There is evidence of rainwater recharge at WIPP test wells. A steady rise in water levels in the Magenta at test wells H-2 and H-3, and in the Culebra at H-1, H-3 and H-4, all located deep within the WIPP site, were recorded between mid-1977 and mid-1981 (SAND 83-0210, pp. 22-25). This occured in undisturbed hydraulic conditions, before the sinking of the first WIPP shaft in July 1981 (CCA, Appendix TFIELD, Table TFIELD-9). DOE says that this rise in hydraulic heads is unexplained. (CCA, Appendix TFIELD, p. TFIELD-17). Phillips (1998a, p. 17) offered an explanation. During this four-year period 68.55 inches of rain (17.14 inches per year) was recorded in Carlsbad, compared to an average of 10.85 inches of rain per year during the preceding 25 years. While the rise in Magenta and Culebra water levels cannot be correlated with individual rainstorms, it can be correlated with short-term trends in precipitation.
DOE's model of the Culebra dolomite as a confined aquifer, receiving negligible rainwater recharge, is inconsistent with groundwater geochemistry. If the Culebra contained only fossil water left over from the ice ages, it would be saturated, or nearly so, with total dissolved solids (TDS). To the contrary, total dissolved solids in Culebra groundwater, within the WIPP site, vary by a factor of 25 -- from 230,000 mg/l at H-15 to 8,890 mg/l at H-2b. These two test wells are only 1.66 miles apart. When the Culebra test wells are plotted on a map (Phillips, 1997b, Figure 1) [Exhibit 34], the contour lines display a zone of high TDS in the northeastern part of the WIPP site, where the Santa Rosa sandstone is present and water is found in the Dewey Lake Redbeds, with TDS steadily decreasing to the southwest, where the Santa Rosa is absent and water is found in the Dewey Lake Redbeds. This is consistent with the interpretation that Culebra groundwater becomes mixed with increasing amounts of fresh water as it approaches Nash Draw, because the hydrologic regime is increasingly karstic (Phillips, 1997b, pp. 9, 11).
When dissolved sodium and chloride in Culebra groundwater are plotted on a map (Phillips and Snow, 1997, Figure 1) [Exhibit 48, the contour lines show that dissolved halite in Culebra groundwater decreases steadily from east to west across the WIPP site. This should demonstrate that Rustler groundwater is not old, and that freshwater recharge is occurring (Id., p. 4). Water saturated with halite, incapable of dissolving significantly more halite, contains approximately 318,000 mg/l of dissolved sodium and chloride (EEG-31, 1985). Within the WIPP site, the highest measurement of dissolved halite in Culebra groundwater, at test well H-5b, is 139,000 mg/l, well below saturation (Phillips and Snow, 1997, p. 4).
When the progression of halite dissolution is plotted on a map, along with the distribution of salinity in Culebra groundwater (Phillips and Snow, 1997, Figure 2) [Exhibit 6], it is shown that some test wells contain dissolved halite in Culebra groundwater where there is no halite in the Rustler. These wells (H-6, P-14, WIPP-25, WIPP-26) are located to the west of the Rustler dissolution front. If groundwater is supposed to be flowing north to south, as DOE contends, how then did dissolved halite appear in these test wells? There is halite in the Rustler only to the east, which implies a westerly component to groundwater flow (Id., pp. 4, 8).
Multiwell pump tests have revealed potential migration pathways for contaminated water from the WIPP site. The procedure is to pump water from one test well, and to monitor the water levels in other wells, to see if there is a response. If the wells are hydraulically connected, the water level will drop in the monitoring well. Likewise, the water level will rise in the monitoring well after the pumping stops. Such multiwell pump tests have revealed hydraulic connections between test wells H-3, DOE-1 and H-11 in the southeastern part of the WIPP site, and between DOE-2, WIPP-13 and H-6 in the northwestern part of the WIPP site (Phillips, 1997a, Figure 2) [Exhibit 49].
These same test wells have shown anomalously high transmissivity in the Culebra dolomite. Transmissivity is the ability of an aquifer to transmit water. It is equal to the thickness of the aquifer multiplied by its hydraulic conductivity. Anomalously high transmissivity in some boreholes is exactly what one would expect in a karstland. A well that taps a connection with the karst conduit system can produce very large quantities of water with negligible drawdown, leading to extremely large calculated transmissivity, whereas a well drilled a short distance away in unfractured, undissolved rock may yield negligibly small amounts of water (Phillips, 1997a, p. 3) (Phillips and Snow, 1998, p. 3).
It is likely that none of the calculated transmissivities in the Culebra east of Nash Draw are representative of karst conditions, because none of these test wells were reported to have encountered cavernous zones in the Culebra. Cavernous zones were found at WIPP-33 in the Magenta and higher strata, but the WIPP-33 borehole was never converted to a test well, despite promises to USGS that WIPP-33 would be available for hydrologic testing (Notes of the Interagency Program Review for WIPP, February 12 and 13, 1979, p. 14). The caverns at WIPP-33 were inferred by: (1) a precipitous drop of the drilling equipment; (2) lost circulation of drilling fluid; and (3) no core recovery (SAND 80-2011, pp. 8, 11, 15-17, C-3, C-4).
If cavernous zones are present in other WIPP boreholes, the same three criteria should apply. Unfortunately, an examination of the basic data reports for other WIPP boreholes reveals that drilling time and lost circulation are rarely noted, so other criteria must be utilized. Lost circulation, washout, loss of core, and/or dissolution residue was reported by drillers at 17 boreholes in the Forty-Niner, 11 boreholes in the Tamarisk, and 22 boreholes in the lower unnamed member, all inside or within one mile of the WIPP site (Phillips, 1997a, Figure 4) [Exhibit 50]. Such consistent occurrence indicates that these zones are poorly consolidated, probably transmissive, and possibly cavernous (Phillips, 1997a, pp. 7-10). Water was observed seeping into the WIPP ventilation shaft from the Forty-Niner member; test well H-1 yielded as much water in the Tamarisk member as in the Magenta or Culebra; and test well H-3 yielded as much water in the lower unnamed member as in the Magenta or Culebra (EEG-32, 1985, pp. 37, 39). This shows that all members of the Rustler are water-bearing, at least in places, and all are involved in groundwater transport.
The groundwater flow path from H-3 to DOE-1 to H-11, primarily through the Culebra and the lower unnamed member, has been modeled by DOE as far as the WIPP site boundary, although DOE has never conceded that this flow path turns westward toward H-7 in Nash Draw. Phillips and Snow (1997, pp. 10-11) have described this groundwater flow path in detail. However, there is another groundwater flow path, from the WIPP repository to Nash Draw, which has not been modeled by DOE. The multiwell pump test centered at WIPP-13 (SAND 87-2456) has demonstrated a hydraulic connection between the WIPP exhaust shaft and WIPP-25 in Nash Draw (CCA, Appendix SUM, Table 4.7), by way of WIPP-13. The response time between WIPP-13 and WIPP-25 was extraordinarily rapid -- a delay in maximum drawdown of only 26 hours between test wells nearly four miles apart. The apparent transmissivity between WIPP-13 and WIPP-25 is extremely high, 650 ft2/day (CCA, Appendix SUM, p. 114), higher than at either WIPP-13 (72.0 ft2/day) or at WIPP-25 (270 ft2/day). Located almost exactly midway between WIPP-13 and WIPP-25 is the WIPP-33 sinkhole, which would explain the extremely high transmissivity.
The WIPP-13 multiwell pump test was centered in the Culebra, and all the monitoring wells were in the Culebra. This is unfortunate, because there is strong evidence that this groundwater flow path is primarily through the Magenta and higher strata. At H-3 the Magenta produced 6 gallons per minute with a 6-foot drawdown, compared to 25 gallons per day in the Culebra (CCA, Appendix GCR, p. 6-53, Table 6.3-4). Transmissivity in the Magenta at H-3 has been calculated at 330 ft2/day (David Snow, personal communication), compared to 19 ft2/day in the Culebra (CCA, Appendix HYDRO, Table 7). At WIPP-13 the Magenta is broken and shattered, whereas the Culebra is not (SAND 82-1880, pp. 11-13). At WIPP-33 five water-filled caverns were found in the Magenta, Forty-Niner, and Dewey Lake; none were in the Culebra (SAND 80-2011, pp. 11, 15-19). At WIPP-25 transmissivity in the Magenta was measured at 375 ft2/day, compared to 270 ft2/day in the Culebra (CCA, Appendix HYDRO, Table 7). It is wrong for the Detection Monitoring Program to disregard these potential pathways and to ignore the Magenta altogether.