CAVERNOUS ZONES AT
THE WIPP SITE
Dr. Richard Phillips
University of Oregon
The Waste Isolation Pilot Plant (WIPP) near Carlsbad, New Mexico, is intended for the permanent disposal of radioactive waste from nuclear weapons production. The project is the brainchild of the United States Department of Energy (DOE). The waste is to be buried in steel drums and plywood boxes in direct contact with salt beds of the Salado Formation. The project lacks engineered barriers; it is the rocks themselves--the Salado Formation and the aquifers of the overlying Rustler Formation and Dewey Lake Redbeds [shown in cross-section in Figure 1]--which are intended to be the effective barriers to radionuclide migration.
The Culebra dolomite member of the Rustler Formation has long been recognized as the most likely pathway for contaminated water to travel from the WIPP site to the accessible environment. The Culebra dolomite is not a porous, homogeneous medium; groundwater does not move uniformly and predictably through interconnected pore spaces in the rocks. The Culebra, in most places, is highly fractured, and the effective groundwater flow paths are through the largest fractures, the paths of least resistance.
Dolomite is a soluble rock; it slowly dissolves when exposed to fresh water. As fractures become enlarged by solution, they become even more effective groundwater flow paths. Over time, more and more groundwater flows through fewer and fewer solution-enlarged fractures. Ultimately, these solution conduits become underground caverns, capable of carrying groundwater quite rapidly with little resistance. This type of groundwater hydrology is known by geologists, world-wide, as karst.
The problem with karst as a waste disposal environment is that some radionuclides may travel unretarded, at the speed of water. The larger the aperture, or diameter, of the solution conduits, the less contact the radionuclides will have with the surrounding rock, and the less the amount of radionuclide retardation. The ability of the Rustler Formation to retard significantly the migration of radionuclides depends upon the absence of karst conditions, of channelized flow, at the WIPP site.
Recent pumping tests at hydrologic test wells in Culebra dolomite at the WIPP site (Beauheim, 1986; Beauheim, 1989) have resulted in unexpectedly short response times between certain test wells. For example, when water was pumped from test well DOE-2, there was a drop in water level within two hours at test well WIPP-13, which is 4835 feet from DOE-2. Test well H-6, which is 10,150 feet from DOE-2, responded within one day. Other test wells, no farther away (e.g. WIPP-12, WIPP-18 and H-5), showed no response at all. This indicates that DOE-2, WIPP-13 and H-6, in the northern part of the WIPP site, are hydraulically connected by channelized flow in the Culebra dolomite. Similar results indicate a hydraulic connection between test wells H-3, DOE-1 and H-11 in the southeastern part of the WIPP site, with measurable response at test wells H-15, H-17, P-17 and Cabin Baby, and little or no response at test wells H-4, H-12, H-14, P-15 and P-18 [Figure 2]. These hydraulic connections are not necessarily fracture networks, as interpreted by Beauheim (1986, 1989); the response was so rapid that they could be karst channels.
Hydraulic conductivity is the velocity at which water moves through an aquifer. Transmissivity is the rate at which water is transmitted by the aquifer; it is equal to the hydraulic conductivity times the thickness of the aquifer. The highest measured values for transmissivity and hydraulic conductivity in the Culebra dolomite aquifer at WIPP test wells are given in Table 1.
It has long been recognized that, in the Culebra dolomite, transmissivity varies by five orders of magnitude in the vicinity of the WIPP site, from 0.002 ft2/day at test well P-18, located 4547.3 feet east of the WIPP site, to 233.0 ft2/day at test well P-14, located 4664.2 feet west of the WIPP site (Haug et al., 1987). It is often represented that transmissivity in the Culebra dolomite increases steadily from east to west, but this is not the case. When measurements of transmissivity are plotted on a map [ Figure 3], contour lines cannot be drawn, not even on a logarithmic scale. Rather, it becomes apparent that, among test wells inside or within one mile of the WIPP site, seven show anomalously high transmissivity in the Culebra dolomite, one to three orders of magnitude higher than in any others. These wells include, besides P-14, the very six test wells shown by pumping tests to be hydraulically connected (DOE-2, WIPP-13, H-6; H-3, DOE-1 and H-11).
Anomalously high transmissivity in some boreholes is exactly what one would expect in a karstland. A borehole which misses one of the active solution conduits should show values which are much less than the average. This applies to almost all boreholes in a karstland because the area of active solution conduits is only a small part of the total area. It is possible, indeed likely, that none of the measured transmissivities within one mile of the WIPP site are representative of karst conditions in the Culebra dolomite, because none of these test wells were reported to have encountered cavernous zones in the Culebra.
However, cavernous zones were encountered at WIPP-33, a borehole located 2753.4 feet west of the WIPP site. There were five caverns in all: four in the Rustler Formation (two in Magenta dolomite, two in Forty-Niner gypsum), and one in the Dewey Lake Redbeds. The caverns were inferred by: (1) a precipitous drop of the drilling equipment (zero minutes per vertical foot); (2) a loss of circulation of drilling fluid; and (3) no core recovery. A camera was then lowered into the borehole, which confirmed the presence of underground caverns. If cavernous zones are present in other WIPP boreholes, the same three criteria should apply. Unfortunately, an examination of the geophysical logs and lithologic descriptions for other WIPP boreholes reveals that the first two criteria--drilling time and lost circulation--are rarely noted. Nor can cavernous zones in the Culebra and Magenta dolomite members of the Rustler Formation be detected by a lack of core recovery alone, because the Culebra and Magenta are typically fractured, which makes core recovery difficult. However, a correlation of geophysical logs and lithologic descriptions does reveal a consistent lack of core recovery at two other stratigraphic horizons in the Rustler Formation: