Sei sulla pagina 1di 8

EXPLORING DIRECTIONAL DIFFERENCES IN

RESISTIVITY RESULTS IN KARST

Ronald Manney, Lafayette College, Easton PA


Mary J.S. Roth, Lafayette College, Easton, PA
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Jonathan E. Nyquist, Temple University, Philadelphia, PA

Abstract

We are involved in a multi-year study to evaluate electrical resistivity as a tool to characterize


shallow karst. As part of this study, we have conducted a 56-line multielectrode resistivity survey (28 x
28 grid of tests) and a 12-line azimuthal survey on a site with a known cave and other karst features.
(The cave is located at the approximate center of the grid and at the center of the azimuthal tests.) In the
inversion results for these lines, the air-filled cave appears to be a high resistivity feature when the tests
are run perpendicular to geologic strike and a low resistivity feature when the tests are conducted
parallel to strike. We have used data from borings taken at the site and an understanding of karst
geology to develop a subsurface model that results in similar anomalous results when used as data for a
resistivity forward modeling program. We conclude that the fracture patterns in the rock and water in
the vadose zone create the conditions causing the anomalous results observed and propose an
investigation approach that would help identify such conditions in the field.

Introduction

Ongoing research at Lafayette College and Temple University has focused on the evaluation of
the reliability of electrical resistivity surveys as a site characterization method for geotechnical projects
in karst terrains. This research has included both 2D and 3D multielectrode arrays to investigate
bedrock depths and to locate karst solution features (Jenkins and Nyquist, 1999; Mackey et al., 1999;
Maule et al., 2000; Roth et al., 2000, Roth et al., 2002). The research area is at Metzgar Field, an
athletic complex owned by Lafayette College. The geology of the complex is characterized by having a
thin mantle of clay soils underlain by limestone (Epler Formation, lower Ordovician). On average, 15
new sinkholes open each year in the complex (Chen and Roth, 1997). The study area is approximately a
one hectare portion in the northwest corner of the site. Previous testing has shown that the bedrock in
the test area has multiple higher-resistive ridges that run along geologic strike with bedrock depth
varying from <1 m to more than 10 m. The bedrock is also known to have soil-filled and open fractures
as well as larger voids.

The Cavity

In 1999, a large anomaly was detected at the site using 2D electrical resistivity measurements
and a boring confirmed the presence of a large void beneath approximately a half meter of soil cover and
seven meters of bedrock. In 2002, a search-and-rescue camera was inserted into the boring and showed
that the void was a large, air-filled cavity with a moist, loose-gravel floor (Figure 1). In 2003, four
borings were drilled into the cavity to allow the search-and-rescue camera and two lasers to be inserted
into the cave to triangulate the cave’s interior dimensions (Roth et al., 2004). The triangulation of the

1117
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 1. Picture taken from interior of cave. Cave is air filled with loose material on the floor.

Figure 2. Point data collected using laser triangulation and a downhole camera (Roth et al., 2004). The
x and y axis of the graph are in PA State Plane Coordinates (ft) and show the orientation of the cave to
be relatively close to geologic strike of N60E.

1118
interior walls gave a series of points (Figure 2) that could be graphed to give the dimensions of the cave.
It was determined from the measurements that the long axis of the cave followed relatively closely to the
known geologic strike of N60E, the long axis of the cave measures approximately 4 m in length, and the
height of the cavity ranges from 0.5 m to nearly 2 m tall.
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

A Problem Arises

To further study the cave, a grid of 56 resistivity lines (28 x 28 lines) was conducted in 2003
centered over the cavity. The tests used 28 electrodes at 3 m spacing and a dipole-dipole array. The
grid was set so that one series of lines ran parallel to geologic strike while the other set of lines was
perpendicular to strike. The results of the tests were analyzed using EarthImager 2D produced by
Advanced Geosciences, Inc. Figure 3 shows the results from the line run perpendicular to strike directly
over the cave at the 45 meter mark and the results from the line run parallel to strike directly over the
cave at the 40 meter mark. It is clear from the results of these two lines that there is a major discrepancy
in the resistivity values of the cavity. In the perpendicular to strike direction the cavity is shown as a
highly resistive body. In the parallel to strike direction the cavity is shown as a low resistive body.
Because the cavity is known to be air-filled, the result from the perpendicular strike test appears to be
more accurate. However, we wanted to understand how the same feature could generate both sets of
data. Our theory is that water is trapped by surface tension in the small cracks and fissures above the
cavity, and this creates an electrical “superhighway” for the current. In addition, because the cavity is
elongated in the direction of strike, the effect on the results is more pronounced in that direction.
The idea of moisture trapped above a void is a well-understood concept and such a void is
typically referred to as a capillary barrier in the literature. The key to understanding this scenario is that
the trapped moisture is under unsaturated conditions, so that the pore pressure is negative (under
suction) due to surface tension/capillarity. Because the cave is at atmospheric pressure, water will never

NS14
South Cave North

West
EW16 East
Cave

Figure 3. Two-dimensional dipole-dipole resistivity inversions for the test perpendicular to strike
(NS14) and the test parallel to strike (EW16). (The horizontal dimensions give in this figure and all
other figures in the paper are in meters.)

1119
flow from lower to higher pressure so it won't go into the cave. This will leave a zone of relatively wet
bedrock (but not saturated) above the cave because it cannot drain freely downward.

Modeling to Test the Idea


Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

To test the viability of our hypothesis, we developed forward models using the Survey Planner
option in Earth Imager 2D. Survey Planner allows the user to create a subsurface using meshes of
varying resistivity values and run a mock test over this model subsurface. The mesh size and
dimensions depend on the forward modeling settings as well as the array type used to test the generated
subsurface. The settings were kept the same in the analysis of the forward models as were used for the
inversion of the field data shown in Figure 3. In the field and the forward modeling analyses we also
used a robust inversion method after Dahlin and Zhou (2004). Dahlin and Zhou found that the robust
inversion worked best for imaging sharp transitions in resistivity as demonstrated by inversions of a
waste pond simulation.
The limitation to this approach is that we are trying to model a three-dimensional problem in
two-dimensions. In the two-dimensional models the anomaly is assumed to be infinite in the direction
perpendicular to the model and that is clearly not the case in our field situation. However, because most
researchers and practitioners are more familiar with two dimensional tests and forward models, we
chose to use this approach first to check our theory. Furthermore, even for lines parallel to strike where
the 2D assumptions are clearly suspect, an air-filled cavity in limestone bedrock should appear as a
resistor, not a conductor.
The model subsurface was setup using typical values of the resistivity of the limestone and clay
obtained from tests at the site. The resistivity of the limestone was selected to be 2500 Ω-m and the clay
resistivity was set at 150 Ω-m. The cavity was modeled at a value of 100,000 Ω-m and the area of
trapped moisture above the cave was modeled 350 Ω-m.
Initial geometries were chosen to model a narrow cavity with a narrow zone of trapped moisture
(Figure 4) to represent the condition perpendicular to strike and a wide cavity with corresponding wide

Figure 4. Subsurface model of narrow cavity (in NS direction, perpendicular to strike) with inversion
results.

1120
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 5. Subsurface model of wide cavity (in EW direction, parallel to strike) with inversion results.

zone of trapped moisture (Figure 5) to represent the condition parallel to strike. The inversion results for
each of these models are also given in their respective figures.
Figure 4 shows that perpendicular to strike, the narrow dimension of the cave, the effect of
moisture above the cave is barely discernable. However, parallel to strike, the long dimension of the
cave (Figure 5), the moisture creates a break in the resistive ridge. The resistive cave buried in resistive
bedrock produces virtually no effect, especially as the conductive clay layer at the surface limits current
penetration, it is the disturbed area above the cave that is critical for detection

Modeling Field Results

Once we understood the effect the trapped moisture above a cave can have on the resistivity
results, we attempted to use forward modeling to develop models of the subsurface at the test site to
match the inversion results shown in Figure 3. It should be noted that our modeling aimed at matching
the inverted resistivity data sections and not the pseudosections or the data directly. This made
comparison between field and model results easier, but apart from the basic problem of non-uniqueness
inherent in all resistivity modeling, matching the inverted section made from the data rather than the
data directly assumes that the inversion is free of artifacts. Two-dimensional models of the subsurface
were developed for both the perpendicular and parallel to strike directions. The models were developed
using geometric information obtained from previous borings that had been conducted at the site. For
simplicity, we limited the number of resistivities used in the model. Bedrock resistivities were again
modeled at 2500 Ω-m, the soil was modeled at 150 Ω-m, the void at 100,000 Ω-m, and areas of trapped
water at 350 Ω-m.
The model for the perpendicular direction is shown in Figure 6 with its inverted results and the
inverted results from the field. The model shows the narrow cave with the trapped water zone as well
as the two bedrock valleys located on either side of the ridge with the cave. The preliminary geometries
for the valleys were developed from the boring results but were then adjusted to create a better fit
between the model inversion and the field inversions.
The model for the direction parallel to strike is shown in Figure 7 with its inverted results and the
inverted results from the field. This model shows more complexity than the previous model. In
particular, while the cave is shown with the trapped moisture above, the soil layer of the model above

1121
the cave was modeled using a slightly higher resistivity. This was done to approximate the very shallow
depth to bedrock observed during boring operations in the area above the cave. In addition, at the 18 m
location, a zone of 350 Ω-m was included to model an area of fractured bedrock. This area was
discovered during a smoke test of the cave conducted during the summer of 2004 (Nyquist et al., 2005).
During subsequent drilling at that location, the rock was found to be significantly fractured.
A comparison of the results for the models and the field results shows many similarities but it is
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

obvious that the models do not fully capture the complexities of the site. This is to be expected due to
the simple resistivity models used (with only four resistivity values). However, the results shown Figure
6 indicates that the model is a good representation of the field with the bedrock surface (its ridge and
valleys) clearly shown and the cave is not evident. A comparison of the inversions of the field results
and the model results in Figure 7 also indicates a relatively close fit. The location of the cave is
indicated by the area of lower resistivity associated with the trapped moisture and the low resistivities at
18 m match with the fractured area. In this direction, the field results do show a decrease in resistivity
values with depth. This decrease is most likely due to current flow in the adjacent soil valleys, a three-
dimensional effect that cannot be reproduced using two-dimensional models.

Inversion of field results (NS14)

Inversion of model results

South Model North

Figure 6. Model of perpendicular to strike direction shown with inversion results from model and field
results.

1122
Inversion of field results (EW16)
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Inversion of model results

Model
West East

Figure 7. Model of parallel to strike direction shown with inversion results from model and field
results.

Conclusions

Trapped moisture above an elongated subsurface void in the vadose zone can cause significant
directional variations in the resistivity results obtained above the void when modeled in two dimensions.
Simple forward models have been used to demonstrate the effect and more realistic models based on
observed field conditions have been developed to illustrate how the presence of trapped moisture may
also be an explanation for the data obtained at the Metzgar Field test site. The modeling is only the first
step in testing our hypothesis. We plan to run moisture profiles in the borings above the cave to confirm
our theory.
Because the depth of investigation of the tests run at the Metzgar site is in the range of only 7 to
8 meters, the presence of the trapped moisture above the cave and the difference in the resistivity results
from the parallel and cross strike tests provide the greatest evidence for the presence of the cave in the
data obtained for this site. Had tests only been conducted in lines perpendicular to strike, the presence
of the bedrock ridge would have been evident, but the large void may have gone undetected. Therefore,
whenever possible, it is recommended that resistivity lines be conducted in both the strike and cross
strike directions. If significantly different resistivity values are obtained in the vadose zone at a given

1123
location, a reasonable hypothesis might be that a similar zone of trapped water is present and that a
larger void may be located at depth. The modeling also suggests that it is the disturbed "halo" around
the cavity that provides the best hope for detection, as it is likely that an air-filled void in resistive
bedrock would otherwise be missed.
Downloaded 02/07/16 to 132.239.1.230. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

References

Chen, L., and Roth, M.J.S., 1997, Sinkhole case study: Athletic fields, Lafayette College, Easton,
Pennsylvania, in The Engineering Geology and Hydrogeology of Karst Terrains, B. Beck and B.
Stephenson, eds., Balkema, Rotterdam, 37-40.
Dahlin, T., and Zhou, B., 2004, A numerical comparison of 2D resistivity imagining with 10 electrode
arrays, Geophysical Prospecting, 52, 379-398.
Jenkins, S. A., and Nyquist, J. E., 1999, An investigation into the factors causing sinkhole development
at a site in Northhampton County, Pennsylvania, in Proceedings of the Seventh Multidisciplinary
Conference on Sinkholes and the Engineering and Environmental Impacts of Karst, 45-49.
Mackey, J.R., Roth, M.J.S., and Nyquist, J.E. (1999). “Case study: Site characterization methods in
karst,” in Geo-Engineering for Underground Facilities, in Geotechnical Special Publication No.
90, G. Fernandez and R. Bauer, eds., ASCE, Reston, VA, 695-705.
Maule, J., Nyquist, J.E., and Roth, M.J.S., 2000, A comparison of 2-D and 3-D resistivity soundings in
shallow karst terrain, Easton, PA, in Proceedings of SAGEEP ’00, Symposium for the Application
of Geophysics to Environmental and Engineering Problems, Eng. Environ. Geophys. Soc., 969-
977.
Nyquist, J. E. and Roth, M. J. S., 2003, Application of a downhole search and rescue camera to karst
cavity exploration, in Proceedings of SAGEEP ’03, Symposium for the Application of Geophysics
to Environmental and Engineering Problems, Eng. Environ. Geophys. Soc., published on CDrom,
8 pp.
Nyquist, J.E., Roth, M.J.S., Henning, S., Manney, R., and Peake, J., 2005, Smoke without mirrors: A
new tool for the geophysical characterization of shallow karst cavities, Proceedings of SAGEEP
’05, in Symposium for the Application of Geophysics to Environmental and Engineering
Problems, Eng. Environ. Geophys. Soc.
Roth, M.J.S., Mackey, J.R., Mackey, C., and Nyquist, J.E., 2002, A case study of the reliability of multi-
electrode earth resistivity testing for geotechnical investigations in karst terrains: Engineering
Geology, 65, 225-232.
Roth, M.J.S, Nyquist, J.E., and B. Guzas, B., 2000, Locating subsurface voids in karst: a comparison of
multi-electrode earth resistivity testing and gravity testing, Proceedings of SAGEEP ’00, in
Symposium for the Application of Geophysics to Environmental and Engineering Problems, Eng.
Environ. Geophys. Soc., 359-365.
Roth, M. J. S., Nyquist, J.E., Faroni, A., Henning, S., Manney, R., and Peake, J., 2004, Measuring cave
dimensions remotely using laser pointers and a downhole camera, Proceedings of SAGEEP ’04,
in Symposium for the Application of Geophysics to Environmental and Engineering Problems,
Eng. Environ. Geophys. Soc., published on CDrom, 8 pp.

Acknowledgements
The authors wish to gratefully acknowledge the support of the National Science Foundation
(Award No. CMS-0201015), Temple University, and Lafayette College.

1124

Potrebbero piacerti anche