
WELL 38
OBSERVATION WELL
WELL 39 WELL 40
WELL 39 - AUGER HOLE
PRE-
DEVELOPMEb
AFTER FIRST
DEVELOPMENT
Q = 7.8 L/min
AFTER SECOND
DEVELOPMENT
Q = 7.7 L/min
AFTER THIRD
DEVELOPMENT
1 2 1020 100200 1,000 1 2 1020 100200 1,000 1 2 1020 100200 1,000 1 2 1020 100200 1,000
TIME (seconds) TIME (seconds) TIME (seconds) TIME (seconds)
PUMPED WELL OBSERVATION WELL
WELL 39 WELL 38 WELL 40
WELL 40 - ROTARY WASH
PRE-DEVELOPMENT
Q = 8.1 L/min//
QL
Q
AFTER FIRST
DEVELOPMENT
,Q = 7.8 L/min//
AFTER SECOND
DEVELOPMENT
Q = 7.8 L/min
AFTER THIRD
DEVELOPMENT
Q = 8.3 L/min
1 2 1020 100200 1,000 1 2 1020 100200 1,000 1 2 1020 100200 1,000 1 2 1020 100200 1,000
TIME (seconds) TIME (seconds) TIME (seconds) TIME (seconds)
PUMPED WELL
WELL 40
OBSERVATION WELL
WELL 38 WELL 39
Figure V-6. Effect of well development on the drawdown values for pumping tests at the 38-39-40 well cluster.
30
-------
CM
co
o
o
LO
5.5
5
4.5
4
3.5
?
t » *
I *
e
Well 38-Arti
* Well 39-Auc
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Well 40-Rot
ficial Filter Pack
er Hole
ary Wash
ii
&
ro
Q.
ro
O
M
'o
Q.
CO
14
19
10
Q
4
o
D
O Well 41 -Artificial Filter Pack
a
A Well 42-Rotary Wash
O Well 43-Auger Hole
n A A
4 6 °
&
Pumping Test Number
Figure V-7. Comparison of specific capacities at different wells.
80 s. Values of early-time transmissivity (Figure V-8, Table
V-2) only increased with successive well development. These
increases are interpreted as reductions in positive skin effects.
There are no similarities evident for early-time transmissivity
increases among wells of a particular construction type.
However, well cluster 38-39-40 exhibits increases in early-
time transmissivity values that are up to an order-of-magnitude
less than estimates for well cluster 41-42-43. This appears to
be a reflection of the low transmissivity region in which
cluster 38-39-40 resides. The geometric means for early-time
transmissivity values were 1.6 cm2/s and 3.2 cm2/s for well
clusters 38-39-40 and 41-42-43. respectively. These averages
are approximately ten times lower than the transmissivity
values measured at late times, and are in good agreement with
that of 3.6 cm2/s calculated by Young (1991a,b) for early-time
slopes from fifty-seven single-well pumping tests and thirty-
one slug tests at the 1-Ha test site.
V-4. Summary
A series of flow-meter tests were conducted at Columbus Air
Force Base to investigate the effects of well construction and
development methods on flow distributions and drawdown
responses in a highly heterogeneous aquifer. Two well
clusters were installed 60 m apart. Each well cluster included
three 5.1-cm diameter PVC wells installed using different
drilling techniques. The three installation techniques included
a modified rotary wash method, a 19.4-cm hollow-stem auger
method, and a 27.0-cm hollow-stem auger method using an
artificial filter pack in well construction. Differences in local-
scale geohydrology are immediately apparent from
observation of the ambient flow profiles. At one well cluster,
all three wells exhibit only downward ambient flow. At the
other well cluster, upward ambient flow was measured.
The test results show that well development has a significant
impact on increasing the magnitude of ambient flows. At
several vertical zones, the directions of ambient flow changed
because of well development. The artificial filter packed and
rotary wash wells displayed the greatest changes in ambient
flow profiles following initial well development by
overpumping and backwashing. The largest changes for the
augered wells with a natural filter pack occurred after the
second well development, which included swabbing with a
surge block accompanied by overpumping and backwashing.
Among the ambient flow profiles, those from the rotary wash
wells had the least sensitivity- to well development beyond the
initial development. At the two rotary wash wells, the ambient
flow profiles stabilized after the initial well development. At
the "19.4-cm augered" wells and one of the augered wells with
an artificial filter pack, the ambient profile patterns remained
similar after the initial well development, but the magnitude of
the flow rates differed. The data clearly indicate that some
type of well development is necessary to obtain useful
flowmeter data for ambient conditions, and suggests that the
amount and method of well development depends on the well
type and aquifer properties.
31
-------
CM
O
ro
LU
0
1 8
1 fi
1 A
1 9
1
ID
fi
0
n
A.
A.
A.
o
A ^
O i
o u n
A
A O
o
A
Well 38 Artificial
Filter Pack
A Well 39-Auger Hole
Well 40-Rotary Wash
D
O
A
Well 41 -Artificial Filter Pack
Well 42-Rotary Wash
Well 43-Auger Hole
Pumping Test Number
Figure V-8. Calculated transmissivities using early time data.
Table V-2. Early-Time CJSL Transmissivity Values
Well No.
38
39
40
41
42
43
initial T
(cm2/s)
1.2
1.7
1.6
3.8
1.4
0.9
Final T
(cm2/s)
1.7
2.1
1.9
8.8
4.9
5.6
Percent
Increase
42
25
18
129
242
494
The flowmeter results for pumping conditions produced
conclusions similar to those obtained for the ambient flow
conditions. However, two differences were observed. The
first difference was a convergence to a stabilized flow profile
for all of the well types. The second was the effect of the well
cluster locations on the relative differences among the flow
profiles. One well cluster displayed the greatest flow profile
changes between the "prc-development" and "1st
development." The other well cluster showed the greatest
differences between profiles for "1st development" and the
"2nd development."
At several elevations in the "pre-development" profiles, a
decrease in net differential flow occurred with increasing
elevation. For this decrease to occur, vertical flow established
within the well must temporarily exit the well. This might
occur near regions where voids existed in the well annulus
and/or the flowmeter caused significant blockage of incoming
flow. The flow bypass problem can be lessened by creating a
more uniform packing in the well annulus. In subsequent tests
after well development, the frequency and magnitude of the
flow bypass problems were greatly reduced.
For each flowmeter test, the drawdown data in the pumped
well were analyzed using the CJSL method to estimate
transmissivity. Typically, the drawdown response was
characterized by a linear slope at early times (<50 s) that was
about five to twenty-five times steeper than the late slope
(after 100 s). The early slope values arc reflective of the
disturbed aquifer material near the well. The late-time values
better represent undisturbed aquifer material at radial distances
further away from the well. The CJSL analysis shows that the
transmissivity values calculated at early times only increased
with additional well development. Although positive skin
effects were not eliminated, these results indicate that the
effects were significantly reduced.
32
-------
Chapter VI
Field of Well Construction
Development at Mobile, Alabama
VI-1 Description of Test Site
a. Site Location
The well field is located at the Bany Steam Plant, which is
owned and operated by the Alabama Power Company.
Geographically, it is approximately 32 km north of Mobile,
Alabama.
b. Aquifer Characteristics
The surface is composed of a low terrace deposit of
Quaternary age consisting of interbedded sands and clays that
have, in geologic time, been recently deposited along the
western edge of the Mobile River. The sand and clay deposits
(Figure VI-l) extend to a depth of 61 m where the contact
between Hie Quaternary and Tertiary formations is located.
Below the contact are deposits of the Miocene series that
consist of undifferentiated sands, silty clays, and thin-bedded
limestones extending to an approximate depth of 305 m.
The shallow subsurface consists of fluvial sediments with a
confined aquifer in the bottom 20 m of the Quaternary
sediments. The aquifer is confined above and below by clay
bearing strata that extend laterally for several thousands of
meters or more. The upper confining layer is located about
40 m deep, and the thickness of the aquifer is relatively
constant at about 20 m. The piezometric surface for the
confined aquifer ranges from land surface elevation to two or
three meters below land surface depending on seasonal and
climatic conditions. In general, the confined aquifer matrix may
be described as a medium sand with silt and clay fractions
ranging from one percent to fifteen percent by weight.
c. Well Installation
Four observation wells at the Mobile test site were designed to
provide two distinct environments for construction and
development studies. Two of the wells (Figure VT-2) were
deep wells and were installed to depths of 60 m, with fully-
penetrating screens in the confined zone between 40 m and
60 m. The other two wells were shallow wells drilled in the
[23 SAND 4 CLAY
I I SAND
EE3 SAND & GRAVEL
HIM CLAY
Figure VI-1. Vertical cross-sectional illustration of the subsurface hydrologic system at the Mobile site.
33
-------
15.2-cm IDPVC casing
(Schedule 40)
Efl
.
0 ol
o>.c
Qco
/
PackerJ.'<
End Cap
^Surface
2.5-cm minimum
grout envelope
,10.2-cm schedule 40
slotted PVC screen
Figure VI-2. Schematic diagram providing the details of the shallow and deep wells constructed at the Mobile site. Well casing and well
screen are the same dimension and schedule. The only difference is the depth dimensions of the casing and screen.
sand, gravel, and clay strata (Figure VI-1) to a depth of 30 in
and screened between 4.5 m and 30 m. The layout of the
wells is in a 10 m xlO m square arrangement with a well
positioned at each comer. The designations for the wells are A
and B for the shallow wells, and C and D for the deep wells.
During construction, a portion of the installed casing of well D
was fractured allowing sediment to enter the well. Thus, well
D was then considered non-functional and removed from the
study. All wells were installed using the direct circulation
rotary drilling method. The shallow wells were drilled to
accommodate a 15.2-cm schedule-40 PVC casing from ground
surface to an approximate depth of 4.5 meters. The casing
was sealed in place by cement grout and left undisturbed for
12 hours. Following the 12-hour period, the well was drilled
to its final depth (30 m) to receive a 10.2-cm slotted schedule-
40 PVC screen. No artificial filter pack material was used in
the construction of these wells. A 3-m section of 10.2-cm
PVC pipe was attached to the screen and extended upward
approximately 3 m from the bottom of the casing. The annular
space between the pipe and the 15.2-cm casing was closed by
a packer. The deep well was constructed in the same manner
using the same size casing and screen. After completion, each
well was developed by forcing air into the casing to agitate,
surge, and lift the water; mildly cleaning the well of drilling
fluid, cuttings, and fines.
VI-2 Test Description
The direct rotary drilling method used a mud slurry as a
drilling fluid during the construction operation. Site
conditions necessitated use of this method. However, the mud
cake has a deleterious effect on the production capacity of the
well due to its very low hydraulic conductivity. Development
was necessary to remove the mud cake to the extent practical.
In this study, air development was used to surge compressed
air into the casing of each well for approximately fifteen
minutes. The key question, of course, was how much
development was necessary to remove the mud without
affecting the hydraulic properties of the aquifer materials
outside the well? This was the main question studied in the
field tests reported below.
The series of testing performed on wells A, B, and C were
identical with a few exceptions and consisted of four sets of
tests (Table Vl-1). The first set consisted of an ambient flow
test and a pump-induced flow test which occurred before any
well development. The second, third, and fourth sets were
identical in procedure; air development, followed by an
ambient flow test, followed by a pump-induced flow test.
Some pump-induced flow tests were repeated following an
overnight rest period (without a repeat of the air development)
34
-------
Table VI-1. Pump-Induced Flow and Ambient Flow Tests Performed on Wells A, B, and C
Ambient Flow Tests
Well A
Well B
Well C
Pre-development
1st Development
2nd Development
3rd Development
X
X
X
X
X
X
X
X
X
X
Pump-Induced Flow Tests
Pre-development
1st Development
2nd Development
3rd Development
X
X
X
X
X
X
X
X
X
X
X
Repeat Tests (Overnight)
Ambient Flow Test
1st Development
Pump-Induced Flow Tests
1st Development
2nd Development
to determine if the results were rcpeatable. The ambient and
pump-induced flow tests were conducted in the same manner
as far as flowmeter position was concerned. The
electromagnetic flowmeter was lowered to the bottom of the
borehole where the first measurement (zero reference) was
taken. Subsequent flow measurements were recorded as the
flowmeter was raised at 1.52-m increments until reaching the
top of the screen. The pump-induced flow was maintained at a
constant rale and averaged about 34 L/min for all tests.
VI-3 Test Results
a. Shallow Well A
Well A served as the standard for measurement procedures at
the remaining wells. Due to the experimental function of this
well, ambient testing was not performed until after the second
development. Pumping tests were completed from the pre-
developed stage through the three stages of development. Net
flow and ambient flow plots (Figure VI-3), as well as
differential net flow charts (Figure VI-4) of the flow
distribution following the second and third developments, are
presented to illustrate the effects of a third air-development.
The ambient flow curve shown (Figure VI-3a) indicates that
either the natural vertical flow has been altered somewhat by
applying a third development, or the hydraulic conditions
causing the ambient flow have changed. The natural flow has
decreased by varying amounts at every measurement point
except the uppermost point and the greater shifts occur
between depths of 15 m and 27 m. For example, at a depth of
18.3 in the ambient flow after the second development was 3.6
L/min, and at the same depth, after the third development it
was 2.9 L/min, an eighteen percent drop. It is possible that
variable hydraulic stresses on the aquifer resulted in the shifts
rather than changes in well development. Several chemical
processing plants and an electrical power plant, some of which
place huge water withdrawal demands on the aquifer, are
located within a 5-km radius of the testing site.
The net flow plot (Figure VI-3b) shows the pump-induced
flow data that have been corrected for the ambient flow
measurements. Both data plots begin at 27.4 m with nearly
the same nominal flow, and increase in tandem to about 15 m.
At this point, the plots indicate that at least some variable
development has occurred between depths of 6 m and 14 m.
The 4.2 L/min flow variation at 9.1 in is a seventeen percent
increase, but there is no clear evidence of its cause. This shift
-------
Q.
-------
-4.6'
-6.1'
0 -15
is -is.
£ -21
[D -24
-27
2nd Development
i
i
i
i
i
i
i
i
i
02468
Differential Net Flow (L/min)
(a)
-4.6
-6.1
? -9.1.
S _12
§ -15'
15 -18.
> -21
[D -24
-27
2nd Development (overnight)
^^^^=^^^^=\
i
i
i
i
i
i
i
i
32468
Differential Net Flow (L/min)
(b)
-4.6'
-6.1'
§ -15'
is -is.
£ -21.
Uj -24
-27
3rd Development
i
i
1
i
i
i
i
i
02468
Differential Net Flow (L/min)
(c)
Figure VI-4. (Well A) Differential net flow (a) obtained after 2nd development, (b) obtained after 2nd development and an overnight
waiting period, and (c) obtained after 3rd development.
37
-------
© O Pre-Development
1st Development
0 0 2nd Development
A A 3rd Development
-5 -
-10
f
2
-20
-25
-30
O G Pre-Development
B B 1st Development
O 0 2nd Development
A A 3rd Development
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
Ambient Flow (L/min)
(a)
0.0
10.0
20.0
30.0
40.0
Net Flow (L/min)
(b)
Figure VI-5. (Well B) (a) Ambient flow as measured by the EM flowmeter and (b) net flow (total flow - ambient flow). Data sets were
obtained prior to development and after 1st, 2nd, and 3rd developments.
changes at the mid-screen level. It appears that for well B,
ambient flow varied more than net flow, as seen between pre-
development first development, and third development.
The largest differences in the distribution of DNF (Figure
VI-6) occur between pre-development and first development
while only minor differences occur after subsequent
developments. In addition, the large shift in ambient flow
occurring after the third development is not obvious in the
corresponding DNF chart. This result supports the conclusion
that the ambient flow variations must be removed from the
total flows to obtain consistent results. Since DNF reflects the
flow entering a screen segment due to pumping, it is logical
that pump-induced flow would be less responsive to
development than ambient flow.
For well B, an additional test was performed after the first
development to determine the effects of an overnight time
period on the ambient flows. No additional development was
performed between these tests. The results (Figure Vl-7) are
quite similar to those obtained the previous day.
c. Deep Well C
The ambient flow in the confined aquifer screened by well C
would be expected to be much smaller than those in the
unconfined or semi-confined sediments above. It is known
also that local industries placed less external demand on this
aquifer than the shallow aquifers. The maximum ambient
flow (Figure VT-8a) was about 1.75 L/min and was recorded
prior to the first development. This flow is in contrast to 7 L/
min recorded as maximum ambient flow in shallow well B.
Development of well C resulted in irregular shifts of the ambient
flows, and the tendency was toward the aquifer having no natural
vertical flows. This result is clearly seen in the measurement
taken after the third development where the majority of ambient
flows are less than approximately 0.5 L/min.
Variability in the net flow curves following the first, second,
and third developments is generally small throughout the
range of measurements (Figure VT-8b). It appears that most
development took place in the upper three measurement
stations, which corresponds to the same area where the larger
shifts in the ambient flow occur.
The DNF profiles for well C are shown in Figure VI-9.
Although the pumping test data before the first development
were not available, the relevant sequential test data following
that stage are presented. A moderate change in the profile is
observed through the progression of developments. Some of
the flow in the upper half of the confined aquifer has shifted to
the bottom half, though most of the flow still remains in the
upper portion. The DNF chart displaying the pump-induced
flow test after an overnight period has the same general
distribution as the first development pump-induced flow test
performed the previous day.
38
-------
-4.61
-7.6"
? -11]
-14"
fe -17-
S -20-
5 -23:
UJ -26"
-29:
Pre-Development
E3
i
0123456
Differential Net Flow (L/min)
-4.61
-7.6"
? -11]
- -14-
fe -17-
IS -20-
5 -23:
HI -26"
-29:
1st Development
= ......
0123456
Differential Net Flow (L/min)
(a)
(b)
-4.61
-7.6"
= -11:
^ -14:
0 -17-
re -20-
S -23:
UJ -26"
-29'
2nd Development
I.MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMJ
'"" '
3 ......
0123456
Differential Net Flow (L/min)
-4.61
-7.6"
E '11-
- -14:
o -17:
w -20-
S -23;
uj -26:
-29'
3rd Development
E3
!=i
0123456
Differential Net Flow (L/min)
(c)
(d)
Figure VI-6. (Well B) Differential net flow obtained (a) prior to development, (b) after 1st development, (c) after 2nd development, and (d) after 3rd development.
-------
0 Q 1st Development
B B 1st Development (overnight)
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Ambient Flow (L/min)
(a)
Figure VI-7. (Well B) Ambient flow obtained after the first
development and repeated after an overnight
period.
VI-4 Summary and Discussion
As noted earlier, the key question is how much development is
necessary to return disturbed porous media surrounding a well
to a near-natural state. This study has presented some initial
observations of llie effects of successive developments on
three wells. Two wells were shallow and screened in phreatic
or semi-confined aquifers, and the third was a deeper, fully -
penetrating well, screened through a confined aquifer.
The results of tests on well A were inconclusive, in part
because of a lack of ambient flow data from pre-development
and first development stages. In addition to the lack of data, a
good correlation between the flow distributions after second
and third development did not exist as shown by the
differential net flow charts. This was probably due, in part, to
the experimental technique used at this well. There was,
however, a similarity between a test repeated after the second
development and the test performed after the third
development This similarity indicates that the first test may
have been influenced by factors other than those imposed by
the pump-induced flow.
The differential net flow charts for well B show that the first
development made a significant difference in the vertical
distribution of flow to the well. Tests after the first second,
and third developments resulted in quite similar DNF profiles.
-35
-40 -
f-45
£
Q.
-50 -
-55 -
-60
© Q Pre-Development
H H 1st Development
2nd Development
3rd Development
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
Ambient Flow (L/min)
(a)
-35
-40
f-45
£
Q.
V
Q -50
-55
-60
0 O 1st Development
H H 2nd Development
0 0 3rd Development
0.0
10.0
20.0
30.0
40.0
Net Flow (L/min)
(b)
Figure VI-8. (Well C) (a) Ambient flow as measured by the EM flowmeter and (b) net flow (total flow - ambient flow). Data sets are
obtained prior to development and after the 1st, 2nd, and 3rd developments for the ambient flow and after the 1st, 2nd, and
3rd developments for net flow.
40
-------
-40-
-43.
-i- -46'
Elevation
U1 U1 4^
CJ1 K> CO
1st Development
,
m
01 23456
Differential Net Flow (L/min)
(a)
-40 '
-43 .
\± <«\
Elevatior
U1 U1 -k
CJ1 K> CO
2nd Development
::=::::::::::: S3
01 23456
Differential Net Flow (L/min)
-40-
-43
1 -46'
O -49.
*j
-52
V
UJ -55-
1st Development (overnight)
0123456
Differential Net Flow (L/min)
(b)
-40'
-43.
& -49-
> -52'
m -55'
3rd Development
i
^
m
0123456
Differential Net Flow (L/min)
(c)
(d)
Figure VI-9. (Well C) Differential net flow obtained after (a) 1st development, (b) 1st development and an overnight period, (c) 2nd development, and (d) 3rd development.
-------
However, as would be expected, the ambient flow data were
less stable. Part of the irregularity may be attributable to the
varying withdrawal demands placed on the aquifer by the local
industry.
The third well was screened in the confined aquifer,
presumably more shielded from the hydraulic stresses beyond
experimental control. The ambient flow in well C was small
compared to the other wells, although measurable shifts were
observed after each development. The irregularity of ambient
flows may be due more to experimental "noise" than to any
significant change in the natural flow. The differential net
flow charts were moderately dissimilar after each
development, indicating that slight adjustments were still
being made to the media even after the third development.
The results of the tests at the Mobile site were generally
similar to those obtained at the Columbus site. Ambient flow
profiles at both sites were more sensitive to effects such as the
state of well development and minor changes in hydraulic
stresses than induced flow profiles. In addition, the state of
well development appeared to vary with differences in aquifer
materials, as is expected.
42
-------
VII
Case Studies
VII-1 Field Applications
Issues related to aquifer heterogeneity are particularly
important at sites where contaminant transport and fate will be
characterized or subsurface remediation will be performed. A
vertical-component borehole flowmeter may be used to
investigate several aspects of ground-water flow relevant to
monitoring and remedial design (Table VII-1). In order to
demonstrate the utility of flowmeter investigations for gaining
quantitative insight into aquifer characterization, results from
several borehole flowmeter applications are described.
Table VII-1. Characterization Objectives for Borehole Flowmeter
Studies
Measured
Flow Log
Phenomena That Can Be Investigated
Ambient
Conditions
Pumping
Conditions
Direction of the vertical hydraulic gradient
Cross-connection among geological units
intersected by a well
Active fracture locations/zones in bedrock
Active fracture locations in bedrock
Horizontal hydraulic conductivity distribution
Identify zones of preferential flow and,
potentially, contaminant transport for design of
monitoring and remediation networks
VII-2 Columbus AFB, Mississippi
As discussed in Chapter V. borehole flowmeter tests were
performed at two sites on the Columbus Air Force Base
(CAFB). The two sites overlie a highly heterogeneous,
unconsolidated and unconfined fluvial aquifer, are about 60 m
apart, and are located approximately 6 km east of Hie
Tombigbee River and 2.5 km south of the Buttahatchee River.
Seasonal water table fluctuations range from 2 m to 3 m. The
aquifer is composed of approximately 10 m of terrace deposits
consisting primarily of poorly- to well-sorted, sandy gravel
and gravelly sand that often occur in irregular lenses and
layers. Visual inspection of the fades exposed at a quarry
located a few kilometers from (lie site shows a complex series
of lenses with significantly different physical and hydrological
properties. The terrace aquifer is unconformably underlain by
the Cretaceous-age Eutaw Formation, an aquitard consisting
primarily of marine clay and silt.
The Electric Power Research Institute performed the MADE
study (Betson et al, 1985) at the first test site, which covers
about 8 Ha. The MADE experiment included a network of
258 multilevel sampling wells designed to monitor a 20-
month, rapid injection, natural gradient tracer study. Boggs et
al. (1992) provide a review of the tracer experiment. The
dominant feature of the tracer plume is a highly asymmetric
concentration distribution in the longitudinal direction. At the
conclusion of Hie experiment, the more concentrated region of
the plume remained within approximately 20 m of the
injection point, while a more dilute portion extended
downgradient a distance of more than 260 in. In certain
respects, the plume configuration would appear to be the result
of a continuous tracer injection rather than the rapid injection
that actually occurred.
One explanation for the skewed plume is that the tracer was
slowly bleeding from a zone of relatively low hydraulic
conductivity into relatively conductive zones, and thereafter
moved rapidly downgradient. This explanation is supported
by the hydraulic conductivity values from borehole flowmeter
tests performed by Boggs ct al. (1989) and Rchfeldt ct: al.
(1989b). These particular flowmeter tests were performed
primarily with an impeller flowmeter. The electromagnetic
flowmeter was used only to supplement the basic data. Figure
VII-1 shows a vertical cross section of the spatially correlated
hydraulic conductivity profile along the longitudinal axis of
the plume. The profile shows that the tracer was injected into
a region of relatively low transmissivity and 25 m upgradient
of a large region of relatively high transmissivity that
contained several conductive lenses. Thus, the flowmeter
measurements provide the basis for understanding the major
characteristics of the tracer experiment.
The first major application of the electromagnetic flowmeter
in a granular aquifer was at the second CAFB site, which
covers about 1 Ha. This is the location of the tests described
in Chapter V. Numerous pumping, flowmeter, and
recirculating tracer tests were conducted using thirty-seven
fully-screened wells. Shown in Figure VII-2 are areal
distributions for the depth-averaged hydraulic conductivity
values for the uppermost and lowermost two meters of the
43
-------
64
62
60
58
56
C
's 54
S
[j 52
50
48
46
h §
ts
-0.2 < log K < 0.7
-0.9 < log K < -0.2
-1.6
-------
saturated aquifer. In the upper portion of the aquifer, a band of
high hydraulic conductivity crosses the test site. The location
of this band matches the location of a former river meander
shown in a 1956 aerial photograph of the site (Figure V-l). In
the lower portion of the aquifer, the only non-clay material
occurs in a depression of the lower aquilard near the center of
the site. The S-shaped band of material with a relatively high
hydraulic conductivity in the central portion of the figure is
most likely sands and gravel deposited by an earlier river
system. Thus, flowmeter data at the 1-Ha test site provided
sufficient detail to delineate the sand and gravel beds of two
former river channels. Being able to observe correlations
between the hydraulic conductivity field and the geological/
depositional history of a study site provides the type of
detailed data required to understand flow and transport in
heterogeneous aquifers.
VII-3 Oak Ridge National Laboratory,
Tennessee
The electromagnetic borehole flowmeter has been used to
characterize the ground-water flow patterns in fractured
bedrock at the Oak Ridge National Laboratory (ORNL)
(Moore and Young, 1992: Young etal, 1991). The laboratory
is located in the Valley and Ridge physiographic province of
the Appalachian thrust belt of eastern Tennessee. The
geological units consist of fractured sequences of calcareous
shale, siltstone, shaley limestone, and limestone, which
typically dip at angles between 45° to 60° from horizontal.
Measurements at outcrops and on cores of the Conasauga
Group show a density of 10-15 joints per meter in shale and 6-
40 joints per meter in siltstone (Sledz and Huff, 1981). The
joints and fractures are oriented parallel to the bedding planes,
strikes, and dips of the lithologic units.
At ORNL, the electromagnetic -flowmeter has been used in
open boreholes and in boreholes with sand-packed PVC
screens. In most of the deep boreholes, the measured ambient
flow profiles in the wells indicate that cross-communication
was occurring between different fracture zones. Figure VII-3
shows a profile of ambient flow where ground-water enters the
well at a depth of about 100 m and exits at a depth of 41 m.
The ambient flow is produced by natural upward hydraulic
gradients, along with hydraulically active fractures at depths
of 100m and 41 m.
At most of the wells, the permeable vertical zones arc
typically greater than 30 cm, which is the minimum vertical
distance required for a steeply dipping fracture (>60°) to cross
through a 17-cm OD borehole, the size of most of the
boreholes at ORNL. Analysis of flowmeter logs indicate that
the orthogonal spacing between fractures is about 0.15 m to
0.44 m in the shallow bedrock, and about 0.44 m to 0.73 m in
the deeper bedrock. Specific information regarding the
location of fractures assists in correlating lithologic structure
and hydraulically-significant fractures. Differences were
observed in the flow profiles between different strata and
regions. In some areas, a permeable flow zone is located near
the top of bedrock. This zone was likely produced by
£
Q.
&
20
40
60
80
100
120
0.012 0.008 0.004
Ambient Flow (Us)
(a)
50
100
150
200
250
0.04 0.03 0.02 0.01
Induced Flow (Us)
(b)
Figure VII-3. (a) Ambient flow distribution in a well and (b) the flow distribution induced by constant-rate pumping.
-------
weathering and serves as a pathway for significant shallow
ground-water flow toward valley streams.
At ORNL, newly drilled boreholes with depths near 500 m
penetrate several fracture zones. In some cases, the different
fracture zones contain ground waters of significantly different
chemistry and from different recharge zones. Among the
applications of the flowmeter data at ORNL is a refinement of
the monitoring program based on the location of hydrauli-
cally active fractures. Once the hydraulically active fractures
are located, specific borehole zones may be isolated so that
selective monitoring can include fluids from fractures or
fracture zones and matrix fluids with long residence times.
These data may be used to identify changes in water chemistry
with depth and water chemistry signatures associated with a
particular rock type.
VII-4 The Oklahoma Refining Company
Superfund Site, Oklahoma
The Oklahoma Refining Company (ORC) Superfund Site is
located in Caddo County on the eastern edge of Cyril,
Oklahoma. Site topography consists of low. rolling hills with
a deeply incised drainage system. Soils are characteristically
red silly clay loams with low to very low permeability.
Alluvium and terrace deposits are present in and around old
stream channel sediments. Soils are underlain by the
Weatherford Member of the Cloud Chief Formation (gypsum)
in the northwestern part of the ORC site, and the Rush Springs
Sandstone elsewhere. Ground water in the Weatherford
commingles with that of the Rush Springs Sandstone, which is
believed to act as an unconfined aquifer. The Rush Springs
Sandstone consists of even-bedded to highly cross-bedded,
very fine-grained, silty sandstone and outcrops on the eastern
side of the ORC site (Bechtel. 1991).
Electromagnetic flowmeter field demonstrations were
conducted at three pairs of ORC wells. One well pair
consisted of wells NE-2 and NE-3, which lies in a region
where gypsum overlies sandstone. Figure VII-4 shows the
flow distributions for ambient and induced-flow conditions at
well NE-2. The flow distributions for well NE-3 were very
similar. The profile of induced flow indicates that about
eighty percent of the total flow to the well originates from the
upper thirty percent of the well screen. This result suggests
that materials in the upper portion of the screened interval are
more permeable than those in the lower portion.
The downward ambient flows indicate downward vertical
hydraulic gradients and possible cross connection between
different hydrostratigraphic units. At wells NE-2 and NE-3,
about 0.2 L/iriin enters the well near the water table and
migrates down the well, where it flows into the aquifer. Of the
0.2 L/min, about half of it enters the sandstone at a depth 18 m
below the top of well casing (TOC). If the recharging ground
water was contaminated, ambient flows in the well would
accelerate the downward migration of ground-water
contamination. If the recharging ground water was clean but
the deeper ground water was contaminated, ambient flows in
the well could dilute ground-water contamination in the
vicinity of well and in samples, hindering characterization of
the ground-water contamination.
In the southeast quadrant of the site, flowmeter tests were
performed on a well pair consisting of wells IBB-4 and SBB-
36. The ambient flows at these wells (Figures VII-4) are
upward. Well IBB-4 was a flowing artesian well at the time of
the test. The flow profiles at well IBB-4 indicate that the
source of the artesian ground water is a narrow zone near a
depth of 36 m. Boring logs indicate that a 0.8-m thick
siltstone layer exists at the 36-m depth, which appears to be
the source of the artesian ground-water flow. Well SBB-36
shows two narrow zones of relatively high flow near depths of
13.5 m and 15.5 m. The limited flowmeter testing at the ORC
test site provided valuable information related to aquifer
heterogeneity and potential contaminant transport.
VII-5 Gilson Road Superfund Site,
New Hampshire
The Gilson Road site at Nashua, New Hampshire, was
originally a 2.4-Ha landfill for refuse and demolition material.
In 1979, it became a site for disposal of other wastes. Ground-
water contamination formed a plume over 450 in wide and 33
m deep, which was estimated to be moving at a rate of 0.6 m/
day (Morrison, 1983). A bentonite shirty wall was constructed
to contain the plume and the surface was capped with a
synthetic membrane to prevent infiltration of rainfall. A
ground-water pump-and-treat system was in operation for
several years.
The Gilson Road site is underlain primarily by stratified,
unconsolidated sand and gravel deposits of glacial origin. The
permeable sand and gravel deposits arc underlain by a thin,
discontinuous, low-permeability glacial till unit, that is up to
3-m thick. Bedrock underlying the till is biotite schist of the
Merrimack Group, which is differentially weathered and
fractured across the site (Morrison, 1983).
Electromagnetic flowmeter tests were conducted at four
monitoring wells and three recover}' wells of the pump-and-
treat remediation system. The recovery wells were designed
to fully penetrate the drift and till overburden. None of the
available well logs indicated intersection of the upper bedrock.
In all three recover}' wells, the measured flow (Figure VII-5)
was significantly greater near the top of the well screen. These
results may be due to use of a coarse-grained filter pack in
well construction and may not be representative of actual site
conditions. However, if the results arc representative of the
shallow hydraulic structure in this aquifer, they would indicate
potential inefficiencies in the recovery of ground water from
the lower portions of the overburden using these wells.
VII-6 Mirror Lake, New Hampshire
The Mirror Lake drainage basin is located in the Hubbard
Brook Experimental Forest in the White Mountains of central
New Hampshire. The drainage basin is a small, well-defined
46
-------
WELL NE-2
E
8
Q.
0
Q
15
20
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0 0.05
Flow (L/min)
WELL NE-2
b)
Pumping Rate = 3.8 L/min
2 3
Flow (L/min)
WELL SBB-36
WELL IBB-4
11
12
~
O 13
R
^14
&
.c
fi
,515
16
-
)
_
-
i
i
c)
T
Top of Screen f
***
r
j
6
' >
i 9
I 0
> c»
IOT-
Bottom of Screen
5 10 15 20 25 30 3
Flow (L/min)
Ambient Pumping
10
11-
12'
jH
s 14-
1 15,
Q 161
i
17
^ Q
WELLK
Top of Screen -
. -""
f
/'
Hi
V
d>
Bottom of Screen
0 5 10 15 20 25 30 35
Flow (L/min)
Figure VII-5. Ambient and induced flow distributions for wells I, J, and K at the Gilson Road site.
47
-------
hydrologic environment covering 85 Ha. A test site was
established in the southwest corner of the basin by the U.S.
Geological Survey for three primary purposes: long-term
monitoring of bedrock environments; maintenance of a con-
trolled field-scale laboratory to test new equipment, methods,
and interpretive models; and characterization of fluid move-
ment and solute transport in fractured rock (Winter, 1984).
Bedrock underlying the Mirror Lake drainage basin is
primarily composed of Silurian-age schist (Shapiro and Hsieh,
1991). The schist is intruded by granite, and less commonly
pegmatites and basalts. The rocks are extensively fractured
from folding and faulting during successive orogenies.
Overlying the bedrock is a layer of glacial drift, predominantly
till, which varies in thickness from zero to 49 m. At the test
site, bedrock is a granite dike, which is overlain by 12 in to 15
m of glacial drift.
Three existing wells, each drilled to a depth of 75 m, were
chosen for demonstration of Hie electromagnetic borehole
flowmeter. Well FSE-06 (Figure VII-6a) exhibited the largest
ambient flow at 0.3 L/min. The results indicate two zones of
significant transmissivity; an upper zone between 30 m and 35 m
below TOC and a lower zone around 57 m to 64 m. The two
zones are probably related to major rock fractures. Further
evidence of transmissive fractures can be seen in the flowmeter
profile obtained during pumping (Figure VII-6b). Ground water
predominantly enters the well from these two intervals.
Well FSE-09 exhibited much lower ambient flow (Figure
Vll-6c) than that detected in FSE-06, but the flows are high
enough to indicate a hydraulically active fracture zone near a
depth of 42 m to 46 m. The presence of this fracture zone is
supported by the induced flow profile (Figure VII-6d). Very
low magnitude ambient flows (Figure VII-6e) were detected in
Well FSE-10. However, a large change in the ambient flow
around a depth of 30 m to 38 m suggests that a transmissive
fracture zone is present. Good correlation with the pumping
induced flow data (Figure VII-6f) verifies the fracture zone.
For comparison, results from acoustic televiewer surveys of
wells FSE-09 and FSE-10 were made available by the U.S.
Geological Survey. The televiewer images many linear
features in the walls of both holes, including a relatively large
feature at a depth of about 44 in in Well FSE-09. Similar
general correlations between televiewer features and the flow
profiles were observed in well FSE-10.
Collective interpretation of the flowmeter logs indicates that
there may be a network of hydraulically-active fractures at
approximately 40 m below the ground surface in this area of
the site. Data from these different logging methods arc highly
complimentary. The acoustic televiewer may be used to esti-
mate the density and orientation of linear features, including
fractures. The borehole flowmeter provides information to
identify which features or zones are hydraulically active.
VII-7 Logan Martin Dam, Alabama
Logan Martin Dam is located in east-central Alabama in a
complex geologic terrain. Situated on the lower Knox Group
of the Pell City Thrust Sheet, the dam rests on a bed
comprised of eighty percent to ninety7 percent dolomite which
ranges from fine- to coarse-grained. The remaining rock is
made up of a collection of cherts, shaley limestones, and fine-
grained silica. The rock is highly fractured and creates a
situation that is extremely conducive to diffuse ground-water
flow. The hydrogeology is complicated further by Hie fact that
certain layers of the underlying material are preferentially
solutioned. This solutioning has created areas of conduit flow-
within the rock.
Logan Martin Dam has experienced extensive problems with
seepage and poor water quality7 downstream of the dam since
its completion in 1964. The rate of leakage under the dam
has increased from 168 m3/min in 1964 to approximately
1104 m3/min, with the flow rate increasing only 84 m3/min
since 1974. In 1.968, it was observed that the flow rate at a
monitoring weir increased sharply from 0.6 mVmin to 10.2
m3/min. A steep-walled sinkhole then developed on the
downstream side of the earthworks. Subsequent
investigations revealed multiple sinkholes in the reservoir
upstream of the dam. Grouting the earthworks and filling the
sinkholes reduced the flow from the monitoring weir to the
original 0.6 m3/min.
Geologic mapping based on outcrops, lineaments, and core
analyses suggested that a zone of rocks within the Knox
Group, called Target Zone #1, was a major source of leakage
(Rcdwinc et al, 1990). The geophysical/hydraulic procedure
to test this hypothesis was to use various existing wells and
coreholes to run a suite of caliper, temperature, and flowmeter
logs in an attempt to detect the vertical movement of water
within the target zone, thereby documenting the qualitative
information resulting from the earlier geologic study. Of the
wells that were tested, most were clogged with mud at various
depths, thus no substantive data were obtained from these
tests. However, two of the wells (Well 332 and Well 301)
were clear for logging.
At Well 332, upward flows of about 130 L/min were
encountered between 78 in and 110 m above mean sea level
(AMSL) (Figure VII-7). The natural flow entered Hie
borehole at 78 m to 80 m AMSL and flowed up the borehole
until reaching an exit point at about 110 m to 112 m AMSL.
These data were interpreted as indicative of large fractures or
conduits at these elevations.
Well 301 provided different results. At 1.5 m AMSL, a flow
of 51 L/min was measured entering the borehole. This flow
steadily decreased to 43.3 L/min at 29.5 m AMSL where the
flow then disappeared completely over a short section of the
borehole. The flow then resumed at 30.5 m AMSL and rose to
a maximum at about 33.5 m AMSL, then decreased, first
slowly and then rapidly, until 48 m AMSL, where flow was no
longer detected. The decrease of flow occurring between 29.5
in and 30.5 in AMSL was due to a large solution opening
where the cross-sectional area for flow was enlarged
dramatically, leading to a substantial decrease in flow velocity
in the wellborc. This expansion in the wellborc was shown on
a caliper log.
48
-------
10
20
1 30
o
0
a) WellFSE-06
Bottom of Casing
emmmmmmmmmmmmmmmmmmm
o
O
tl
40
50
60
70
80
Bottom of Borehole
-0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0
O
0
!-
^o
O
ti
10
20
30
40
50
60
70
80
C) Well FSE-09
Bottom of Casing
Bottom of Borehole
-0.1
-0.05
0
0.05
0.1
10
20
70
80
6) WellFSE-10
Bottom of Casing
-0.02 -0.015
-0.01 -0.005
Flow (L/min)
0.15
0.005
Ambient
b) WellFSE-06
Bottom of Casing
Bottom of Borehole
8 10 12 14
d) Well FSE-09
Bottom of Casing
temmmmmmmmmmmmmmmmmmmmmmmmmmmm
mmmmmmmmmmmmmf^t
9
\
--6
o
I
Bottom of Borehole
f) WellFSE-10
Bottom of Casing
I O
Bottom of Borehole
0.5
1 1.5 2
Flow (L/min)
2.5
---O
Pumping
Figure VII-6. Ambient and induced flow distributions for wells FSE-06, FSE-09, and FSE-10 at Mirror Lake, New Hampshire.
49
-------
Well 332
Well 301
<
E
CO
_0>
LU
120
110
100
90
70
60
60
50
40
30
20
10
25 50 75 100 125
Flow (L/min)
150
10 20 30 40 50
Flow (L/min)
60
Figure VII-7. Substantial ambient flow moving from one stratum to another as detected by an impeller flowmeter. The flow is moving
under a dam, the base of which is at an elevation of about 140 m AMSL. Flowmeter data were used to help select a
geologic model for explaining the large leakage of water low in dissolved oxygen that was observed below the dam.
All detected flows were upward, indicating that water leakage
was from around and/or under the reservoir. It was concluded
that the borehole flowmeter was a valuable device for locating
transmissive features and directly measuring flows of ground
water in fractured rock systems.
VII-8 Cape Cod, Massachusetts
In a study by Hess et al. (1992). a comparison of the
variability of hydraulic conductivity was presented for in situ
borehole flowmeter measurements and lab pcrmeameter
measurements. The theory and application pertaining to the
flowmeter data are those techniques given by Hufschmied
(1983, 1986) and Rehfeldt et al. (1989a, b). "Much of what
follows comes directly from Hess et al. (1992).
Field studies into the characterization of aquifer heterogeneity
were conducted in an abandoned gravel pit south of Otis Air
Force Base at Cape Cod, Massachusetts. The unconfined
aquifer is located in a glacial outwash plain and is composed
of clean, medium-to-coarse sand and gravel, containing
typically less than one percent silt and clay. The aquifer is
about 30 m thick and is underlain by less permeable silty sand
and till deposits. The water table is about 14 m AMSL and
about 6 m below land surface. The hydraulic gradient is
approximately 0.0015 m/m and indicates a southward flow
direction. Ground-water seepage velocity averages 0.42 m/d
as calculated from average properties of the aquifer (LeBlanc
et al.. 1991). Surface exposures of the outwash reveal
interbedded sand and gravel deposits. Crossbedded troughs up
to 1 m wide, but typically less than 0.5 m high, are observed in
exposures perpendicular to the hypothesized paleocurrent
direction, which is north to south. Parallel to this direction,
troughs exhibit a tubular form and arc several meters long
(Hess and Wolf, 1991).
Sixteen wells were installed for flowmeter tests. Fifteen of the
wells were installed by a drivc-and-wash technique to
minimize the disturbance of the aquifer region surrounding the
well. One well was installed using a hollow-stem auger so
that core material could be obtained for permeameter analysis,
which could then be directly compared to the flowmeter
analysis. In addition, core samples were taken from the upper
6 m of each of the boreholes. Each 5.1 -cm diameter
flowmeter well was screened over the upper 12 m of the
saturated zone. Only the top 6 m to 7 m of the profiles were
used for comparisons. These intervals correspond to the core
sample zones, and also to the horizon of the aquifer through
which die tracers moved during an earlier tracer test.
Multiport, constant-head permeameter analysis of the core
samples was the second method used in this study to obtain a
measure of hydraulic conductivity. Measurements were taken
from each section of the core between adjacent manometer
probes inserted at 5-cm to 10-cm intervals along the core
sample length. Summary statistics, mean, variance, and range
for the hydraulic conductivity data sets from the two methods
are presented in Table VII-2. The flowmeter data have
significantly higher mean and variance values than do the
permeameter data. The flowmeter geometric mean of
0.11 cm/s lies within the range of values previously estimated
for this aquifer and is only slightly lower than the value
50
-------
Table VII-2. Summary Statistics of Hydraulic Conductivity Data
Obtained Using a Borehole Flowmeter Method and
a Permeameter Analysis Method (from Hess et al.
(1992))
Flowmeter Permeameter
Number of wells or coring locations
16
Number of hydraulic conductivity values (K) 668
Mean vertical spacing (cm)
Geometric mean of K (cm/s)
Variance of In K (of2)
Range
Minimum K(cm/s)
Maximum K (cm/s)
Range in horizontal spacing (m)
15
0.11
0.24
0.013
0.37
0.9-24
16
825
7.3
0.035
0.14
0.006
0.14
0.9-24
estimated from the aquifer and tracer tests (0.13 cm/s). The
permeameter mean of 0.035 cm/s is significantly lower than
previously reported values. Several potential reasons for the
low permeameter calculations include: local-scale anisotropy,
nonreprcsentative sampling techniques, and the influence of
compaction during the coring process. In comparison, the
flowmeter appears to supply more representative data, and is
an easier and cost-effective alternative to core sampling and
permeameter measurements.
VII-9 Summary
Sensitive borehole flowmeters, including the electromagnetic
flowmeter described in this report, are valuable tools for
detailed characterization of hydrogeology. In many cases,
these tools may be used in existing wells to obtain such
information as natural ground-water flow patterns within the
well, identification of hydraulically active fracture zones in
rock, and, potentially, estimates of the relative hydraulic
conductivities of materials adjacent to the well screen. In
conjunction with other data, including information from
geologic and geophysical logs, these results may be used to
define the hydrostratigraphy of aquifer formations and
evaluate such issues as contaminant transport and fate,
efficient design of extraction and injection systems for
subsurface remediation, and monitoring network design.
51
-------
52
-------
VIII
References
Aller, L., T. W. Bennett, G. Hackett, R. J. Petty, J. H. Lehr, H.
Sedoris. D. M. Nielsen, and J. E. Denne, 1989. Handbook
of Suggested Practices for the Design and Installation of
Ground-Water Monitoring Wells, National Ground Water
Association, Dublin, Ohio.
Barcelona. M. J., and J. A. Helfrich, 1986. Well construction
and purging effects on ground-water samples,
Environmental Science and Technology, 20:1179-1184.
Beclitel Environmental. Inc., 1991. Remedial investigation
report: Oklahoma Refining Company (ORC) Superfund
Site, Cyril, Oklahoma, Contract No. BP02358, Houston,
Texas.
Betson, R. P., J. M. Boggs, S. C. Young, and L. W. Gelhar,
1985. Macrodispersion Experiment (MADE): Design of a
field experiment to investigate transport processes in a
saturated groundwater zone, EPR1 Interim Report, EA-
4082, Electric Power Research Institute, Palo Alto,
California.
Boggs, J. M., S. C. Young, D. J. Benton, and Y C. Chung,
1989. Hydrogeologic characterization of the MADE site,
EPRI Interim Report, EN-6915. Electric Power Research
Institute, Palo Alto, California.
Boggs, J. M., S. C. Young, L. M. Beard, L. W. Gelhar, K. R.
Rehfeldt, and E. E. Adams, 1992. Field study of
dispersion in a heterogeneous aquifer, 1. Overview and
site description, Water Resources Research, 28:3281-
3292.
Boman, O.K., FJ. Molz, andK.D. Boone, 1997. Borehole
flowmeter application in fluvial sediments: Methodology,
results, and assessment, Ground Water, 35:443-450.
Chu, W. C., J. Garcia-Rivcra, andR. Raghavan, 1980.
Analysis of interference test data influenced by wellbore
storage and skin at the flowing well, Journal of Petroleum
Technology. 32(1): 171.
Church, P.E., and G.E. Granato, 1996. Bias in ground-water
data caused by well-bore flow in long-screen wells.
Ground Water, 34:262-273.
Collar, R.J., and PA. Mock, 1997. Using water-supply wells to
investigate vertical ground-water quality, Ground Water,
35:743-750.
Collinson, J. D., andD. B. Thompson, 1989. Sedimentary
Structures, Routledge, Chapman & Hall, Inc., New York,
New York.
Cooper, Jr., H. H., and C. E. Jacob, 1946. A generalized
graphical method for evaluations of formation constants
and summarizing well-field history, Transactions of the
American Geophysical Union, 27:526-534.
Driscoll, F. G., 1986. Groundwater and Wells, Johnson
Filtration Systems, St. Paul, Minnesota.
Dudgeon, C. R., and P. S. Huyakorn, 1976. Effects of near-
well permeability variation on well performance,
Australian Water Resource Council Technical Paper No.
18, Canberra, Australia.
Person, L. M.. ed., 1983. Standards and Practices for
Instrumentation, Instrument Society of America, Research
Triangle Park, North Carolina.
Hackett G., 1987. Drilling and constructing monitoring wells
with hollow/stem augers. Part 1: Drilling considerations,
Ground Water Monitoring Review, 7:51 -62.
Hackett, G., 1988. Drilling and constructing monitoring wells
with hollow/stem augers. Part 2: Monitoring well
installation, Ground Water Monitoring Review, 8:60-68.
Haley, J. L., B. Hanson, C. Enfield, and J. Glass, 1991.
Evaluating the effectiveness of ground-water extraction
systems, Ground Water Monitoring Review, 11:119-124.
Hanson, R.T., and T. Nishikawa, 1996. Combined use of
flowmeter and time-drawdown data to estimate hydraulic
conductivities in layered aquifer systems. Ground Water,
34:84-94.
Harlan, R. L., K. E. Kolrn, andE. D. Gutentag, 1989. Water/Well
Design and Construction. Elsevier, New York, New York.
Hawkins, M. F., Jr., 1956. A note on the skin effect, Trans. Am.
lust. Min. Metall. Pet. Eng., 207:356-367.
Hess, A. E., 1982. A heat-pulse flowmeter for measuring low
velocities in boreholes, U.S. Geological Survey Open-File
Report 82-699, U.S. Geological Survey, Reston, Virginia.
Hess, A. E., 1986. Identifying hydraulically conductive
fractures with a slow-velocity borehole flowmeter,
Canadian GeotechnicalJournal. 23:69-78.
53
-------
Hess, A. E., and F. L, Paiiiet, 1990. Applications of the
thermal-pulse flowmeter in the hydraulic characterization
of fractured rocks, in Geophysical Applications for
GeotechnicalInvestigations, Paillet and Saunders, ed.,
ASTM STP 1101, ASTM, Philadelphia, Pennsylvania, 99-
112.
Hess, K. M., and S. H. Wolf, 1991. Techniques to determine
spatial variations in hydraulic conductivity of sand and
gravel, EPA/600/2-91/006, U.S. Environmental Protection
Agency, Ada, Oklahoma.
Hess, K. M., S. H. Wolf, andM. A. Cclia, 1992. Large-scale
natural gradient tracer test in sand and gravel, Cape Cod,
Massachusetts, 3. Hydraulic conductivity variability and
calculated macrodispersivities. Water Resources
Research, 8:2011-2027.
Hufschniied. P., 1983. Ermittlung der durchlassigkeit von
lockcrgesteins-grundwasscrlcitcrn, cine vcrglcichcndc
untersuchcung verschiednener feldmellioden, Ph.D.
Dissertation 7397, Eidgenossische Technische
Hochschule, Zurich, Switzerland.
Hufschniied, P., 1986. Estimation of three-dimensional
statistically anisotropic hydraulic conductivity field by
means of a single well pumping test combined with
flowmeter measurements, Hydrogeologie, 2:163-174.
Javandel, L, and P.A. Witherspoon, 1969. A method of
analyzing transient fluid flow in multilayered aquifers,
Water Resources Research. 5:856-869.
Kabala, Z.J., 1994. Measuring distributions of hydraulic
conductivity and specific storage by the double flowmeter
test, Water Resources Research, 30:685-690.
Kaye, J. M., 1955. Certain aspects of the geology of Lowndes
County, Mississippi, M. S. Thesis, Mississippi State
University.
Keely, J. F., and K. Boateng, 1987a. Monitoring well
installation, purging and sampling techniques - Part 1:
Conceptualizations, Ground Water, 25:300-314.
Keely, J. F., and K. Boateng, 1987b. Monitoring well
installation, purging and sampling techniques - Part 2:
Case histories, Ground Water, 25:427-439.
LeBlanc, D. R., S. R. Garabcdian, K. M. Hess, L. W. Gelhar,
R. D. Quadri, K. G. Stollenwork, and W. W. Wood, 1991.
Large-scale natural gradient tracer test in sand and gravel,
Cape Cod, Massachusetts, 1. Experimental design and
observed tracer movement, Water Resources Research,
27:895-910.
Liptak, B. G., andL. Venczel, 1982. Instrument Engineers'
Handbook, Chilton Book Company. Radnor,
Pennsylvania.
Long, J.C.S., J. S. Remer, C. R. Wilson, and P. A.
Witherspoon, 1982. Porous media equivalents for
networks of discontinuous fractures. Water Resources
Research, 18:645-658.
Martin-Hayden, J.M., and G.A. Robbins, 1997. Plume
distortion and apparent attenuation due to concentration
averaging in monitoring wells. Ground Water, 35:339-
346.
Mercer, J.W., D.C. Skipp, and D. Griffin, 1990. Basics of
pump-and-treat remediation technology, EPA/600/8-90/
003, U.S. Environmental Protection Agency, Ada,
Oklahoma.
Minning, R. C., 1982. Monitoring well design and installation,
Proceedings of the Second National Symposium on
Aquifer Restoration and Ground Water Monitoring,
Columbus, Ohio, National Ground Water Association,
Dublin, Ohio, 194-197.
Moench, A. F., and P. A. Hsieh, 1985. Analysis of slug test
data in a well with finite thickness skin, in Memoirs of the
17th International Congress on the Hydrogeology of
Rocks of Low Permeability, 17(1): 1729, Int. Assoc.
Hydrogeol., Tucson, Arizona.
Molz, F. J., and S. C. Young, 1993. Development and
application of borehole flow meters for environmental
assessment, The Log Analyst, 34:13-23.
Molz, F. J., O. Giiven, J. G. Melville, J. S. Nohrstedt, and J. K.
Overholtzer, 1988. Forced-gradient tracer tests and
inferred hydraulic conductivity distributions at the Mobile
site, Ground Water. 26:570-579.
Molz, F. ]., R. H. Morin, A. E. Hess, J. G. Melville, and O.
Giiven, 1989. The impeller meter for measuring aquifer
permeability variations: Evaluations and comparison with
other tests, Water Resources Research, 25:1677-1683.
Molz, F. J., O. Giiven, J. G. Melville, (with contributions by I.
Javandel, A. E. Hess, and F. L. Paillet), 1990. A new
approach and methodologies for characterizing the
hydrogeologie properties of aquifers, EPA/600/2-90/002,
U.S. Environmental Protection Agency, Ada, Oklahoma.
Moore, G. K., and S. C. Young, 1992. Borehole flowmeter
data and interpretation, Report ORNL/ER-91, Oak Ridge
National Laboratory, Oak Ridge, Tennessee.
Morin, R. H., A. E. Hess, and F. L. Paillet, 1988a. Determining
the distribution of hydraulic conductivity in fractured
limestone aquifers by simultaneous injection and
geophysical logging, Ground Water, 26:587-595.
Morin, R. H., D. R. LeBlanc, and W. E. Teasdale, 1988b. A
statistical evaluation of formation disturbance produced
by well/casing installation method. Ground Water,
26:207-217.
54
-------
Morrison, A., 1983. Arresting a toxic plume, Civil
Engineering, ASCEr 53(8):35-37.
Moss, R, Jr., and G. E. Moss, 1990. Handbook for
Grounchvater Development, John Wiley & Sons, New
York, New York.
National Ground Water Association, 1989. Development
methods for water wells - an anthology, National Ground
Water Association, Dublin, Ohio.
Ogbe, D.O., 1984. Pulse testing in the presence of wellbore
storage and skin effect, Ph.D. Dissertation, Stanford
University, Stanford, California.
Ogbe, D.O., and W. E. Brigham. 1984. A model for
interference testing with wellbore storage and skin effects
at both wells, Society of Petroleum Engineers 59th
Annual Technical Conference, Houston, Texas.
Paillet, F. L., 1991a. Use of geophysical well-logs in
evaluating crystalline rocks for siting of radioactive waste
repositories. The Log Analyst, 33:85-107.
Paillet, F. L., 1991b. High-resolution flow logging in
observation boreholes during hydraulic testing of
fractured-rock aquifers, Society of Professional Well Log
Analysts 32nd Annual Logging Symposium Transactions,
Midland, Texas, L1-L23.
Paillet, F. L., 1991c. The 1st core hole at Mirror Lake, New
HampshireComparing geophysical logs to core and
cross-hole flow logging, U.S. Geological Survey Toxic
Substance Hydrology Conference, Monterrey, California,
Proceedings, Water Resources Investigations Report 91-
4034, U.S. Geological Survey, Reston, Virginia, 162-171.
Paillet, F. L., and A. E. Hess, 1987. Geophysical well log
analysis of fractured granitic rocks at Atikokan. Ontario.
Canada, U.S. Geological Survey Water Resources
Investigations Report 87-4154, U.S. Geological Survey,
Reston, Virginia.
Paillet, F. L., andK. Kapucu, 1989. Characterization of
fracture permeability and fracture flow modeling at
Mirror Lake, New Hampshire, U.S. Geological Survey
Water Resources Investigations Report 89-4058, U.S.
Geological Survey, Reston, Virginia.
Paillet, F. L,, and W.S. Keys, 1984. Applications of borehole
geophysics in characterizing the hydrology of fractured
rocks, Proc. of NWWA/EPA Conference on Surface and
Borehole Geophysical Methods in Ground-Water
Investigations, San Antonio, TX, 743-761.
Paillet, F. L., A. E. Hess, C, H. Cheng, and E. L. Hardin, 1987.
Characterization of fracture permeability with high-
resolution vertical flow measurements during borehole
pumping. Ground Water, 25:28-40.
Paillet, F. L., K. Novakowski, and P. Lapcevic, 1992. Analysis of
transient flows in boreholes during pumping in fractured
formations, Society of Professional Well Log Analysts 33rd
Annual Logging Symposium Transactions, S1-S22.
Paul, D. G., C. D. Palmer, andD. S. Cherkauer, 1988. The
effect of construction, installation and development on the
turbidity of water in monitoring wells in fine-grained
glacial till, Ground Water Monitoring Review, 8:73-82.
Redwine, J. C., R. R. Parizek, andF. J. Molz, 1990. Controls
on porosity and permeability in fracture-flow/conduit-
flow rocks of the Knox Group, Southern Appalachian
Fold-and-Thrust Belt, U.S.A., Southern Company
Sen-ices, Birmingham, Alabama.
Rehfeldt, K. R., L. W. Gelhar, J. B. Southard, and A. M.
Dasinger, 1989a. Estimates of macrodispersivity based on
analysis of hydraulic conductivity variability at the
MADE site, EPRI Interim Report EN-6405. Electric
Power Research Institute, Palo Alto, California.
Rehfeldt, K. R., P. Hufschmied, L. W. Gelhar, and M. E.
Schaefer, 1989b. Measuring hydraulic conductivity with
the borehole flowmeler, EPRI Topical Report EN-6511,
Electric Power Research Institute, Palo Alto, California.
Ruud, N.C., and Z.J. Kabala, 1996. Numerical evaluation of
flowmeter test interpretation methodologies, Water
Resources Research, 32:845-852.
Ruud, N.C., and Z.J. Kabala, 1997. Numerical evaluation of
the flowmeter test in a layered aquifer with a skin zone,
Journal of Hydrology', 203:101-108.
Sandal, H. M., R. N. Home, and H. J. Ramey, Jr., and J. W.
Williamson, 1978. Interference testing with wellbore
storage and skin effects at the produced well, Society of
Petroleum Engineers, 53rd Annual Technical Conference,
Houston, Texas.
Shapiro, A.M., and P.A. Hsieh, 1991. Research in fractured
rock hydrogcology: Characterizing fluid movement and
chemical transport in fractured rock at the Mirror Lake
Drainage Basin, New Hampshire, in U.S. Geological
Survey Open File Report. Proc. of USGS Toxic
Substances Hydrology Program Technical Meeting,
Monterey, California, March 1991, U.S. Geological
Survey, Reston, Virginia.
Sledz, J. J., and D. D. Huff, 1981. Computer model for
determining fracture porosity and permeability in Hie
Conasauga Group, Report ORNL/TM-7695, Oak Ridge
National Laboratory, Oak Ridge, Tennessee.
Strauss, M. K, S. L. Story, and N. E. Mehlhorn, 1989.
Applications of dual-wall reverse-circulation drilling in
ground-water exploration and monitoring, Ground Water
Monitoring Review, 9:63-71.
-------
Tongpenyai, Y., and R. Raghaven, 1981. The effect of storage
and skin on interference test data, Journal of Petroleum
Technology, 33:151-160.
van Everdingcn, A. E, and W. Hurst. 1949. The application of
Laplace transforms to flow problems in reservoirs,
Transactions of the American Institute for Min. Me tall.
Pet. Eng., 186:305-320.
Winter, T. C., 1984. Geohydrologic setting of Mirror Lake,
West Thorton, New Hampshire, U.S. Geological Survey
Water Resources Investigations Report 84-4266, U.S.
Geological Survey, Reston. Virginia.
Xiang, I, 1995. The evaluation of the flowmeter test in three-
layer aquifers and the influence of disturbed zones,
Journal of Hydrology, 1.66:127-145.
Young, S. C., 1991a. A site characterization methodology for
aquifers in support of bioreclamation activities. Vol. I:
Well network design, well equation, and aquifer multi-
well and single-well tests, U.S. Air Force Engineering and
Sendees Center, Tyndall AFB, Florida.
Young, S. C,, 199Ib. A site characterization methodology for
aquifers in support of bioreclamation activities, Vol. II:
Borehole flowmeter technique, tracer tests, geostatistics,
and geology, U.S. Air Force Engineering and Sendees
Center, Tyndall AFB, Florida.
Young S. C., H. S. Pearson, G. K. Moore, and R. B. Clapp,
1991. Application of the electromagnetic borehole
flowmeter technique at the Oak Ridge National
Laboratory. Report WR28-1-900-247, Tennessee Valley
Authority7, Morris, Tennessee.
Zemanek. J., E. E. Glenn, L. J. Norton, and R. L. Caldwell,
1970. Formation evaluation by inspection with the
borehole televiewer, Geophysics, 35:254-269.
56
-------
Appendix A
Field for Borehole Flowmeter Tests
TVA ENGINEERING LABORATORY
EM FLOWMETER FIELD DATA SHEET
Project Name / Project Abbr. (xxx)
Date Start Time End Time
Pre-Calibration Ref No. Flowm
Gain Switch Integration Switch
| | HI | | LOW I Ms CZ]lOs Q
Packer Collar
| | Yes | | No | | Yes |
Flowmeter Serial # Electronic Seria
I I Pump | | Injection I I Ambient P
Rate
QA Number (xxx....)
Survey By
eter Cal Flowmeter Zero
LPM/Volt Volts
Excitation Voltage
I] 20s VRMS
(Serial #)
| No
I # Cable Serial #
ump Intake Depth GW Temperature
Well Number: Depth to Water:
Top of Reference Elevation: Depth to Top of Screen-
Construction Tvpe: Depth nf Bottom "f fi
-------
Appendix A
Field for Borehole Flowmeter Tests (continued)
TVA ENGINEERING LABORATORY
SINGLE-WELL FIELD DATA SHEET
Project Name / Project Abbr. (xxx)
QA Number (xxx....)
Well Number
Survey By
Date
Pump Type:
Pump Serial #
Transducer Type:
Transducer Serial #
Flow Check
Time
Volume
Time to Fill
Flow Rate
Depth of Transducer:
Initial Transducer Reading:
Datalogger File Name:
Pumping Schedule
Time (hr, min) Rate (LPM)
1.
2.
3.
Drawdown Check
Time
Electric Tape
Depth
Datalogger
Reading Depth
Comments:
58
-------
Appendix B
Equipment Checklist
_ Address & phone number of contact
_ Site maps
_ Well specifications
_ Calibration sheets
_ Test forms
_ Equipment manuals
_ 1/2-inch flowmeter
_ 1-inch flowmeter
_ Flowmeter electronic system
_ Computer interface (6B12, HPIB)
_ Flowmeter spare parts
_ Flowmeter collar
_ Collar weight
_ Packer assembly
_ Packer inflator
_ Air compressor
_ Airtank
_ Pressure inflator tank
_ Inflation tubing, fittings, and valves
_ Flowmeter cables
_ 5 psi pressure transducer
_ 10 psi pressure transducer
_ 20 psi pressure transducer
_ 30 psi pressure transducer
_ Flowmeter computer
_ Electronic data logger
_ Printer
_ Computer, printer, transducer, & electronics cables
_ Cable for downloading data logger
. Floppy disks
_ Application and operating software
_ Flowmeter software
_ Datalogger software
_ Other software
_ Peristaltic pump
_ Submersible pump
_ Centrifugal pump
_ Other pumps
_ Valves
_ Inlet hoses
_ Discharge hoses
_ Checkvalve
_ Hose clamps and fittings
_ Pump calibration containers
_ Flowmeter calibration cylinder
Water containers
OSHA Training Card
Stop Watch
120 ACV generator
220 ACV generator
50 ft 120 ACV extension cord
50 ft 220 ACV extension cord
25 ft 120 ACV extension cord
Special extension cords
Ground fault box
Power strip
Gas can
Weight for steel tape
Engineering rule
Steel tape
Tools (electrical and mechanical)
Hack saw
Drill and bits
Tie wraps
Electrical tape
Flashlight
Electrical supplies (splicing kit, solder, wire, etc.)
Misc hardware (screws, bolts, eta)
Marking tape
Indelible pen
Rope
Tarpaulins
Canopy
Tent
Spring clamps
Parachute cord
Multimeter
Oscilloscope
Rain suit
Boots
Cold weather gear
Sun screen
Insect repellent
Cooler and drinking water
Stools, chairs
Portable table
Decontamination supplies (buckets, water, soap, brushes)
Rubber gloves
Work gloves
Rags, paper towels
First aid kit
59
-------
60
------- |