USGS Home
Contact USGS
Search USGS

 Home  Resources  Research  Publications  About  Contact Us  Groundwater Information

Integration of Surface Geophysical Methods for Fracture Detection in Bedrock at Mirror Lake, New Hampshire

By C.J. Powers **, Kamini Singha, and F. Peter Haeni

ABSTRACT

Five surface geophysical methods were used to determine the locations of fracture zones in crystalline bedrock for predicting fluid flow and chemical migration at the U.S. Geological Survey Fractured Rock Research Site at Mirror Lake, Grafton County, New Hampshire. Two methods of direct-current (dc) resistivity (two-dimensional (2D) and crossed square-array profiling), two methods of inductive terrain conductivity, and very-low-frequency electromagnetics (VLF) were used over survey lines extending about 200 meters. The results of the five methods were correlated to locate fracture zones; anomalies that were detected in one or two of the results were eliminated, increasing the confidence in the interpretation of anomalies detected in all of the results.

Two low resistivity anomalies were detected with all the geophysical methods in the southeast part of the study area. Based on the geophysical, outcrop, and photolinear data, the anomalous areas were interpreted as steeply dipping fracture zones approximately 10-meters wide. One interpreted fracture zone strikes approximately north 45 degrees east and the other strikes approximately north 17 degrees east.

Results of dc-resistivity surveys were analyzed to estimate the secondary porosity of the two interpreted fracture zones. Crossed square-array dc-resistivity profiling data indicates the secondary porosity is between 0.65 to 0.75 percent, whereas the 2D dc-resistivity profiling results indicate the secondary porosity to be 1.6 to 1.9 percent. Estimates from the 2D dc-resistivity profiling could indicate the effects of alteration and/or iron precipitate observed in outcrops near the survey lines.

INTRODUCTION

The location and orientation of fracture zones is important for modeling fluid flow and contaminant transport in fractured rocks. Surface geophysical methods are a rapid, inexpensive addition to drilling for determining the locations and orientation of fractured zones in bedrock. Surface geophysics can be used in conjunction with geologic, hydrologic, and borehole-geophysical investigations to optimize well siting (Jansen and Jurcek, 1997), or as a stand-alone method of fracture detection (Lewis and Haeni, 1987; Lieblich and others, 1991; Haeni and others, 1993).

Five different surface geophysical methods were used during the summer of 1997 at Mirror Lake, Grafton County, New Hampshire to determine the location of saturated fracture zones in crystalline bedrock. 2D dc-resistivity profiling, crossed square-array dc-resistivity profiling, two inductive terrain conductivity methods (EM-34¹ and Slingram) and very-low-frequency electromagnetics (VLF) were used to detect low resistivity zones that indicate the location of fracture zones. By comparing the results from all of the methods, anomalies induced by noise were identified and eliminated, increasing the confidence with which fracture zones were interpreted from the remaining low resistivity anomalies. The study site was selected to profile across several photo-lineaments identified by Clark and others (1999).

SITE CHARACTERIZATION

The USGS Fractured Rock Field Research Site is located at Mirror Lake, in the Hubbard Brook Valley, Grafton County, New Hampshire. The data presented in this paper were collected approximately 1 kilometer (km) west of Mirror Lake (fig. 1). At this location, results of 2D dc-resistivity surveys indicate that a layer of glacially deposited sediment 3- to 6- meters (m) thick covers the bedrock. Overlying the till is a layer of organic material, approximately 0.2-m thick, consisting mainly of organic-rich soil and leaf matter. Small cavities in the organic layer were observed in several locations within the study site.

Figure 1. Site map of the study site, Mirror Lake, Grafton County, New Hampshire.

Geophysical data were collected along two parallel lines oriented north 43 degrees west, about 200-m long and 20-m apart. The average slope of the land surface along the lines dips 8 degrees to the southeast. The survey lines were located and oriented to intersect the maximum possible number of geologic features and lineaments observed in air-photos. Lineaments observed in low altitude aerial photos served to focus the initial study site selection, but due to the lineament map scale, the intersection of the lineaments with the survey lines could not be accurately determined. The first 140 m of Line 1 were 10 to 40 m from the northeast of the edge of the Hubbard Brook gorge. Preferential erosion, rockslides, and weathered zones along the walls and edge of the gorge likely mark the location of several large fracture zones.

PRINCIPLES OF SURFACE GEOPHYSICAL TECHNIQUES

Dc resistivity

Dc resistivity methods measure the electrical resistivity distribution of the subsurface using current transmitted into the ground from dc- or low-frequency sources, by two electrodes (C1 and C2), and measuring the potential difference between a second pair of electrodes (P1 and P2) (fig. 2). The apparent resistivity of the subsurface can be calculated by applying a geometric correction (K) to Ohm's law (R = DV/I, where R is the resistance, DV is the measured potential difference, and I is the injected current), based on the specific electrode spacing and geometry. These geometrically corrected measurements are defined as apparent resistivities rather than true resistivities because a resistively homogeneous subsurface is assumed. Measured resistivity values are controlled by material resistivity, and the presence, quality, and quantity of ground water (Haeni and others, 1993). The resistivity of a fracture zone is controlled by the secondary porosity, and the presence of altered secondary minerals and/or precipitate. The maximum penetration depth is directly proportional to the electrode spacing and inversely proportional to the subsurface conductivity (Edwards, 1977).

Figure 2. Schematic of Schlumberger, dipole-dipole, and square-array dc-resistivity electrode configuration.

Two dc-resistivity methods were used at Mirror Lake: 2D dc-resistivity profiling and crossed square-array dc-resistivity profiling. 2D dc-resistivity profiling is conducted by making many measurements at different locations along the profile and at different offsets. The 2D dc-resistivity profiling data are inverted to create a tomogram-like model of resistivity along a section of the subsurface that can be used to detect and define individual fracture zones. The crossed square-array profiling method measures changes in apparent resistivity with measurement direction along a profile.

The 2D dc-resistivity profiling system manufactured by Advanced Geosciences, Inc. consists of a linear array of 28 electrodes that are connected to a measurement control unit by a specially designed cable. The control unit uses an automated data-collection program to control the location of current and potential electrodes. Two types of arrays were used for profiling: a dipole-dipole array and a Schlumberger array (fig. 2). The dipole-dipole array has better horizontal resolution, but poorer depth of penetration compared to the Schlumberger array (Loke, 1997). By using an iterative smoothness-constrained least-squares inversion method (deGroot-Hedlin and Constable, 1990; Sasaki, 1992), apparent resistivity data collected by the 2D dc-resistivity system are inverted to create a model of subsurface resistivity that approximates the true subsurface resistivity distribution (Loke, 1997). Linear zones of low resistivity that are continuous with depth are interpreted as fracture zones.

The crossed square-array profiling method uses two squares of electrodes of equal side length "a" rotated by 45 degrees around a center point, defined as the measurement location (fig. 2). Apparent resistivity is measured along the lengths of the arrays (the r_a and r_b measurements) and also across the diagonals (r_g) for each square. A series of crossed squares are collected along the length of a line as shown in figure 3. In a layered medium:

The crossed square-array data provide information about the resistivity as a function of direction (every 45 degrees) at each station along the profile. From this data, the mean apparent resistivity, and the magnitude of apparent anisotropy in the resistivity are determined for each crossed square. The magnitude of the anisotropy can be used to calculate a secondary porosity in saturated bedrock. A decreased mean resistivity and an increased secondary porosity can indicate the location of a fracture zone.

Figure 3. Crossed square-array dc-resistivity profiling.

Inductive terrain conductivity

Inductive terrain conductivity is an electromagnetic method that measures subsurface electrical conductivity. Two inductive terrain-conductivity instruments were employed: the EM-34 from Geonics and the Slingram from ABEM Instrument AB. Terrain-conductivity instruments consist of a transmitting coil, a receiving coil, a control unit for each, and intercoil cables. The coils are held coplanar at a constant offset, and data are collected at discrete intervals along a survey line.

For this study, the EM-34 was used in the vertical dipole configuration, where the coils are held horizontally. The vertical dipole configuration of the equipment is most sensitive to material at a depth of 0.45 times the coil spacing, and can penetrate depths about 1.5 times the transmitter-receiver spacing. This configuration of the EM-34 is fairly insensitive to near-surface material (McNeill, 1980a).

The Slingram is used in a similar manner, with the transmitter and receiver coils kept at a constant separation, and held parallel to the ground. The depth of penetration of the Slingram is about 0.75 times the coil spacing (Borje Niva, Swedish Geological Company, written commun.,1997).

A qualitative method was used to interpret the inductive terrain-conductivity data. Fracture zones were identified by comparing measured data to instrument response calculated for models of conductors with similar shapes. The vertical dipole response over a thin, vertical conductor is shown in figure 4. Vertical dike-like conductive bodies, such as water-bearing fractures, produce negative anomalies in the conductivity response curve; the true conductivity is not measured (McNeill, 1980b). Based on the modeling results, anomalies were selected that had negative or low measured conductivity values over several data points.

Figure 4. Vertical-dipole response of an inductive terrain-conductivity instrument to a thin vertical conductor.

Very-low-frequency electromagnetics (VLF)

The VLF method is a passive electromagnetic (EM) method that utilizes powerful military transmitters operating between 15-30 kilohertz (kHz) as the primary EM wave source. VLF methods can be used to determine the locations of saturated, sub-vertical conductive zones in which the primary EM wave induces current flow. The field radiated from a VLF transmitter over a uniform or horizontally layered earth consists of a vertical electric field component and a horizontal magnetic field component, each perpendicular to the direction of propagation (McNeill, 1990). Because the source of the electromagnetic field is usually greater than 50 miles away, the long wavelength EM wave approximates a plane wave. In this study, the VLF was used in two modes: tilt angle and resistivity.

Although the primary magnetic field is oriented horizontally and perpendicularly to its source, induced current flowing in fracture zones produces a secondary magnetic field that is out-of-phase with the primary magnetic field and is oriented in any direction (McNeill and Labson, 1990). The vector sum of the two fields traces out an ellipse over time, the tilt of which is measured in the VLF tilt angle mode (fig. 5). The tilt angle is approximately equal to the in-phase part of the vertical component of the magnetic field ellipse. (Iris Instruments, 1993).

Figure 5. Tilt angle and ellipticity of a VLF field.

For tilt angle measurements, magnetic field coupling with the fracture zone is important. Therefore, the VLF-transmitter should be located along the strike of the target. The depth of investigation is dependent on the frequency used and the resistivity of the host medium. At the study site, the local resistivity minimum is 400 ohm-m (based on the 2D dc-resistivity data), therefore, the VLF depth of investigation is more than 65 m (Iris Instruments, 1993).

Measuring resistivity with the VLF requires measuring the electric field with two electrodes connected to the control unit by a wire. In the resistivity-mode, the VLF is similar to the high frequency magnetotelluric method (Kaikkonen and Sharma, 1997) and measures apparent resistivity and phase angle (time-delay) between the electric and magnetic fields. For resistivity measurements, the electrical field coupling with the fracture zone is important, therefore, the transmitter should be located in a direction perpendicular to the target (Iris Instruments, 1993).

A qualitative method was used to interpret the VLF tilt-angle data to identify fracture zones by comparing measured data to instrument response calculated for models of conductors with similar shapes.

A conductive fracture zone produces a tilt-angle anomaly that looks like a distorted sinusoid (fig. 6). The top of the fracture zone is centered at the inflection point of the anomaly. In the VLF resistivity data, low resistivity zones were interpreted as possible fracture zones.

Figure 6. VLF response to a thin vertical conductor.

DATA ACQUISITION AND PROCESSING

2D dc-resistivity profiling

Two 275-m 2D dc-resistivity profiles were collected. The 2D dc-resistivity profiles extend beyond the ends of each survey line to provide 200 m of complete data coverage because of a limited imaging depth for about 35 m on each end of the profile. For each line, a dipole-dipole array and a Schlumberger array were used. The apparent resistivity data were inverted to create a model of the resistivity of the subsurface using Res2dinv. Res2dinv uses an iterative smoothness-constrained least-squares method (deGroot-Hedlin and Constable, 1990; Sasaki, 1992).

To test interpretation, resistivity models were created based on the inversion results. The resistivity models were used to generate synthetic apparent resistivity data. The synthetic apparent resistivity data were inverted using Res2dinv and the resulting inversions were compared with the original inverted resistivity section. The resistivity models were adjusted and simplified to qualitatively match the field-data inversions. Generating resistivity models helped constrain interpretation of the field-data inversions to identify locations and orientations of anomalies.

The 2D dc-resistivity field-data inversions, resistivity models and synthetic-data inversions are shown in figure 7. Bedrock is calculated to have an average resistivity of 3,000 and 4,000 ohm-m along Line 1 and Line 2, respectively. Within the bedrock there are several linear features of low resistivity (400 ohm-m) that are interpreted as fracture zones. Line 1 has a blocky anomaly, dipping to the southeast from 40 to 30 m, a vertical linear anomaly from 40 to 55 m, a dipping linear anomaly from 60 to 100 m and another dipping anomaly from 140 to 190 m. Line 2 has vertical linear anomalies from 0 to 30 m and 55 to 70 m, a sub-horizontal linear anomaly from 80 to 135 m and a sub-vertical linear anomaly from 155 to 175 m.

Figure 7. Results of 2D dc-resistivity data and modeling from Mirror Lake, New Hampshire

(Click here for a larger version of this figure)

Crossed square-array dc-resistivity profiling

A dc-resistivity system manufactured by ABEM Instruments AB was used to collect crossed-square data at 10-m intervals along Line 1, with the first square centered on 5 m. A 10-m a-spacing and a current of 5 mA were used. Data were collected four times and averaged at each station.

The mean apparent resistivity for each crossed square was calculated to determine an azimuthally independent value for each square (Habberjam, 1972):

The secondary porosity was calculated following Habberjam (1975) from apparent anisotropy in the resistivity, which is a measure of the magnitude of resistivity variation as a function of direction. The calculation of secondary porosity assumes that apparent anisotropy is due to fractures and does not account for anisotropy due to schistosity or heterogeneity.

The crossed square-array dc-resistivity profiling data are shown in figure 8. A decreased mean resistivity, and an increased secondary porosity are interpreted as fracture zones. Areas of decreased mean apparent resistivity are located at 5, between 45 and 55 m, and possibly at 155 m. Areas of increased anisotropy and, therefore, increased calculated secondary porosity are located at 5, 45, and 175 m.

Figure 8. Results of square-array profiling at the Mirror Lake, New Hampshire.

Inductive terrain conductivity

The EM-34 was used in the vertical dipole configuration, with a 20-m coil separation. Data were collected every 2.5 m. The 20-m EM-34 data is shown in figure 9. To minimize the effects of random noise, an eight-point moving average function was applied to the data. With a 2.5-m station increment and 20-m intercoil spacing, an anomaly due to a dipping conductor should extend at least 8 data points. Anomalies appear at 5, 55, 135, and 170 m on Line 1 and approximately 15, 60, and possibly at 155 m on Line 2.

Figure 9. Results from EM-34 data collection at Mirror Lake, New Hampshire.

The Slingram data were collected every 2.5 m, using a 40-m coil separation. The Slingram data are shown in figure 10. Slingram data are recorded as a percent of the primary magnetic field intensity at the receiver coil, which is directly proportional to the apparent conductivity. Because of a higher signal-to-noise ratio, only a 2 point moving-average function was applied to the data. Anomalies were detected at 15, 55, and, possibly, at 145 m on Line 1. On Line 2, anomalies were detected at 10, 60, and, possibly, at 160 m.

Figure 10. Results from Slingram data collection at Mirror Lake, New Hampshire.

Very-low-frequency electromagnetics (VLF)

A VLF transmitting station in Cutler, Maine (24 kHz) was the only VLF station with a strong enough signal for VLF data acquisition. At the Hubbard Brook Site, the direction to the Cutler, Maine transmitter approximately east 10 degrees north. The direction to the transmitter intersects the survey line at an angle of less than 60 degrees to the strike of the suspected fracture zone; suitable for the resistivity survey, but less suitable for the tilt angle survey. However, precise tilt-angle measurements were obtained using the Cutler, Maine station.

Tilt angle and resistivity data were collected along the survey lines every 2.5 m. For the resistivity survey, the electrodes were separated by 10 m. A minimum of three measurements were collected and averaged at each station to reduce the effects of random noise. The VLF data is shown in figure 11. A two point moving average function was applied to the tilt angle data. Resistivity data collected with the VLF were first converted to conductivity values in order to more clearly show the anomalies. A four point moving-average function was applied to the resistivity data. Tilt angle anomalies were detected at 50, and, possibly, at 20 m on Line 1 and approximately 60 m, and, possibly, at 20 m on Line 2. VLF resistivity anomalies were detected at 80 m, and, possibly, at 135 m on Line 1 and 15, 60, and 100 m on Line 2. An area of higher conductivity is measured from 140 to 190 m on Line 2.

Figure 11. Results of VLF data from Mirror Lake, New Hampshire.

INTEGRATED RESULTS

Because of the robust modeling program and the high quality data, the integrated results are based primarily on the 2D dc-resistivity inversions supported by the results of the other geophysical methods. The integrated results are summarized in figure 12. Two anomalous areas were detected by all the geophysical methods. These areas, labeled Fracture Zone 1, and Fracture Zone 2, indicate the presence of fracture zones from 0 to 20 m, and 45 to 55 m on Line 1, and from 0 to 25 m, and 55 to 65 m on Line 2. Several other anomalies interpreted with one or two methods in other locations along the survey lines were eliminated because no clear correlation between methods or from one line to the other was found.

Figure 12. Interpreted cross-section of Lines 1 and 2, Mirror Lake, New Hampshire.

(Click here for a larger version of this figure)

Fracture Zone 1, located from 0 to 20 m on Line 1 and from 0 to 25 m on Line 2, was detected by all geophysical methods (fig. 12). The strike of this zone across Line 1 and Line 2, is approximately north 45 degrees east. 2D dc-resistivity modeling suggests the fracture zone is irregular, indicating that more than one fracture zone may be producing the anomaly. The blockiness of Fracture Zone 1 on Line 1 may actually be two fracture zones, one dipping 21 degrees to the southwest and the other sub-vertical. On Line 2, the anomaly that characterizes Fracture Zone 1 is more vertical, suggesting that the sub-vertical fracture zone extends between the two lines, while the shallowly dipping feature does not.

Fracture Zone 2 is located from 45 to 55 m on Line 1 and from 55 to 65 m on Line 2. Fracture Zone 2 is the most prominent anomaly in most of the geophysical surveys and was detected by all geophysical methods. The strike of this zone across Line 1 and Line 2 is approximately north 17 degrees east. Results of 2D dc-resistivity modeling indicate that the anomaly is sub-vertical. Observations on the face of the Hubbard Brook Gorge show a sub-vertical fracture zone 5 to 6 m wide, striking north 34 degrees east (+/- 10 degrees). The zone consists of heavily fractured, incompetent schist thinly coated with iron precipitate. The fracture zone projects between 45 to 50 m on Line 1 and 46 to 56 m on Line 2, consistent with the location of Fracture Zone 2. A lineament observed in low altitude aerial photos also projects through this approximate area.

A group of anomalies, interpreted as a probable fracture zone, is located on Line 1 from 135 to 170 m and on Line 2 from 160 to 170 m. Within this group, all the geophysical methods detected an anomaly on Line 1 and most methods detect an anomaly on Line 2, but the interpreted location on Line 1 is inconsistent (the anomalies were detected from 135 to 170 m), and all of the anomalies are low magnitude. Factors that may cause the anomalies to be poorly formed include multiple interfering anomalies, a greater overburden thickness, or a decreased fracture thickness. Projection of a possible fracture zone observed in the Hubbard Brook gorge intersect Line 1 at 145 to 150 m, and Line 2 at 150 to 160 m, which is consistent with the anomaly grouping. A second lineament observed in low altitude aerial photos also projects through the northwestern part of the survey lines and may correlate with this zone.

Modeling of the 2D dc-resistivity data indicate that the resistivity of the fracture zones is approximately 400 ohm-m, whereas the resistivity of the surrounding rock is approximately 3,000 to 4,000 ohm-m. Assuming pore fluid resistivity is 3.3 ohm-m, a secondary porosity of both fracture zones of 1.6 to 1.9 percent was calculated using values based on 2D dc-resistivity and square array dc-resistivity data and Archie's Law (Archie, 1942):

The value of m was calculated using Archie's law, by setting r_r to 3,000 to 4,000 ohm-m (average bedrock based on 2D resistivity modeling) and r_f to 3.3 ohm-m. A secondary porosity of 0.29 percent was assumed based on the average value of the secondary porosity calculated from the square-array dc-resistivity data collected over non-anomalous areas. The coefficient of saturation was assumed to be equal to 1.

Using the calculated cementation factor of 1.17 to 1.22, the secondary porosity of 1.6 percent to 1.9 percent for the fracture zone was calculated from the 2D dc-resistivity modeling. Crossed square-array dc-resistivity profiling data indicates the secondary porosity is between 0.65 to 0.75 percent in the same location.

The secondary porosity calculated from the 2D dc-resistivity data is higher than that calculated from the square-array dc-resistivity data in the same area. One reason is that the value of the secondary porosity determined from the square-array dc-resistivity is an average of the entire volume under the array, whereas the secondary porosity value determined from the 2D dc-resistivity inversion only includes the volume from the fracture zone.

The high secondary porosity calculated from the 2D dc-resistivity modeling may indicate that the fracture zones contains rubble, or iron precipitate. On the face of the gorge, both rubble and iron precipitate were observed in Fracture Zone 2.

CONCLUSIONS

Five surface geophysical methods were used to determine the locations of fracture zones in crystalline bedrock for predicting fluid flow and chemical migration at the U.S. Geological Survey Fractured Rock Research Site at Mirror Lake, Grafton County, New Hampshire. Two methods of direct-current (dc) resistivity (two-dimensional (2D) and crossed square-array profiling), two methods of inductive terrain conductivity (EM-34 and Slingram), and very-low-frequency electromagnetics (VLF) were used over survey lines extending about 200 meters. The results of the five methods were correlated to locate fracture zones; anomalies that were detected in one or two of the results were eliminated, increasing the confidence in the interpretation of anomalies detected in all of the results.

Two low resistivity anomalies were detected with all the geophysical methods in the southeast part of the study area. Based on the geophysical, outcrop, and photolinear data, the anomalous areas were interpreted as steeply dipping fracture zones approximately 10-meters wide. One interpreted fracture zone strikes approximately north 45 degrees east and the other strikes approximately north 17 degrees east.

Results of dc-resistivity surveys were analyzed to estimate the secondary porosity of the two interpreted fracture zones. Crossed square-array dc-resistivity profiling data indicates the secondary porosity is between 0.65 to 0.75 percent, whereas the 2D dc-resistivity profiling results indicate the secondary porosity to be 1.6 to 1.9 percent. Estimates from the 2D dc-resistivity profiling could indicate the effects of alteration and/or iron precipitate observed in outcrops near the survey lines.

ACKNOWLEDGEMENTS

The U.S. Geological Survey Fractured Rock Research Site is located within the Hubbard Brook Experimental Forest. The Hubbard Brook Experimental Forest is operated and maintained by the Northeastern Forest Experiment Station, USDA Forest Service, Radnor, Pennsylvania.

REFERENCES

Archie, G.E., 1942, The electrical resistivity log as an aid in determining some reservoir characteristics: Trans. Am. Inst. Min., Metal. and Petr. Eng., v. 146, p. 54-62.

Clark, S.F., Jr., Ferguson, E.W., Short, H.A., Marcoux, G.J., and Moore, R.B., 1999, Lineament map area 9 of New Hampshire bedrock aquifer assessment, north-central New Hampshire: U.S. Geological Survey Open-File Report 99-63, scale: 1:48000.

deGroot-Hedlin, C. and Constable, S., 1990, Occam's inversion to generate smooth, two-dimensional models from magnetotelluric data: Geophysics, v. 55, p.1613-1624.

Edwards, L.S., 1977, A modified pseudosection for resistivity and IP: Geophysics, v. 42, no. 5, p. 1020-1036.

Habberjam, G.M., 1972, The effects of anisotropy on square array resistivity measurements: Geophysical Prospecting, v. 20, p. 249-266.

Habberjam, G.M., 1975, Apparent resistivity, anisotropy, and strike measurements, Geophysical Prospecting, v. 23, p. 211-247.

Haeni, F.P., Lane, J.W., Jr., and Lieblich, D.A., 1993, Use of surface-geophysical and borehole-radar methods to detect fractures in crystalline rocks, Mirror Lake area, Grafton County, New Hampshire in Banks, D. and Banks, S., eds., Hydrogeology of Hard Rocks, Memoires of the XXIVth Congress, Oslo, Norway: International Association of Hydrologists, p. 577-587.

Iris Instruments, 1993, T-VLF Operating manual (Release 1.0), Orleans, France.

Jansen, J., and Jurcek, P., 1997, The application of surficial geophysics to well site exploration and wellhead protection in fracture controlled aquifers in The Symposium on the Application of Geophysics to Environmental and Engineering Problems '97 Volume II: Wheat Ridge, Colorado, Environmental and Engineering Geophysical Society, p. 635-644.

Kaikkonen, P., and Sharma, S.P., 1997, Delineation of near-surface structures using VLF and VLF-R data-an insight from the joint inversion results: The Leading Edge, v. 16, no. 11, p. 1683-1686.

Lewis, M.R., and Haeni, F.P., 1987, The use of surface geophysical techniques to detect fractures in bedrock-an annotated bibliography: U.S. Geological Survey Circular 987, 14p.

Lieblich, D.A., Lane, J.W., Jr., and Haeni, F.P., 1991, Results of integrated surface-geophysical studies for shallow subsurface fracture detection at three New Hampshire Sites in Expanded Abstracts with Biographies, SEG 61st Annual International Meeting Volume I: Society of Exploration Geophysicists, p. 553-556.

Loke, M.H., 1997, Electrical imaging surveys for environmental and engineering studies: a practical guide to 2D and 3D surveys: Unpublished short training course lecture notes, Universiti Sains Malaysia, Penang, Malaysia.

McNeill, J.D., 1980a, Electromagnetic terrain conductivity measurements at low induction numbers: Geonics Ltd. Technical Note TN-6, Mississauga, Ontario, Canada.

McNeill, J.D., 1980b, EM 34-3 survey interpretation techniques: Geonics Ltd. Technical Note TN-8, Mississauga, Ontario, Canada.

McNeill, J.D., 1990, Use of electromagnetic methods for groundwater studies, in Ward, Stanley H. ed., Geotechnical and Environmental Geophysics, Volume 1: Tulsa, Oklahoma, Society of Exploration Geophysicists, p. 191-218.

McNeill, J.D., and Labson, V.F., 1990, Geological mapping using VLF radio fields, in Nabighian, M. ed., Electromagnetic Methods in Applied Geophysics, Volume 2, Part B: Tulsa, Oklahoma, Society of Exploration Geophysicists, p. 521-640.

Sasaki, Y., 1992, Resolution of resistivity tomography inferred from numerical simulation: Geophysical Prospecting, v. 40, p. 453-464.

AUTHOR INFORMATION

C.J. Powers, U.S. Geological Survey, Storrs, CT

Kamini Singha, U.S. Geological Survey, Storrs, CT

F. Peter Haeni, U.S. Geological Survey, Storrs, CT

[1] The use of trade names is for descriptive purposes only and does not constitute endorsement by the U.S. Geological Survey.

Final copy as submitted t for publication as: Powers, C.J., Singha, Kamini and Haeni F. Peter, 1999, Integration of Surface Geophysical Methods for Fracture Detection in Bedrock at Mirror Lake, New Hampshire, in Morganwalp, D.W. and Buxton, H.T., eds., U.S. Geological Toxic Substances Hydrology Program -- Proceedings of the Technical Meeting, Charleston, South Carolina, March 8-12, 1999: USGS Water-Resources Investigations Report 99-4018C, v. 3, p. 757-768.