Tests of Ground-Penetrating Radar and Induced Polarization for Mapping Fluvial Mine Tailings on the Floor of the Couer d'Alene River, Idaho

by: David L. Campbell, Jeffrey C. Wynn, Stephen E. Box, Arthur A. Bookstrom, and Robert J. Horton
Mineral Resources Program, U.S. Geological Survey

The following is an abbreviated version of a paper presented at the SAGEEP symposium, Reno, Nevada, on 25 March 1997

|| Abstract || Historical background || Field experiments || Data examples || Conclusions || Acknowledgements || References ||


Figure 1, map of Idaho showing location of Couer d'Alene in northwest part of the state.
Figure 1. Location map showing Couer d'Alene in the northwest part of Idaho.

In order to investigate sequences of toxic mine tailings that have settled in the bed of the Coeur d'Alene River, Idaho, (see figure 1) we improvised ways to make geophysical measurements on the river floor. To make ground penetrating radar (GPR) profiles, we mounted borehole antennas on a skid that was towed along the river bottom. To make induced polarization (IP) profiles, we devised a bottom streamer from a garden hose, lead strips, PVC standoffs, and insulated wire. Each approach worked and provided uniquely different information about the buried toxic sediments. GPR showed shallow stratigraphy, but did not directly detect the presence of contaminating metals. IP showed a zone of high chargeability that is probably due to pockets of relatively higher metal content. Neither method was able to define the base of the fluvial tailings section, at least in part because the IP streamer was deliberately designed to sample only the top three meters of sediments to maximize horizontal resolution.

Return to top of page.

Historical background

Beginning in 1884, lead-zinc-silver was produced from the historic Coeur d'Alene mining district in Idaho (Hobbes and Fryklund, 1968). Tailings from the old mining operations have often flushed down the Coeur d'Alene River, especially during times of heavy winter floods, depositing in the river channel, flood plain, and even as far as Lake Coeur d'Alene 50 miles (80 kilometers) away. These fluvial deposits are reworked by each big flood, and some contain enough heavy metals, including lead and cadmium in particular, to be dangerous or even fatal to fish and grazing herbivores that ingest the sediment. Box and others (1994) discussed some of the resulting environmental problems, as well as prospects for their remediation. Drilling and assaying are reliable for investigating the fluvial tailings from spot to spot, but are slow, difficult, probably not representative, and expensive. A fast, cheap method was needed to map their location, thickness, and compositions. This paper reports on tests we made to do that using ground penetrating radar (GPR) and a modified form of induced polarization (IP).

Return to top of page.

Field experiments

Our field tests were conducted in September 1996 at two sites on the Coeur d'Alene River above Lake Coeur d'Alene. We report here one of those sites, the Killarney profile about 1.5 miles (2.5 kilometers) due east of Killarney Lake.

Radar Gear:

Our GPR equipment (figure 2) consisted of a GSSI SIR-10A+ unit and several GSSI antennas, including an 80 MHZ surface antenna and a pair of 200 MHZ borehole antennas.

Figure 2. Photo of Dave Campbell holding the underwater towed antenna. Steve Box, USGS, is the boat operator.

The surface-towed 80 MHZ antenna clearly saw gross sedimentary structures along the entire line, and it also showed water depth (which we already knew from a fish-finder profile) and diffraction V's from (metallic?) junk in the river bed. To get more structural details in our GPR profiles, we made a skid to tow along the river floor, in which we mounted our borehole GPR antennas.

Return to top of page.

IP Gear:

Our IP equipment consisted of a Zonge AMG-3 generator, a GGT-3 transmitter, and a GDP-16 receiver (figure 3). Building from a garden hose (to ensure fixed distances between dipole elements) we constructed a streamer that would rest on the river bottom. For electrodes we used thin lead strips wrapped around the hose and held in place with cable ties. The streamer was designed for use in a dipole-dipole mode with N = 1, 2, and 3. The electrodes were spaced 3 meters apart. Potential wires were held 40 centimeters from the current wires by PVC standoffs spaced every meter along the streamer. This separation was necessary in order to cut down on capacitive coupling or cross-talk between transmitter and receiver circuits. We would start each profile by laying out the streamer in the water next to the opposite shore. After making a measurement, we would pull the array in a distance equal to one electrode spacing, repeat the process, and continue. Measurements were made in the time-domain mode, with more detailed Spectral IP measurements taken when we encountered pockets of high-chargeability sediments.

Figure 3, induced polarization  gear laid out at the river's edge.
Figure 3. Photo of Jeff Wynn standing with the IP equipment as it was deployed on the river bank at the Killarney site.

Return to top of page.

Figure 4, the modified induced polarization array, A=3 meters, N = 1, 2, 3 deployed in the river.
Figure 4. Photo shows the modified IP array deployed in the river.

Data examples

Figure 5, the underwater induced polarization profile at the Killarney site.
Figure 5 shows IP chargeabilities taken using the river-bottom streamer, averaged using the method of Eaton and Nabighian (1996). These averages roughly apply to Edwards (1977) depths below the river floor of 1.25 meters (for dipole-dipole pair N = 1), 2.1 meters (for N = 2) and 2.9 meters (N = 3). Resistivities over the entire line varied slowly from 12 ohm-meters on the eastern shore, probably caused by higher clay content there, to 45 ohm-meters on the west side of the river.

Return to top of page.

Over much of the Killarney IP profile the measured chargeabilities are low and flat, with a general tendency for chargeability to increase slightly with depth. Note that the fixed geometry of our streamer (A=3 meters, N=1,2,3) probably did not let us detect features below about 3 meters, so that we could not resolve pre-mining sediments beneath the river. The interval between about X=-40 meters and X=-20 meters has much higher chargeabilities, with the highest values at shallow depth, interpreted to be a shallow pocket of high-metal-content tailings in this place. Compare this with the radar profil:

Figure 6, ground-penetrating radar profile at the Killarney site.
Figure 6 shows GPR data processed and displayed as a gray-scale image draped at river bottom depth. The image was cut off below about 3.8 meters because of increasing noise. The system of wave trains slanting steeply inward from the east (especially) and west ends of the image represent reflections from the water surface, not sedimentary structures there.

The interval from X =-40 meters to X =-20 meters shows GPR waveforms that are jumbled; this coincides with unusually high IP chargeabilities (See the earlier IP profile Figure). Waveforms in this interval almost certainly reflect real stratigraphic structures in the river bed; they contain particular arrivals that are identifiable from scan to scan as they move up and down in time. Such arrivals very likely represent reflections from interfaces between particular depositional units (sand and silt, for example), interfaces that are continuous but not flat. We believe this zone resulted from the river scouring its channel down into a bar of the fine-to-very-fine sand and then redepositing a medium sand in it during the great 1974 flood. We think the east-dipping GPR reflector might represent the edge of the 1974 channel scour, and the jumbled GPR arrivals between about X=-32 meters and X=-20 meters could reflect blocks of pre-1974 units that slumped towards the channel as the river cut into the bar. This slumping re-exposed more deeply buried units of fluvial mine tailings with high metal values.

Return to top of page.


The IP work clearly detected a high-chargeability zone which we think reflects shallow deposits of fluvial mine tailings with high metal concentrations. We interpret these shallow deposits to have resulted from reworking of river sediments during a flood. GPR is insensitive to high metal values, but it does seem to map some shallow stratigraphy, probably reflecting textural differences between depositional units. The GPR stratigraphy showed jumbled reflectors in the high-chargeability zone that may represent slumping.

GPR failed to detect the base of the fluvial mine tailings under the Killarney profile, however. This failure probably stems from the fact that there may not have been any contrast for GPR to see. Laboratory measurements showed no significant contrast of electrical properties, at least at GPR frequencies, between the deepest fluvial mine tailings and the pre-mining river bed deposits.

Return to top of page.

IP streamers like the one used here should work to find other shallow high-chargeability targets of this type. Such IP surveying (with or without auxiliary GPR) might help direct core sampling work, which must be done in any case to develop a remediation strategy for the Coeur d'Alene River basin. Ideally, however, fluvial mine tailings in all of the river basin need to be investigated and remediated, including units now buried deeper than we could see using our particular IP array size and GPR frequencies. Deeper-seeing IP units (e.g., with larger electrode spacings) could be used to help find deeper-buried high-chargeability units, especially ones that are thick or that have particularly high metal concentrations. Conventional IP data are typically difficult to characterize, however, and more expensive spectral IP would have to be used in selected instances. GPR work does give better depth resolution than IP, but may only be useful in places where there are sufficient electrical-property contrasts between tailings and pre-mining deposits, and where formation resistivities are high enough to allow good GPR depths of investigation.


We thank Don McNair for many practical suggestions in the design and fabrication of the IP streamer. We also thank Don McNair, Samantha Korash, and H. Richard Blank for helping collect the IP and GPR field data.

Return to top of page.


Box, S.E., Bookstrom, A.A., Lindsay, J., and Smith, C., 1994, River dispersal of mine tailings downstream from the Coeur d'Alene mining district and Bunker Hill Superfund site, Idaho -- Part 1. Geologic overview and outlook for remediation: Geol. Soc. America Programs with Abstracts, v. 26, no. 7, p.506.

Eaton, P.A., and Nabighian, M., 1996, Some quasi-quantitative approaches used to interpret IP data for gold exploration (abs.): Expanded Abstracts with Authors' Biographies, Society of Exploration Geophysicists International Exposition and Sixty-sixth Annual Meeting, Novenber 10-15, 1996, Denver CO, p. 607-610.

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

Hobbes, S.W., and Fryklund, V.C., Jr., 1968, The Coeur d'Alene District, Idaho, in Ridge, John D., ed., Ore Deposits of the United States, 1933-1967, the Graton-Sales volume: New York, American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc., v. 2, p. 1417-1435.

Powers, M.H., and Olhoeft, G.R., 1995, GRPMODV2 -- One-dimensional full waveform forward modeling of dispersive ground penetrating radar data, version 2.0: U.S. Geological Survey Open-file Report 95-58, 41p. and floppy disk.

Roth, M.E., 1996, Sample analysis and modeling to determine GPR capacity for mapping fluvial mine tailings in the Coeur d'Alene River channel: U.S. Geological Survey Open-file Report 96-515, 35 p.

Return to top of page.

For more information, contact Dave Campbell; phone (303)236-1380, FAX (303) 236-1425, email davec@musette.cr.usgs.gov