|Resistivity Meters & Imaging Systems|
|Magnetic Susceptibility Meters|
|Magnetic Resonance Systems|
Hickam Air Force Base,
Underground Storage Tank Project
Results of Geophysical Investigation
James C. Hasbrouck
John W. Dickerson
H. David MacLean
NOTE: The following is a summary of the Results of the Geophysical Investigation.
This paper can be found in its entiretyj at the National Technical Information Service
(NTIS) under the identification number DOE/ID/12584-37.
Magnetic total field, magnetic gradient, and horizontal
loop electromagnetic geophysical survey measurements were carried out at Hickam Air Force
Base for the purpose of identifying and determining the location of abandoned underground
storage tanks that might be in-place at several locations on the base. Steel underground
storage tanks are ferromagnetic, and are good electrical conductors; unexplained electric
conductors of the appropriate size that can be shown to be ferromagnetic might indicate
the presence of an underground storage tank.
The geophysical surveys were conducted at twenty-three
sites that had been targeted as a result of records examinations and interviews with
Hickam personnel. The magnetic and electromagnetic instruments responded
characteristically to the presence of known tanks. An equivalent magnetic susceptibility
of 0.08 cgs units, and a conductor EM response parameter of 100 to 200 were found to be
characteristic of underground storage tanks; other metal objects situated within the
survey areas also affected the instrument readings, but these objects usually exhibited
significantly different response parameters, or were otherwise identifiable by their
dimensions or depth of burial.
The geophysical survey results are presented through this
report. Data recorded at survey grid points are stored on magnetic media, along with the
appropriate location information. A summary of the data for each survey area is presented
as maps and profiles. Data deemed to be essential for interpretation purposes have been
presented in this discussion; other data can be reviewed by the interested reader by means
of the referenced surface contouring and profiling software.
The conclusions reached as a result of consideration of the
survey data discussed are summarized in Table 1.
In September, 1988, geophysical methods that could
determine the presence of buried underground storage tanks were considered for use at
Hickam Air Force Base (AFB). Survey procedures were proposed to test for underground tanks
that may have been abandoned at several suspected locations throughout the base, to
confirm the location of such tanks at other locations where records were thought to be
inadequate, and to confirm that tanks at certain other locations had been removed, as
indicated by existing records.
Subsequent to a visit to Hickam AFB and a review of the
applicability of geophysical methods, a program of surface geophysical surveys was
recommended. Steel storage tanks are strongly magnetic and highly conductive; accordingly,
a suite of magnetic and electromagnetic surveys was recommended. Details of the survey
methods and field procedures that were used are provided in Section 3 of this report.
Twenty-three sites on Hickam AFB were designated to be
investigated for the presence or absence of underground storage tanks. The sites and
survey area to be investigated were selected by the project manager on the basis of the
lack of definitive records of removal of tanks that were once located in an area, or by
the historical land uses at particular areas.
Field work for the geophysical surveys was conducted during
July and August of 1989. The magnetic surveys were conducted with a
magnetometer/gradiometer. An EM-31 induction conductivity meter (Slingram type EM
instrument) was used to collect electromagnetic data.
1.1 Location and Access
The geophysical work was carried out within the confines of
Hickam AFB. A of the base area is included as Figure 1. The locations of the survey grids
established to accommodate the geophysical work can be determined from Figure 1, using the
site location key (Table 2) and the index bars that appear on the map. Each letter and
number bar on Figure 1 can be considered as a bar scale, subdivided into 10 units, with 0
at the top or left of the bar. A site with an origin that projects left to the mid-point
of the bar, and upward to the midpoint of the "5" bar would be located at D-0.5.
5-0.5. Access to all sites was coordinated with Hickam AFB personnel; only areas 5 and 6
required access to a fenced and controlled enclosure.
2.0 Purposes of Surveys
2.1 Confirmation of removal of tanks
Field surveys were conducted at the following sites to
confirm that the tanks located in areas had been removed.
||Map Coordinates (fig.1)
||JR Tank Farm
2.2 Confirmation of Location of Tanks Abandoned in Place
Field work at the following sites was carried out to verify the
location of tanks that had been abandoned in-place after completion of accepted
||Map Coordinates (Fig.1)
||Bldgs. 1250, 1257
2.3 Investigations for the Presence of "Unkown"
Tests were conducted to determine whether tanks were present at the
||Map Coordinates (Fig.1)
3.0 Survey Procedures
3.1 Grid Preparation
A grid of lines with ten foot spacings, with stations marked along
the, lines at ten foot intervals, was established over the areas of interest. These grids
were laid out relative to prominent landmarks or structures within the survey area. The
survey grids are shown on detailed drawings prepared from Hickam AFB facility drawings.
3.2 Magnetic method
Magnetometer readings were taken at two measurement heights with the
magnetometer at all points on the survey grids established over each area of interest. The
magnetic field read at the upper sensor was recorded as the standard field strength; the
difference between the readings at the ripper and lower sensors provided a measure of the
vertical gradient of the magnetic field. Total magnetic field and gradient values were
posted for each survey point.
In order to verify data integrity and quality, magnetic data were
posted and contoured immediately after the survey of each area was completed. It is noted
without further comment that the gradients recorded were actually gradients per half
meter, since the sensor separation of the (38-06AX was 0.5 meters.
Variations in the intensity of the observed magnetic field at grid
points caused by diurnal changes in the earths magnetic field were corrected. The
change in magnetic field intensity with elapsed time at the base station was subtracted
from the synchronous field intensity readings obtained in the, survey areas. One of two,
centrally located, base stations was used for each survey at Hickam AFB.
The average magnetic intensity read at the base stations was 35300
nanoteslas (nT), which agrees well with values published on regional magnetic field maps.
Although the inclination of the magnetic field was not directly measured, an inclination
of 41°, rather than 39° value extracted from the NOAA maps, appears to be more
All data were recorded and checked to ensure that the readings fell
within the acceptable field standardization ranges currently used by UNC Geotech. A
description of the quality control and standard ecceptance procedures that were utilized
is provided in appendix D.
3.3 Electromagnetic Method
Electromagnetic (EM) field measurements were taken at grid
intersection points for all areas surveyed. Measurements were collected using vertical and
horizontal dipole configurations. Conductivity, as well as in-phase data were digitally
recorded using a Polycorder.
Measurements were initially made at intervals of 10 feet, which
suffices for anomaly recognition, but is only marginally adequate for definition of
complex EM profile shapes. Anomalous areas and other areas of interest were detailed using
a 5-foot measurement interval.
3.4 Data Sets
Complete magnetic and EM data sets exist for all survey areas, and
are included as data files. The entire set has not been reduced and displayed as maps and
profiles; only pertinent portions of the data set have been presented. For example, all of
the horizontal dipole EM-31 data have not been presented, only those profiles relevant to
an interpretation have been included. All of the data are of course available on disk
On occasion, the field crew found it necessary to conduct duplicate
"detail" surveys within a previously defined survey area. To the extent
possible, these "detail survey" data sets have been edited and integrated with
the first, or reconnaissance survey; however, inevitable errors in positioning and other
factors prevented this approach in all cases and separate "detail area surveys are
included for some locations.
Fig. 4.1-1a Magnetization of ferromagnetic
4.0 Theory fo Survey Methods
4.1 Magnetic Method
Ferromagnetic objects in the earths magnetic field become
magnetized. The induced component of magnetization causes an anomalous magnetic field in
the vicinity of the object, which can be measured with a magnetometer. By convention, the
direction of a magnetic field is the path that a free magnetic pole would travel.
Accordingly, the magnetic field in the northern hemisphere trends from south to north and
has an inclination of 90° at the poles and 0°) at the magnetic equator. At the latitude
of Hawaii, the mean inclination of the magnetic field is 39° .
A ferromagnetic body will be magnetized as shown in Figure 4.1-1a:
magnetic profiles that pass over, or off to the sides of the body will have the general
shapes shown in Figure 4.1-lb. The east-west profiles shown as Pl, P2, and P3, and the
north-south profiles shown as P4, P5, and P6 are fairly symmetrical, and easily
understood. Where the secondary field vector opposes the primary, the resultant field
measured at a point will be less than the primary, or anomalously low; where the secondary
field vector is in the same direction as the primary field vector, the fields will add,
and the measure field will be anomalously high.
Steel storage tanks are ferromagnet; large magnetic total field and
vertical magnetic gradient anomalies can be expected from such tanks. By their geometry,
tanks buried at a shallow depth (i.e., a depth that is much less than the long dimension
of the tank) will appear as horizontal cylinders; magnetically, they will resemble a line
of dipoles. Tanks buried at greater depths may appear as single dipoles; the line analogy
for the conditions set forth above, however, will remain valid.
The observed vertical magnetic gradient can also be used in the
interpretation of magnetic anomalies that might be related to tanks. Convenient
relationships exist between the gradient and a single or line of dipoles (Breiner. 1974).
For a line of dipoles or a horizontal cylinder,
For a single dipole,
||G is the magnetic gradient
||T is the total field anomaly
||D is the depth to the source
In instances where the sensor geometry does not support the gradient
equations above, such as when the 0.5 meter sensor separation approaches 0.2 times the
depth to the top of the tank, the depth can be calculated from the two magnetic readings,
gradiometer sensor separation
sensor field/upper sensor field)1/3 -1]
Though the precise theoretical magnetic response of a hollow
cylindrical tank has not been formulated for all situations encountered in this project,
it is possible to synthesize a solid body with an effective volume and equivalent
susceptibility by modeling the observed magnetic field variation at several known tank
locations. The permanent, or remnant magnetization of the tanks is included in effective
susceptibility calculations. Tanks of a similar type and composition that were buried for
similar lengths of time under identical conditions might reasonably be expected to have
similar components of remnant magnetization.
Fig.4.1-1b. Generalized magnetic response for
representative magnetic profiles.
Use of this equivalent magnetic solid greatly simplifies the task of
analysis and interpretation of anomalies from hollow steel cylinders because software for
modeling magnetic anomalies induced in the earths field by solids is readily
available. Treating the tank as a solid is an acceptable method of approximating the
magnetic field anomaly caused by a tank. For this project, it was assumed that the
cylindrical axis of symmetry of the tanks was horizontal.
The anomalous magnetic field expected over a solid ferromagnetic
body of uniform susceptibility can be calculated using commercial software, such as the
"MAGIX" package that was used for this study. The theoretical response along
east-west and north-south profiles over a solid body with the effective susceptibility of
a buried tank are shown in Figures 4.1-2a and 4.1-2b, respectively. The change in response
with depth along the north-south profile of a hollow cylinder is presented in Figure 4.1-3
(D Snyder, 1990) . Under field conditions where a high degree of noise can be expected,
the differences between the anomalous profiles over hollow and solid objects are slight.
Therefore, the representation of the buried tanks as solid objects is valid for the
purpose of this report.
During the process of interpretation, the magnetic response of
several bodies equivalent in volume to standard underground storage tanks were calculated.
Representative tank volumes and dimensions are summarized in Table 1. All equivalent
magnetic susceptibility was utilized for a solid body with dimensions of the known tanks
and the calculated magnetic profile over the tanks was compared to the observed magnetic
The models presented as Figures 4.1-2a and 4.1-2b for a 50,000
gallon tank suggest that the equivalent solid material susceptibility is on the order of
0.08 cgs units. The theoretical curves derived for these models are not greatly different
from the curve that was calculated for the hollow cylinder model that is shown in Figure
4.1-3, especially for cases where the tank is at a depth of one meter or greater. Because
the sensor of the magnetometer was three meters above the ground, and the tank assumed to
be at a depth of one meter below the surface, it is reasonable to use the simpler solid
model for predicting the response of a buried tank.
Fig. 4.1-2a. Magnetic response for an east-west
central profile that transects a solid magnetic cylinder.
Fig. 4.1-2b. Magnetic response for a north-south
central profile that transects a solid magnetic cylinder.
Fig. 4.1-3. Magnetic response for a north-south
central profile for the object modeled in Figure 4.1-2a treated as a hollow cylinder.
The terrain of Hickam AFB is extremely flat. Though it is recognized
that some errors will be introduced if parts of a particular traverse are at different
elevations, we are generally concerned only with interpreting those objects that are below
the line of traverse. For this study, a simplification has been made when comparing
theoretical to observed magnetic fields; all measurements may be considered to have been
made on a plane. The profile will be distorted at elevation changes, but only along the
elevation gradient. Any such distortion will be treated as a fitting error.
4.2 Electromagnetic Method
Steel tanks are very good electrical conductors. Such conductors can
be detected much as are electrically conductive ore bodies using standard electromagnetic
prospecting devices. The EM-31, though not designed specifically for underground tank
location, is generally suited to the purpose, and interpretation of the data can be
modified to be effective in this application.
The EM-31 is a slingram, or moving horizontal coil EM system. The
receiver portion of the system senses eddy currents induced in conductive media by the
transmitter portion of the system. These signals can be characterized in terms of the
size, conductivity, and location of the conductor. The basic theory of operation of this
type of system is provided in standard geophysical text books. A good introductory
reference is provided by Parasnis (1966).
The theory of electromagnetic induction, and of the operation of the
EM-31 is important to the understanding of how anomalies encountered in the of this
project have been interpreted and categorized. Accordingly, a brief outline of the
principles involved in the application of the, instrument to tank detection follows.
A typical EM system array is shown in Figure
4.2-1. An alternating current flowing in the coil marked "Tx" produces an
alternating magnetic field, in turn induces eddy currents in any nearby conductors. The
secondary magnetic field (S) has the same frequency as the transmitter, but i out of
phase. The secondary field is sensed by the receiver coil ("Rx") along with the
primary field (P) from the transmitter. The field at the receiver a resultant, or vector
sum of the two components. Elementary considerations of EM theory require that the
secondary field component lag the primary by no less than 90° ; if the secondary field
comes from a very good conductor, the lag will be almost 180°.
Fig. 4.2-1. Typical Horizontal
loop EM configuration
Fig. 4.2-2 Vector Diagram of EM responses.
The various vectors mentioned are represented in Figure 4.2-2. To
the extent that the vector S, produced by conductors of interest, can be measured and
quantified, certain parameters of the conductor can be determined. It should be noted that
when the receiver and transmitter are on the same side of the conductor, the vectors P and
S will add, and the resultant field is greater than what is measured in the absence of a
conductor; when they are, on opposite sides, the vector S opposes P and a
"negative" anomaly results.
Fig. 4.2-3a. Typical in-phase EM response over a
The magnitude of the vector S can be described as a percentage of
the primary field P from the transmitter. The phase angle, a (Figure 4.2-2), represents
the difference of phase of the resultant vector, relative to the primary. A component of
the vector S is represented by the projection of S onto the horizontal axis of the figure.
This component, (S sin theta) is 180° out of phase with the primary, or transmitter
field; this component is the in-phase or real component, as it has the same phase sense as
the primary field. The projection of S onto the Y axis of the figure, (S cos theta),
represents a vector component of S that is 90° out-of-phase with the primary, this is the
quadrature (imaginary) or out-of-phase component. The in-phase and out-of-phase
components, expressed as a percentage of the primary field, are convenient forms of
describing the magnitude and phase relationship of the secondary field vector S.
For tank location applications, the conductor of Figure 4.2-1 would
be the tank. A typical buried storage tank might be considered as a conducting sphere, or
as a thick tabular body. From an analysis of the vector S, the quality, size and position
of the conductor can be determined. If the electrical parameters determined fall in a
range known to be compatible with underground storage tanks, the presence of a tank might
be inferred from these factors.
The response to a permeable conducting sphere with a given radius
(a), by a Slingram system with variable ratios of depth (z) versus coil separation has
been well documented in the geophysical literature (Fuller, 1971, Rai & Verma, 1982,
and Wait, 1960), as has the EM response of a thin sheet in a uniform and a layered
conducting half space (Hanneson and West, 1984). Reasonable estimates of the response of a
hollow conducting cylinder can be made from theoretical work and empirical observation.
Typical in-phase responses over a sphere are shown in Figure 4.2-3a
and 4.2-3b. A typical response for a plate is shown in Figure 4.2-4. It should be noted
that the response of both plates and spheres is influenced by the conductivity of the half
space in which the conductor occurs, and by the permeability of the conducting body. The
relationship is very complex; the interested reader is referred to insightful articles by
Hanneson and West (1984), and by Rai and Verma (1982) for clarification.
Fig. 4.2-3b. Effects of increasing permeability on
the in-phase EM response over a sphere.
The host or half space conductivity is an important consideration,
and cannot be dismissed. The conductor response, RC (conductor), is affected by the host
conductivity, Rh. When evaluating conductors, the following EM responses for a two coil EM
system are pertinent:
|| = µwshs2
||= µwscts2 for a plate
||= µwsca2 for a sphere
| µ = 4n10-7
||w = 2nf (f=frequency), and
||sh is the host conductivity
||sc is the conductor conductivity
||s is the coil separation
||t is the thickness of a plate conductor
||a is the radius of the sphere (in coil length)
A tank, whether treated as a tabular body, a plate, or as a sphere,
will have a characteristic response parameter, RC, that can be used to identify and
characterize Unknown conductors.
An important aspect of EM profiles is their behavior with respect to
increasing depth to the target. Unlike magnetic anomaly profiles, the amplitude of the
anomaly is affected by depth, not width. This feature is illustrated by Figures 4.2-3 and
4.2-4. The wavelength, or width of an EM anomaly is controlled by the coil spacing. All of
the anomalous response ocurrs within four coil lengths of the EM system, in the case of
the EM31, this distance is 48 feet.
The EM-31 device employs two coplanar coils that are normally in a
horizontal plane (vertical dipoles), but the coils can be oriented vertically as well,
providing a horizontal dipole configuration. Since the response of Certain types of
conductors to the two types of coil configurations is significantly different, information
concerning the conductor can be determined from measurements with the different coil
The depth of investigation of an EM system is dependent upon the
coil separation, operating frequency, and host conductivity. The two coils of the EM-31
are set 12 feet (3.66 meters) apart and the instrument operates at a frequency of 9.8 kHz.
As such, the maximum depth of investigation is approximately 20 feet (6 meters), which is
in the range of the depth of burial of most underground tanks.
The EM-31 measures both conductivity and in-phase component values.
Conductivity is linearly related to the out-of-phase component. The
out-of-phase component value, in percent of the primary field, is 0.23 times the
(conductivity value, in milli-Siemens per meter (McNeill, 1980). The in-phase component
value is measured as a percentage of the primary field.
Although EM theory is complicated, a few
simplifications can he made that will ease the understanding of the method in the tank
detection application. Under certain restricted conditions (McNeil, 1980) the ratio S/P is
a function of the coil spacing, the frequency and the soil conductivity. Since the first
two parameters are fixed, a linear relationship exists between the ratio S/P, and the soil
conductivity. Though it is convenient in many applications to determine the soil
Conductivity directly, the response of a tank may not approximate a layer of uniform
conductivity; indeed, i tank is best modeled is a sphere. The soil conductivity is
relevant only to the extent that this parameter affects the EM response of the target
within the soil (i.e., RI, versus RC). The commonly used secondary field in-phase and
out-of-phase components must be determined in order to interpret the instrument response
in terms of plate, sphere or tabular conductors. McNeil (1980) has shown the relationship
between the out-of-phase component of the secondary field and the undisturbed primary to
s/p = µws2s
Since the conductivity is read in milli-Siemens/meter, the
out-of-phase component of S is simply 0.281s, in tenths of a percent, or in thousands of parts per million.
When the response at the receiver indicates the presence of a
conductive body, the response parameter, RC, determines whether or not the response falls
within a range that is indicative of a tank. Published values of these parameters for
hollow spheres or cylinders are not available. However, during the course of this
interpretation effort, RC parameters were obtained for a number of tanks. A RC of 100 to
200 ohm-meters- appears to be reasonable for tanks; a permeability of 2 was commonly
The tanks that are the subject of these investigations probably lie
above the water table, since they would not have been installed underground if the water
table were high. It is assumed then that the tanks are conductive lying in a relatively
nonconductive environment. To the EM-31, a tank will resemble a thick, highly conductive
material. Although the tank is hollow, it will appear as a highly conducting solid sphere
or oblate spheroid. At the operating frequency of the EM-31, electrical currents flow as
sheets at the outer edge of the conductors. Large tanks, with a diameter of 4 meters or
more, may resemble a conducting layer, rather than a plate or sphere. In this case, the
response parameter RC will still be of the order used for tanks.
Anomalies in the conductivity profile will resemble the out-of-phase
portion of familiar horizontal loop anomalies. Because the tank conductors are thick
relative to the coil spacing, the negative going portion of the curve may not always be
observed. The anomalies may resemble those that are associated with airborne or helicopter
surveys, more than standard ground horizontal loop EM surveys. Figure 4.2-5, adapted from
Fraser (1979), illustrates disappearance of the negative portion of the anomaly curve for
a whale tail, or horizontal loop EM system. It should be noted that the anomalies
expressed in parts per million, rather than percent, for conductors that are at a depth of
2 to 3 times the coil spacing.
5.0 Interpretation Procedure
5.1 Magnetic Method
Magnetic anomalies with the general shape and pattern shown in
Figure 4.1-lb are identified Central azimuth and transverse profiles are then compared to
a theoretical anomaly from a modeled solid with the approximate dimensions of a buried
tank, and with an effective susceptibility in the appropriate range. Other profiles are
then compared to the models in Figure 4.1-lb to ensure that they are consistent with a
tank. The edges of the tank can be located by peaks in the gradient profile; this
parameter can also be used to estimate the depth to the top of the magnetic source. If the
model developed from this inversion process is reasonable for a tank, an indicator of the
presence of a tank has been established.
The magnetic response of a large ferromagnetic object like a steel
tank is highly characteristic. There are few geological features that could resemble a
steel tank; however, other buried steel objects could produce similar anomalies, and much
larger geologic bodies with much lower susceptibility can also approximate a given
5.2 Electromagnetic Method
Profiles of the in-phase and conductivity measurements are examined
for anomalies. When an anomaly is recognized, a response parameter, Rc, is determined from
curves of the type published by Rai and Verma (1982), and Hanneson and West (1984). A
depth to the conductor, in terms of coil separation is then determined, along with the
position of the edges of the conductor. The radius and response parameter of a conducting
sphere that could produce the same anomaly are also determined from Argand diagrams
developed for conducting spheres. A tank might be inferred from survey results that match
As mentioned above, it is convenient to plot conductivity profiles,
and convert the measured anomalies to out-of-phase for use on the Argand diagrams in-phase
is read directly. It is also convenient to determine the host response directly from the
instrument, without recourse to Argand diagrams for half spaces. Knowing sh, the host
response parameter (Rh) can be readily calculated.
Fig. 4.2-5. Typical horizontal loop EM curves,
adapted from Fraser, 1979.
5.3 Integration of Data
The interpretation of the presence of a tank does not rely on the
analysis of a single set of either magnetic or electromagnetic data; the strongest
indications of an underground tank occur when both data sets suggest the same
interpretation. A conductor with tank characteristics is unlikely to be a tank unless it
is magnetic, with a susceptibility in the range that has been empirically determined for
tanks. Other information, such as historical records, must also be considered in the
evaluation of the data sets.