Airborne electromagnetics (EM) as a three-dimensional aquifer-mapping tool

Jeff Wynn, USGS, 954 National Center, Reston, VA 20192
Don Pool, Mark Bultman, and Mark Gettings, USGS, Tucson, Arizona,
and Jean Lemieux, Fugro Airborne Geophysics, Ottawa, Canada

[Proceedings Volume, SAGEEP-2000 Conference, 20-24 Feb 2000, p. 93-100]

Contents: Abstract || Introduction and background || Mapping the basement from magnetic data || Mapping the aquifer using airborne EM || Conclusions || References


The San Pedro River in southeastern Arizona hosts a major migratory bird flyway, and was declared a Riparian Conservation Area by Congress in 1988. Recharge of the adjacent Upper San Pedro Valley aquifer was thought to come primarily from the Huachuca Mountains, but the U. S. Army Garrison of Fort Huachuca and neighboring city of Sierra Vista have been tapping this aquifer for many decades, giving rise to claims that they jointly threatened the integrity of the Riparian Conservation Area. For this reason, the U. S. Army funded two airborne geophysical surveys over the Upper San Pedro Valley (see figure 1), and these have provided us valuable information on the aquifer and the complex basement structure underlying the modern San Pedro Valley. Euler deconvolution performed on the airborne magnetic data has provided a depth-to-basement map that is substantially more complex than a map obtained earlier from gravity data, as would be expected from the higher-resolution magnetic data. However, we found the output of the Euler deconvolution to have "geologic noise" in certain areas, interpreted to be post-Basin-and-Range Tertiary volcanic flows in the sedimentary column above the basement but below the ground surface.

Figure 1, map of Arizona, showing location of study area in southeastern Arizona.
Figure 1. Map showing locations of 1997 and 1999 surveys in southeastern Arizona.

Return to top of page.

The EM component of the airborne geophysical survey used the Geoterrex-Dighem 20-channel multicoil GEOTEM® system1 configured in a CASA aircraft (figure 2). We inverted these airborne EM data using the proprietary Geoterrex-Dighem Conductivity-Depth-Transform (CDT) algorithm and Encon's EM-Flow1 inversion software, and compared the resulting resistivity-vs-depth sections to the limited E-Logs available in the survey area. We find that the CDTs map the conductive aquifers to deeper depths, but the EM-Flow Conductivity-Depth-Inversions (CDIs) provide somewhat better vertical resolution at shallower depths. The CDTs generally map the water table down to about 150 meters depth except in areas with high resistivity, low-porosity overlying rocks (a confined aquifer situation). Here the CDTs apparently map where water exists below a silt-clay, mid-basin unit that because of lower porosity keeps the bulk of the water below the equipotential line. The CDTs show an increasingly higher resistivity above the E-Log values for depths below 150 meters; we believe, however, that we still have qualitative information down to about 400 meters depth in areas where there is no cultural interference such as power lines and pipelines. Experiments suggest that it may be possible to "tune" the inversion process to provide more accurate resistivity data at these greater depths

Figure 2, photograph of the CASA aircraft U. S.ed in the survey.
Figure 2. Photo of the CASA aircraft in flight over Sierra Vista

Return to top of page.

Introduction and Background

Airborne geophysical surveys, including both magnetic and GEOTEM1 (time-domain electromagnetic) components, were flown in 1997 and 1999 over the Upper San Pedro River drainage in southeastern Arizona, with the objective of mapping a poorly-understood but ecologically sensitive aquifer (see figure 1). The Upper San Pedro aquifer (Pool and Coes, 1999) is bracketed on one side by crystalline and sedimentary rocks of the Huachuca Mountains on the west and the Mule Mountains and Tombstone Hills on the east. The U. S. Army Fort Huachuca Garrison and neighboring city of Sierra Vista lie east of the Huachuca Mountain front, but west of the San Pedro River Riparian Conservation Area (see figures 3, 4, and 6; the heavy aqua band in figure 6 marks the center of the narrow Riparian Conservation Area). The San Pedro River is an important migratory bird flyway, and surface water is essential for its continued integrity. Despite conservation measures, groundwater discharge probably still exceeds the estimated recharge, which is mainly from rainfall and snow on the Huachuca Mountains to the west. The historical flow has apparently been little affected by water withdrawal by large copper mines in neighboring Mexico to the south beginning in the late 1940's (Brown and others, 1966). Gravity studies (Gettings and Gettings, 1996) have provided a general map showing depths to crystalline basement beneath the basin sediments. The processed depths-to-magnetic-source from the recently-acquired magnetic data provide much higher resolution to the structure of the crystalline basement, but also show geologic "noise" in the form of volcanic flows within the sedimentary column above the basement. One feature that has significantly changed the original basement morphology is the Cretaceous. age Tombstone Caldera (Moore, 1993) in the northeastern part of the study area (see figure 4).

Under these circumstances, the U. S. Army, a major user of the aquifer, contracted with the U. S. Geological Survey to conduct a detailed study of the Upper San Pedro basin to better understand the structure and stratigraphy of the basin, and where the water might be. This was done using airborne magnetic instrumentation to map the crystalline basement underlying the basin sediments, and airborne electromagnetic (EM) instrumentation to map the water. The ground water in the area has quite low resistivities, ranging from 8 to about 42 ohm-meters (Pool and Coes, 1999, plate 3), making the EM component of the airborne geophysical study an effective means for mapping the water in the upper 400 meters of the basin.

Return to top of page.

Mapping the basement from magnetic data

Figure 3, thumbnail image of depth-to-magnetic-source map for survey area with link to larger, 85 kb, version of the map.
Figure 3. Depth-to-magnetic-source map for the entire survey area, and link to larger, 85 kb, version of map.

We used Euler deconvolution (Reid, 1990) to obtain these depths using the combined magnetic data from the two airborne surveys. Gravity data (Halvorsen, 1984; Gettings and Gettings, 1996) had earlier indicated that there was a deep basin underlying Huachuca City in the north of the study area and a raised platform (a horst?) beneath the city of Sierra Vista, among other features. The magnetic data (not shown here) only shows cultural anomalies over Sierra Vista and Fort Huachuca, and a large magnetic anomaly over the Tombstone volcanic field shown in blue in the upper right in figure 4).

Figure 4, thumbnail image of 1999 Airborne electromagnetic channel 10 z-axis grid (shallow horizontal conductor) map with link to larger, 108 kb, version of the map.
Figure 4. 1999 Airborne electromagnetic channel 10 z-axis grid (shallow horizontal conductor) map and link to larger, 108 kb, version of map.

The depth-to-crystalline-basement map (figure 3) shows complex structures in the basement, and agrees well with the gravity-modeling-derived map except around Huachuca City. The deep basin "pothole" in the gravity modeling (Gettings and Gettings, 1996) is also visible in the Euler depths derived from the magnetic data, but with an apparent ridge crossing through it (the red-yellow structure running from southwest to northeast, in the northwest part of figure 3). This apparent ridge is almost certainly a Tertiary volcanic flow positioned at an intermediate depth in the sediment stack (Tertiary volcanics outcrop on the edge of the survey area in the Mustang Hills as well as farther northwest; see the geologic map by Kneale and others, 1998). Because of features like this, depth-to-basement maps derived from magnetic data should be interpreted with care, and when available the gravity-model should be the ultimate arbiter. Nevertheless, and despite decreasing porosity with increasing lithostatic pressure, the basement structure reported here is ultimately crucial to a full assessment of the water stored in the Upper San Pedro Basin because it gives a sense of the available volume of porous materials that could hold water.

Return to top of page.

Mapping the aquifer using airborne EM

Modern 20-channel, low-noise digital airborne multicoil EM systems can provide direct information on the pervasive horizontal conductor that we interpret to be the Upper San Pedro Basin aquifer. We can effectively view depth-slices in this conductor by comparing short-time vs. long-time gate-channels in the received secondary signal. Figure 4 shows the channel-10 conductor over the survey area; this is an early-time channel, generally indicative of sources having a shorter time-constant, reflecting higher conductivity and shallower depth extent. We interpret this conductor to be a view of the shallow part of the Upper San Pedro aquifer. Note the Fort and the city of Sierra Vista showing up as "speckly" electrical-culture anomalies on the left side of the figure; the roads-with-power-lines are particularly noticeable. Note also how the Tombstone Hills volcanic province is cut by the San Pedro River (the heavy north-south black line), but nevertheless effectively deflects the shallow Upper San Pedro aquifer (the red zone) away from the northeast while the river nevertheless runs right through it.

Figure 5, small image of conductivity depth profile (C.D.T.) from 1997 flight line #122 with link to larger, 20 kb, version of the diagram.
Figure 5. Conductivity depth profile (CDT) derived from 1997 flight line #122, with link to larger, 20 kb, version of the diagram.

The multi-channel, low-noise nature of these data also allow for the generation of a relatively new product - Conductivity-Depth-Transforms, commonly called CDTs (Wolfgram and Karlik, 1995; see figure 5 for an example CDT with an interpretation. When parameters are properly set, these can allow us to map in three dimensions the horizontal conductors such as the aquifer underlying the Upper San Pedro Valley. In the area of interest the water has been measured at 236 - 1,210 micro-Siemens/cm; see Pool and Coes, 1999. Viewed in cross-section, we can use CDTs to observe remarkable details in the upper San Pedro aquifer, including a band of thick unsaturated sediments between the recharge areas at the Huachuca Mountain front in the southwestern part of the study area and the San Pedro aquifer farther east, as well as the shallowing of the water-table as it approaches the San Pedro River farther east still. figure 6 shows a fence-diagram assembled using a series of CDT images, where the red appears to map the conductive water-silt-clay mixture of the Upper San Pedro aquifer. Comparisons of these CDTs with resistivities from E-Logs show that they generally correlate quite well down to about 150 meters depth. While some CDTs (those free of cultural interference) show conductivities down to 400 meters depth, we have found that the resistivities on the CDTs increase with depths of more than about 150 meters - and increasingly diverge from the E-Logs at greater depths. Nevertheless, we believe we can qualitatively use the CDTs to map the aquifer down to 400 meters depth in areas where there is no human cultural interference. In general, sands and gravels accumulate near the mountain fronts, while silts and clays tend to dominate the basin fill in the center of the valley.

Figure 6, conductivity versus depth fence diagram, 1997 and 1999 electromagnetic surveys, and link to larger, 579 kb, version of the diagram.
Figure 6. Conductivity versus depth fence diagram using data from the 1997 and 1999 elelctromagnetic surveys, with link to larger, 579 kb version of diagram. Diagram is superimposed on a topographic map of the study area.

Return to top of page.

The main problems encountered using CDTs to interpret the Upper San Pedro aquifer are that human cultural interference (power lines, pipelines, etc.) can locally render the CDT results largely unusable. This is because induced currents through man-made conductors, with ground-return through the Earth, give rise to strong, short-time secondary signals that saturate the measured response from deeper conductors; in effect, the shallow cultural noise swamps the deeper geologic signal.

In addition, near the Huachuca Mountain front there is a deep unsaturated zone, and we can use the CDTs to map what may be a deep recharge path to the Upper San Pedro aquifer farther to the east through a deeper, older lower-porosity unit called the Pantano Formation. In most cases, the first conductive zone in the CDTs maps the water table in sands and gravels. In the few cases where the CDTs map a conductor that lies below the apparent water table, we believe this is due to the presence of a confined aquifer dominated by clay in the center of the basin. In this environment, the well has penetrated a sand-and-gravel zone below the equipotential, and thus frees the water to fill up the well (only) to the equipotential level; the CDTs apparently map where the bulk of the water is, in one case up to 50 meters below the equipotential. In this confined aquifer zone the term "water table" is no longer really relevant; the water simply moves to the equipotential whenever it is freed to do so by a well.

The issue of matching the CDTs to the E-Logs raises the question of whether we might be able to "tune" the airborne EM inversion process to provide more accurate information below 150 meters depth. In an effort to explore this, we used Encon's EM-Flow1 software to process a test profile (the "Mexico Line" in figure 6) acquired parallel to, but several kilometers north of, the Mexican border just south of the main survey area. figure 7 shows a comparison between the Geoterrex-Dighem CDTs1 and the EM-Flow CDIs (which stands for "Conductivity Depth Inversions," see Macnae and others, 1991). From this figure we can see that the EM-Flow inversion provides somewhat more resolution at shallow depths, but the Geoterrex-Dighem CDTs appear to reach deeper and provide information on a second conductor, a deeper branch of the main aquifer on the western (left) edge of the profile.

Figure 7, thumbnail image comparing the E.M.-flow inversion C.D.I.s with the Geoterrex proprietary inversion C.D.T.s and link to larger, 14 kb, version of the image.
Figure 7. Thumbnail image that compares Geoterrex-Dighem CDTs (conductivity depth profiles)1 and the EM-Flow CDIs (conductivity depth inversions, see Macnae and others, 1991), with link to larger (14 kb) version of image.

Return to top of page.


From our experience with the Upper San Pedro Valley airborne geophysical surveys, we have learned the following:

1. Magnetic data can be used to extract depth-to-crystalline basement beneath the sediments, subject to the proviso that in some cases we are actually seeing shallower Tertiary volcanic flows intercalated in the sediment-stack as a sort of "geologic noise".

2. Early and late off-time channels in the airborne EM data can be used to map different levels of the conductive body that we interpret to be the Upper San Pedro Valley aquifer.

3. Conductivity Depth Transforms (CDTs) can be used to derive cross-sections of the aquifer - in effect, we can derive a 3-D map of the subsurface conductivity. However, where the aquifer is confined, shallow sediments above it are dry and hydraulic transmissivity is restricted by clays. In this case, the "water table" measured in wells is actually the hydraulic equipotential line, and the CDT maps where the bulk of the water lies below it.

4. The CDTs agree quite well with E-Logs where we were able to compare them (note the equipotential exception in item #3 above, however). The CDT and E-Log data agree down to about 150 meters depth, but below this the CDT resistivities increase with increasing depth down to the noise limit (up to 400 meters in many cases). This means we only have quantitative data down to 150 meters, and qualitative data from 150 to 400 meters depth. It is theoretically possible to "tune" the EM inversion process to better match the conductivities below 150 meters, but our own testing was inconclusive due to the proprietary nature of the CDT algorithm and inherent constraints of the EM-Flow inversion software.

Return to top of page.


Brown, S.G., E.S. Davidson, L.R. Kister, and B.W. Thomsen, 1966, Water Resources of Fort Huachuca Military Reservation, southeastern Arizona: USGS Water-Supply paper 1819-D, 57 p.

Gettings, P.E., and Gettings, M.E., 1996, Modeling of a magnetic and gravity anomaly profile from the Dragoon Mountains to Sierra Vista, southeastern Arizona: U.S. Geological Survey Open-File Report 96-288, 15p., 3 pl.

Halvorsen, P.H., 1984, An exploration gravity survey in the San Pedro Valley, southeastern Arizona: Master's thesis, University of Arizona, 70 p.

Kneale, Sean; Bultman, Mark; Chan, Tammy; Grisolano, Anna; and Hirschberg, Doug, 1998, Coronado National Forest Geology, U.S. Geological Survey Administrative Report, (CD-ROM release).

Macnae, J.C., Smith, Richard, Polzer, B.D., Lamontagne, Y., and Klinkert, P.S., 1991, Conductivity-depth imaging of airborne electromagnetic step-response data; Geophysics, 56, pp. 102-114.

Moore, R.B., 1993, Geologic map of the Tombstone volcanic center, Cochise County, Arizona: U.S. Geological Survey Miscellaneous Investigations Series Map I-2420, scale 1:50,000

Pool, D.R., and Coes, A.L., 1999, Hydrogeologic investigations of the Sierra Vista subbasin of the Upper San Pedro River Basin, Cochise County, Arizona: U. S. Geological Survey Water-Resources Investigations Report 99-4197, 47 p., 3 plates.

Reid, A.B., Allsop, J.M., Granser, H., Millett, A.J., and Somerton, I.W., 1990, Magnetic interpretation in three dimensions using Euler deconvolution: Geophysics 55, pp. 80-91

Wolfgram, P., and Karlik, G., 1995, Conductivity-depth transform of GEOTEM data: Exploration Geophysics, 26, pp. 179-185.

Return to top of page.

1 Disclaimer: References to companies or specific equipment in this paper are provided for information purposes only, and do not imply endorsement by the U. S. Geological Survey of the Department of the Interior.