13th EEGS Symposium on the Application of Geophysics to Engineering and Environmental Problems

The sandstone aquifer is the major source for ground water in eastern Wisconsin. Over
the last decade, salinity has risen in several wells and several new wells have encountered
saline water in areas thought to contain fresh water. TEM surveys are being used to map
the location of saline water zones, as a tool to select new well locations, and to develop
better management methods for the aquifer. This paper presents a case history for a small
village which unexpectedly encountered poor water quality in a new municipal well. A
TEM survey suggested that the poor water quality was due to a stagnation zone behind a
previously unknown mound on the Precambrian surface. A new well location was
selected on the basis of the survey results. In addition, a regional TEM survey is being
completed for the western suburbs of Milwaukee to determine distribution of saline water
in that portion of the aquifer.

Developing technologies in UXO investigations requiring mapping, are utilizing Differential Global Positioning Satellite (DGPS) techniques to improve the quality of data collection. Additional benefits include accurate geographic referencing and anomaly re-acquisition during intrusive stages in the investigation. Until recently, differential GPS geographical referencing was typically merged with geophysical data during post processing allowing the user to view results without the capability to analyze the data with regard to Geographical Information Systems (GE) on larger scales. Recent developments to improve UXO geophysical investigations include multiple array sensors connected to real-time kinetic (RTK) GPS systems via a lap top computer. This method allows real-time data analysis, which reduces and/or eliminates much post processing stages of data analysis. The data collected by the RTWGPS and multiple array system is simultaneously imported into a spatial data collection and management product designed specifically to work with relational databases to support GIS applications. Since the goal of most UXO investigations involving map generation is to apply GIS techniques in risk assessment scenarios, rapidly created and adaptable databases are highly desirable. The integration of system components in real-time mode not only saves time in the analysis of the data, but also reduces the cost of the investigation and simultaneously enters data into a GIS database.

Electromagnetic methods provide a relatively quick way to survey a large area for UXO but
often suffer from a high rate of false detections due to clutter. Seismic methods detect the
mechanical properties of objects below the surface, providing an independent “look” and another
dimension in measurement space. The merging of the mechanical response with the
electromagnetic response may aid in target classification and a reduction in the number of
expensive excavations.
In this SERDP funded effort, we approach the problem from a systems point of view. The goal is
to estimate the excess SNR for a given target/environment combination, taking into account the
source characteristics, source coupling, forward propagation, target response, return propagation,
reverberation, and receiver array gain. Models and measurements of each of these factors aid in
the design of the prototype system. For example, finite element modeling of a 155~mm shell in a
realistic soil indicates a target resonance at about 1600 Hz. The target response characteristics, as
well as the target imaging, may thus be useful features, which can be merged with the EM data.

The objective of UXO investigations is to minimize risk related to the location,
characterization and removal of unexploded ordnance. Successful UXO remediation is a
function of two distinct processes - knowledge building and decision making - which in
turn, commonly rely on spatial data (i.e. located measurements) acquired from surface,
marine or airborne geophysical surveys.
Initially we rely on technical “knowledge experts” to acquire and transform spatial data,
such as magnetic or electromagnetic measurements, into information and combined with
their knowledge, be used for UXO decision-making and actions. From a risk
minimization perspective, the knowledge expert acts as the “custodian” for spatial data -
maintaining the integrity of original measurements, and recording all assumptions and
transformations from start to finish.
As projects advance, we rely on technical (or non-technical) “decision-makers” to review
results and recommendations, question their validity and direct actions. From a risk
minimization perspective, the decision-maker is the “arbitrator” over spatial data -
ensuring that all conclusions and recommendations are validated or at least understood in
the context of the original spatial observations.
In this paper, we examine the roles of spatial data, knowledge building and decisionmaking
in UXO risk minimization via the conceptual framework of the Integrated
Knowledge / Decision (IKD) model. The IKD model presented here describes knowledge
building and decision-making processes and the role of spatial data as the nucleus around
which all knowledge and decisions are formed.

We report the results of tests where electrical impedance tomography (EIT) was used for
detecting and locating buried unexploded ordnance (UXO). The method relies on the
polarization induced at the boundary between soil and buried metal to produce a
measurable phase difference between the injected current and the measured voltage.
When natural sources of induced polarization are smaller than those due to buried metal
objects, then tomographs of impedance phase indicate regions were metal-soil
polarization may be present. Relatively large negative phases may indicate regions were
the buried UXO is located.
Unexploded ordnance is typically detected using magnetic surveys or conventional metal
detectors. These techniques provide limited information regarding the depth of burial of
potential targets and are adversely impacted by the presence of metal objects near the
surface, such as fences, building foundation, or buried utilities. The EIT method can
provide depth and position information on objects located below the surface, and can be
deployed around buildings, providing information regarding what lies beneath them.
Two controlled tests were performed at a field site containing UXO buried in known
locations. Both tests produced a phase anomaly of about 20 milliradians, which closely
matched the known location of buried UXO objects.

Cost effective cleanup of lands contaminated with buried UXO requires significant advances
in our current capability to discriminate and identify UXO in cluttered environments.
Recently completed field tests have demonstrated that frequency domain sensors, such as
Geophex Ltd’s GEM-3, can reliably separate UXO targets from clutter objects based on their
complex broadband EM signatures. These field tests have also identified a number of areas
where improvements are needed before this technology is transitioned to full-scale UXO
cleanup applications. This paper describes our progress in addressing these needs, including
(a) improving our understanding of the GEM-3 phenomenology, (b) the development and
validation of physics-based analytical and numerical models of UXO target responses, and
(c) the development of improved multifrequency EM sensor prototypes. Recent
enhancements to our multifrequency EM data collection, visualization, and analysis systems
are described.

Currently, most unexploded ordnance (UXO) remediation is carried out with magnetic and
electromagnetic induction (EMI) sensors. While highly effective in detecting metallic objects
such as UXO, present field techniques also result in many false targets from metallic scrap. To
reduce the cost of digging non-UXO, discrimination techniques are required. One approach to
UXO discrimination is to recognize features from broadband EM1 data that reflect the shape of
the target only, while filtering out other features which may relate to target depth, orientation,
sensor-dependent signals, or combinations of these factors. A thorough calibration of the sensor
against targets of known shape and material properties is required for proper interpretation of
field data. Toward this goal, controlled measurements were made using the GEM-3 (FDEM)
sensor on spherical conductors of various sizes at several distances. These data generally
compare very well against the analytic solution for a sphere in a spatially uniform, time varying
magnetic field, despite the fact that the GEM-3 sensor produces a primary field that is not
spatially uniform.

Although the concept and use of soil-cement for ground improvement, excavation
and liquefaction migitation has been available for several decades, a reliable, routinely applied
Non-Destructive Testing WT] method has yet to be defined. This thesis looks at two
approaches for examining the in-situ material as to its Unconfined Compressive Strength PCS].
The dynamic properties of the mixtures are examined by measuring shear wave velocity. Cored
samples were taken from the material at the 28-day mark for laboratory testing of UCS. These
results were correlated to the measurements made by the NDT device.

A 1.5 GHz ground-coupled antenna was used with a digital ground penetrating radar system to
evaluate the amount of deterioration within an aboveground concrete holding tank. Because of
the corrosiveness of the solution held within and the structural design of the 72 year-old tank,
deterioration could manifest itself by delamination and/or by surface cracking. On the GPR
record, potential areas of deterioration appear as zones of attenuation. Delamination is most
likely to occur at the inner wall where the tile and concrete meet, but can also occur at either the
bottom or top rebar schedules within the concrete. Surface cracking can indicate both a suticial
stress problems, caused by the elliptical shape of the structure, as well as more severe
voiding/delamination problem. Over 50 vertical profiles were conducted on the 17-foot high
walls using GPR to more accurately assess the deterioration associated with failure of the
structure’s integrity.
Attenuation, measured as dB loss relative to the transmitted pulse when the antenna is coupled to
the concrete surface, was mapped and contoured for the tile/concrete boundary and the upper
rebar schedule. Deterioration at the concrete surface was achieved by calculating the real
concrete dielectric permittivity (i.e. dielectric “constant”) from the reflection coefficient of the
surface reflector when the antenna was mounted on a 12” thick foam block. Attenuation and
dielectric information were then compared with visual observations of the data to determine the
overall deterioration of the structure. Overall, the curved walls revealed more deterioration, over
25%, while the straight sections of wall had about 10 to 14% deterioration.

Marine geophysics has played an important role in a lo-year investigation and repair program at
the Ludington Pumped Storage Plant, an 1872 megawatt hydroelectric pumped storage facility
located on the eastern shore of Lake Michigan (Figure 1). The upper reservoir is 842 acres in size
and lined with clay on the bottom and lower inside slopes. Excessive seepage was noted in 1973
after the reservoir was filled for the first time. In 1975, trench-like features were first observed
extending through the clay liner. Geophysical and diver investigations later found that these
“trench features” ranged from four to ten feet in width, three to forty feet in depth, and hundreds
of feet in length (Figure 2). A geophysical testing program conducted in 1990 showed that sidescan
sonar and fathometer surveys combined with an accurate vessel positioning system were the
most effective methods to map the trench features. Annual side-scan sonar and fathometer
surveys have been performed since 1992 to assess the current status of the trench features and
provide a reliable map for diver inspections and underwater repairs. The digital trench feature
map is used as control for the real-time positioning of the boats and divers using differential GPS
and ultra-short baseline (USBL) tracking systems, respectively.