Review of U.S. Department of Energy Technical Basis Report for Surface Characteristics, Preclosure Hydrology, and Erosion (1995)

This chapter provides a review of the sections of the technical basis report (TBR) related to preclosure hydrology, including flooding potential, ground water conditions in the unsaturated zone, and water resource potential. These issues are discussed in Chapter 2 and Chapter 3 of the TBR (see Table 1.1 of this report), but are grouped and discussed together in this chapter because they address a common topic—the potential impact of hydrology on the proposed repository during construction and operation.

The preclosure hydrology sections of the TBR address the following three questions:

1. Can surface floods enter the access tunnels or shafts or jeopardize surface operations at the portals during construction and operation of the proposed repository?

2. Can the repository be constructed, operated, and closed by using reasonably available technology to control ground water intrusions?

3. Is the ground water supply at Yucca Mountain sufficient to provide for the needs of the proposed repository during construction and operation?

The potential for surface flooding at Yucca Mountain is discussed in Section 2.6 of the TBR. The focus of this discussion is whether the proposed repository can be constructed, operated, and closed using reasonably available technology¹ (RAT) to protect against surface flooding. The TBR presents estimates of the water surface elevations associated with the probable maximum flood (PMF) to demonstrate that flooding does not pose a threat to three repository facilities: (1) Shaft Site No. 2, which is potentially affected by flooding in Drillhole Wash or Coyote Wash; (2) South Portal Site and Pad, which are potentially affected by flooding in an adjacent wash; and (3) North Portal Site and Pad, which are potentially affected by flooding in Midway Valley Wash (see Figure 2.6.2-1 of the TBR and Figure 1.2 of this report).

The probable maximum flood is an estimate of the largest flood that can occur at a site under current climatic conditions. The PMF is estimated by using a number of very extreme or upper bounding values for precipitation and runoff. The PMF procedure is used routinely by engineers in the design of flood-prone engineering structures such as dams.

Estimation of the PMF for a site can be envisaged as a three-step process: (1) the probable maximum precipitation (PMP)² event at the site is estimated by using extreme but possible events based on regional weather records; (2) surface runoff and the resulting streamflow from this precipitation event are calcu-

lated; and (3) the flow depth at selected channel cross sections is estimated from this discharge.

PMFs for selected areas at Yucca Mountain were estimated using these procedures. Precipitation estimates for Yucca Mountain were taken from a National Weather Service report (NWS, 1977) that provides PMP estimates for the Colorado River and Great Basin drainages. The PMP event estimate was used to calculate the volume of runoff from specific drainage areas at Yucca Mountain. Unit hydrographs³ were developed based on the standardized regional flood graphs for the southwestern United States used by the U.S. Army Corps of Engineers. The unit hydrographs and runoff estimates were then used to generate flood hydrographs for a number of drainages. The flood hydrographs were estimated by using standard Bureau of Reclamation procedures as described by Bullard (1991). The maximum discharge rate on the hydrograph is called the peak discharge rate, the peak discharge, or the peak runoff rate. This peak discharge rate was converted to a depth of flow at selected stream channel cross sections through flood routing procedures, which require estimates of water velocity and channel roughness. The maximum predicted depths of flow at these cross sections were compared directly to the elevations of the portals, pads, and shaft to determine their flooding potential.

The procedures used to estimate the PMP and peak discharge rates are relatively straightforward engineering calculations. However, flood routing is less standardized and requires expert judgment in the estimation of input parameters to a hydraulic

simulation model, in this case the Flood Hydrograph and Routing algorithm of the Bureau of Reclamation. The critical parameters of interest in the model are a hydraulic resistance coefficient, called Manning 's n, and the Froude number, a physically significant dimensionless ratio that determines the depth and velocity characteristics of the flow. ⁴ A third important input parameter to this model, the “bulking factor,” is used to represent the effects of entrained sediment, air, and debris carried by the flowing water on the flow depth.

The application of probable maximum flood procedures in the TBR to estimate flood events is consistent with practices used to design civil structures such as bridges and dams. However, the TBR descriptions of these procedures, particularly the data and assumptions used in flood routing and water surface elevation estimation, are too brief to allow evaluation of the adequacy of data collection and analysis. The committee had to consult references cited in the TBR (i.e., Bullard, 1991; Blanton, 1992; Glancy, 1994) to determine exactly which procedures and data were used. For example, the TBR does not mention an important assumption used in the calculations: namely, that critical flow depths are assumed in cases where the Manning equation⁵ predicts supercritical flow velocities. The committee discovered this assumption during its review of the references cited in the TBR.

The TBR is limited in its descriptions of PMF calculations. Consequently, it does not provide explicit support for the details of the PMF procedure, including the runoff response, flood routing, and parameter estimates and assumptions. Estimates of PMF flows were made for several locations in the washes near the North Portal Pad, South Portal Pad, and Shaft Site No. 2 by using Blanton's (1992) procedure, which assumed a bulking factor of 2. Although not stated explicitly in the TBR, it appears that many calculations were based on a Manning's n value of 0.045. The assumption of critical flow velocity is conservative for this assumed value of Manning's n (as noted in the last section, critical flow depths were assumed in cases where the Manning equation predicted supercritical flow velocities). That is, the use of this Manning's n value with a critical flow velocity produces a greater flow depth (i.e., a “larger” flood) than would be expected under supercritical flow conditions.

Bounding curves for maximum peak discharge shown in Figure 2.6.3-1 of the TBR and those from Glancy (1994, derived from the equation given on p. 28 and the table on p. 29 of that paper) are generally consistent by a factor of 2 or less, except on very small watersheds. Given the nature of these kinds of calculations, these results seem reasonable to the committee.

It would have been helpful to readers to provide more discussion regarding the procedures and rationale for the selection of Manning 's n value and the bulking factor. Also, the committee notes that the water surface elevations for the PMFs shown in Figure 2.6.2-1 and Figure 2.6.2-2 of the TBR are difficult to interpret. Clearer graphics showing PMF elevations also would be helpful to readers.

Comparison of PMF peak discharge estimates (open square symbols in Figure 2.6.3-1 of the TBR) with the curve used to represent maximum regional floods (Figure 2.6.3-1, upper curve) shows that the PMF-based water surface elevations are conservative: that is, they provide maximum depths of flooding. Given the high level of uncertainty in estimating flood peaks in arid areas (e.g., by applying six methodologies, Glancy, 1994, p. 29, produced estimates with variations as great as a factor of 3 for a specific site), the differences in flood peak discharges using the methods cited in the TBR appear reasonable.

Figure 2.6.2-2 of the TBR shows that parts of the pad at the North Portal Site are within PMF inundation boundaries. The last two sentences of Section 2.6.2 of the TBR (p. 2-12) state, “A portion of the existing North Portal Pad to support ESF⁶ operations is located in the flood-prone area (Figure 2.6.2-2). This will be further evaluated in the Guideline Compliance Assessment to ensure that RAT can mitigate potential hazards for the repository surface facilities.” The committee was unable to determine whether this pad could be protected from flooding by using RAT, although such protection would appear to the committee to be relatively straightforward.

An alternative hypothesis is that the PMF is significantly different than estimated because of uncertainties in runoff response, Manning 's n, and the bulking factor. This could affect the interpretation of flooding potential at critical surface facilities at Yucca Mountain.

The alternative hypothesis can be tested easily through sensitivity analysis of the estimated water surface elevations to assumed values of runoff response, Manning's n, and the bulking factor, and the assumption of critical flow depth at all locations in the stream channels. Bounding calculations involving the systematic variation of the Manning coefficient around the value of 0.045 assumed in the TBR and the bulking factor from 1.0 to more than 2.0 would show the sensitivity of estimated water surface elevations to the assumed values of these parameters. This would provide additional insight into the uncertainty of the estimated water surface elevations.

The use of probable maximum flood estimate procedures in the TBR is consistent with general engineering practice. Assumed values of runoff response, Manning's n (0.045 in all channels), critical velocity, and the assumed bulking factor of 2, while likely conservative, are not supported by data or documentation in the TBR. Nor were they justified by field measurement or citation to the literature in the references provided in the TBR. More documentation and explanation of procedures used to calculate PMFs are needed in the TBR, as are graphics to illustrate PMF elevations in relation to critical facilities.

More work should be done to assess the sensitivity of the estimated water surface elevations to the assumed values of Manning's n and the bulking factor, and the assumption of critical flow depth at all locations in the stream channels. This testing will

provide additional insight into the effects of these assumptions on estimates of flooding potential at the sites of interest.

A description of the unsaturated zone hydrostratigraphy at Yucca Mountain and the potential for intrusion of ground water into the proposed repository is addressed in Section 3.2 of the TBR. Subsurface water intrusions during repository construction and operation could conceivably arise from three sources: (1) deep percolation of infiltrating surface water through continuous cracks or fractures; (2) discharge of subsurface water under positive pressure stored above the repository (perched water); or (3) rise of the underlying water table to the level of the repository. Because the TBR deals only with preclosure issues, potential adverse effects are to be evaluated within the projected 50- to 100-year life of repository operation prior to closure. ⁷

The TBR does not explicitly address whether rising water tables or deep percolation of infiltrating surface water could cause flooding of the repository during the preclosure phase. Although neither of these two sources appears to the committee to pose a problem for construction or operation, the report should have demonstrated this. The percolation rate in this area is extremely

low (Montazer and Wilson, 1984), and there is no evidence of steady deep seepage in appreciable amounts (Czarnecki, 1985). Substantial amounts of data have been collected on infiltration rates at the site, and age dating of water in the infiltration zone and water extracted during deep borehole construction has been attempted, but these data are not discussed in the TBR.

The proposed repository is to be located about 200 m (about 650 feet) above the present level of the water table, so significant volumes of recharge would be required to raise the water table level to the repository. It has been argued that there is historical evidence for mean annual precipitation amounts up to 100 percent above the current mean annual precipitation in the area (Spaulding, 1985). Modeling of the steady-state position of the ground water table for a doubling of mean-annual precipitation showed that the repository remained in the unsaturated zone, about 30 m (approximately 100 feet) above the final position of the water table (Czarnecki, 1985). However, it is worth noting that in these simulations, the predicted steady-state ground water recharge rate increased by up to a factor of 15 compared to present-day recharge.

A number of untested simplifying assumptions were made in this model. The model calculation assumed that the ratio of runoff to infiltration was identical to present-day conditions. The calculation also does not include the effects of increased precipitation on unsaturated zone fluxes through large flow channels or on perched water storage changes. A change in hydrologic response, such as decreased runoff with increasing precipitation (e.g., due to the retardation by vegetation), would have resulted in increased recharge. An empirical relationship between precipitation and recharge was also used in the model. This relationship has not been tested over the range of inputs used in the calculation. Further, the

model predicts only the steady-state position of the water table. It does not address the question of how much time would be required for the water table to re-equilibrate after precipitation is doubled.

Thus, the model exercise has limited value for making quantitative estimates of water table responses to a doubling of mean annual precipitation. Because the time frame of the preclosure phase is short (50-100 years), water entry into the repository from infiltrating water does not appear likely. This could have (and should have) been demonstrated in the TBR.

Section 3.2.1 of the TBR describes the principal hydrostratigraphic units in the unsaturated zone of Yucca Mountain. The narrative descriptions of these units are probably adequate for the purposes of this TBR —although they rely on papers by Montazer and Wilson (1984) and Weeks and Wilson (1984), which were published prior to most of the detailed site characterization studies of the unsaturated zone. It would have been useful to include additional estimates of hydraulic conductivity, porosity, and water saturation that have been determined from more recent core sampling and borehole logging. The “conceptual” hydrogeologic section presented in Figure 3.2.1-1, which is based on the early work of Montazer and Wilson (1984), is grossly inadequate for evaluating the potential relationships between perched water zones and stratigraphic or structural features. An accurately scaled cross section that shows elevations of the land surface and locations of boreholes, the proposed repository, and the water table should have been provided in the TBR to facilitate understanding.

The TBR discusses the appearance of perched water in various boreholes and concludes that (1) all perched water is below the proposed repository depth, although it is possible that it could be encountered during construction and operation, and (2) perched water does not appear in sufficient quantity to constitute a

problem that cannot be handled by RAT. The sources of information used to identify subsurface water storage are various deep boreholes in the vicinity of Yucca Mountain extending to the water table and shallow neutron access tubes confined to the surface alluvium. The methods of data collection appear to be up-to-date. Dry drilling technology was used to construct most of the boreholes, which eliminated drilling fluid contamination problems. Standard methods of isotope analysis were used estimate water ages. However, little of this information is provided in the TBR.

As noted above, the TBR concludes that perched water does not pose a problem for preclosure operations that cannot be solved by RAT, and the committee believes that this conclusion is reasonable. However, the TBR does not use available data to best advantage in drawing this conclusion, and some statements of questionable accuracy appear in the discussion. The brief discussion of perched water on pages 3-4 and 3-5⁸ of the TBR would have been clearer and more effective if the report had included a figure showing the locations of the wells and the geologic cross sections identifying the perched water locations. The committee was shown a poster during the field excursion that would have been adequate for this purpose. The statement made in the TBR on page 3-4 that perched water was encountered 100 m (approximately 330 feet) above the water table in all cores is wrong, since UZ-14 encountered perched water at an altitude of 965 m (3,165 feet) above mean sea level (TBR, p. 3-5), which is some 187 m (613 feet) above the water table at that location and

between 29 m (95 feet) and 87 m (285 feet) below the proposed repository. ⁹ This hole is about 1 km (approximately 3,300 feet) north of the site. The absence of graphics or even a comprehensive table summarizing perched water locations made it extremely difficult for the committee to discern the spatial distribution of perched water from the discussion in the TBR.

No attempt was made in the TBR to analyze perched water locations relative to geologic features in the unsaturated zone, even though such an analysis might assist in estimating its distribution and extent. The zones of perched water appear to occur above hydrostratigraphic units of lower permeability, which is logical, and may also accumulate at or near fault boundaries. Accurately drawn cross sections would have illustrated such relationships.

No information on perched water volumes or its rate of entry to the boreholes is provided in the TBR. Although most of the perched water discovered was insignificant in volume, wells UZ-14 and SD-7 produced large volumes of perched water [6,000 and 22,000 gallons (23,000 and 83,000 liters)] in short (67- and 90-hour) pumping tests. Moreover, the perched water in UZ-14 behaved as though it were part of a large reservoir (Burger and Scofield, unpublished manuscript cited in the TBR). This information could have been useful in making bounding calculations about the possible impacts of encountering perched water zones during repository construction and operation.

The discussion of RAT is very brief and would have benefited from a discussion of water volumes typically encountered in mining operations and how they are handled. Rather than take an analytical approach, the TBR simply states that RAT could deal with any anticipated problems. The committee has made no attempt to evaluate RAT and merely notes that to enhance the completeness and credibility of the analysis, the TBR should explicitly address RAT in the context of the available information on perched water at the site.

There are two places in which the TBR has glossed over points of relevance to the issues surrounding perched water. First, the extremely high water table levels in wells G-2 and WT#6 to the north of Yucca Mountain are never mentioned in the perched water discussion, even though one alternative interpretation for their existence is that they are in fact perched water, and not part of the regional water table (e.g., Ervin et al., 1994). The proposed location for the repository is at least 200 m (approximately 650 feet) above the water table that lies directly beneath it. However, the elevation of the water table in wells G-2 and WT#6 is about 20 m (66 feet) below the elevation of the northern end of the upper block of the proposed repository and about 36 m (118 feet) above the elevation of the northern end of the lower block of the proposed repository,¹⁰ although at least 2 km (1.2 miles) north of the repository boundary.

The paper by Fridrich et al. (1994), which was not cited in the TBR, discusses several possible explanations for the apparent rise in water table elevation north of Yucca Mountain. These authors did not discuss perched water as one of the explanations, although it was proposed by Ervin et al. (1994), which also was not cited in the TBR. Interestingly, the water table elevations observed in wells G-2 and WT#6 are not too different from the elevation of perched water (965 m [3,165 feet] above mean sea level; TBR, p. 3-5) in the northernmost well (UZ-14) in which it has been detected.

The second omission in the TBR is the absence of any discussion of the age or genesis of the water found in either the perched water zones or the shallow neutron access holes. There is substantial information from ongoing tritium, ¹⁴C, deuterium, strontium, uranium, and ³⁶Cl analyses of subsurface pore and perched water that suggests different hypotheses for the origin of this water. These data should have been discussed in the TBR, particularly as they bear on the rate of supply of water to the perched zones. The information obtained to date (which was presented to the committee at its information-gathering sessions) suggests that the perched water appears to be quite old. This information would have supported the conclusion that perched water is not a problem for preclosure and would have offered some assurance that the zones are not connected by short travel times to the surface.

More work needs to be done to the north of Yucca Mountain to investigate the nature of the high water table in that area. By pumping wells G-2 or WT#6 and looking for discrete fractures refilling the pumped zone, it should be possible to determine whether the saturated zone is perched. This technique was used to identify the perched water in well UZ-14 (Burger and Scofield, unpublished manuscript cited in the TBR). The information that is used to delineate the high-water-table-gradient zone north of the repository location is very sparse, and the contour lines are subject to great uncertainty. The potentiometric surface determination would, therefore, benefit from an additional well placed in the high-gradient zone. The committee understands that both of these activities are currently under way at the site (Russ Patterson, DOE, personal communication).

Of the three potential sources of subsurface water that could flood the proposed repository (deep percolation, rising water table, and perched water), only the latter is discussed in the TBR. Inclusion of an analysis of the first two sources would have been possible, given the body of available information, and would have added to the completeness of the discussion.

The brief discussion of perched water in the TBR contains no analysis. The TBR should have included figures showing topography, geology, structure (e.g., faults), well locations, perched water, and the ground water table. It also should have provided quantitative estimates of perched water and data sum-

maries of relevant information where available. It does not appear to the committee that perched water will pose problems that cannot be handled by RAT within the 50- to 100-year life of repository operation. Nevertheless, the TBR could have made far more effective use of existing information in making this point.

Water resource potential at Yucca Mountain is addressed in Section 3.3 of the TBR. The questions related to water resource potential are not stated explicitly in the TBR but can be inferred from information in other references. According to the early site suitability evaluation (Younker et al., 1992), two questions must be addressed to evaluate water resource potential:

1. Is adequate water for repository development available from surface or ground water sources?

2. Can legal rights to available water be obtained?

The second of these questions is outside the scope of a scientific peer review, although the answer to this question may depend on a technical evaluation of whether existing ground water resources in the area are being mined (i.e., being pumped at a rate that exceeds the long-term perennial or sustained yield of the aquifer). This report, therefore, concentrates on the first question related to analysis of ground water availability.

Another question related to preclosure water supply is whether the water is of sufficient quality for use at the repository site. The issue of water quality is not addressed in the TBR, pre-

sumably because water quality is known to be adequate or water quality is not of particular concern given its intended use at the repository site.

An explicit definition of the basic questions related to water resource availability is essential to judging the adequacy of data and syntheses; different types of data and different analyses would be required depending on the nature of the technical question. Adequacy of available ground water for repository needs can be assessed by either (1) accurate estimates of the total amount of water available for use or (2) a comparison of bounding estimates of water requirements to bounding estimates of water supplies. Some statements in the TBR suggest that the total available water is to be estimated, but other statements indicate that bounding estimates of requirements and supplies are the objective. It is not clear if “available water” is assumed to include all water in the aquifer or if available water is limited to that which could be produced without significant mining.

At least three aquifers could be tapped for water supply at Yucca Mountain: a shallow alluvial aquifer in Fortymile Wash, the tuff aquifer (shown on the cross sections in Figure 1.3 and Figure 1.4), and a lower confined carbonate aquifer. The production rates and modeling results described in the TBR are only for wells completed in the tuff aquifer, implying that this is the aquifer that will be used to supply repository construction needs. However, the TBR makes no explicit statement that this is the only aquifer being considered for that purpose.

This review of the adequacy of data collection and analysis focuses on three issues: (1) well and production data, (2) numerical modeling, and (3) water requirements for repository construction and operation. These issues are addressed in the following subsections.

The section of the TBR on water resources potential begins on page 3-6 with a review of water supply and monitoring wells within the Jackass Flats area (see Figure 1.1). Wells are described in a chronological narrative rather than in a tabulated summary, which hinders comparisons and spatial understanding. Production rates are reported in a variety of units (liters per second, gallons per minute, cubic meters per day, and acre-feet per year), which also makes direct comparisons difficult.

Many relevant and available well data are not discussed in the TBR. These include numerous data from wells completed as part of the Yucca Mountain site characterization project (e.g., Ervin et al., 1994), as well as regional water level data (e.g., La Camera and Westenburg, 1994). Water levels measured in these wells can provide important constraints on conceptual ¹¹ and numerical models of the ground water flow system.

The TBR cites, without any supporting data, an estimate by Young (1972) of the interval of time required to dewater the aqui

fer at a particular pumping rate. Given that considerable uncertainties still exist regarding the conceptual model of the flow system (Fridrich et al. 1994), this estimate may not be accurate.

The summary of annual production rates and periodic water level measurements from La Camera and Westenburg (1994), which are shown in Figures 3.3-1 through 3.3-3 of the TBR, provides the most quantitative data on water availability in the vicinity of Yucca Mountain. The usefulness of this information would be enhanced by the inclusion of a map in the TBR indicating pumping and observation well locations and a cross section showing open intervals and production intervals of monitoring and pumping wells, respectively.

The final section of the discussion of preclosure water supply refers to the simulation results of Czarnecki (1991). These simulation results are presented without discussion of the conceptual model of the ground water system or of the data used to estimate model parameters. Although they are not presented in the TBR, basic data for conceptual model development and model calibration are discussed in publications describing the original numerical model developed by Waddell (1982) as well as in Czarnecki and Waddell (1984). These publications note that although there is a paucity of water level data north of the proposed repository, available data in this area indicate an anomalously steep gradient. Czarnecki and Waddell (1984) also note that there is a general lack of transmissivity data throughout the modeled area.

Use of the modeling results of Czarnecki (1991) to assess the adequacy of available water requires that the pumping rates simulated [up to 7 × 10⁸ gallons per year (2.6 × 10⁹ liters per year)] be related to those anticipated for repository construction and operation. No estimates of repository requirements are provided in the TBR. Table 3.1 of this report summarizes estimates of water requirements derived from values cited in other documents including the final environmental assessment (DOE, 1986), site characterization plan (DOE, 1988), the early site suitability evaluation (Younker et al., 1992), and a memorandum prepared by the Department of Energy (DOE, 1995) in response to committee questions. The wide range of these estimates suggests considerable uncertainty in the water requirements for the repository. In addition to estimates of repository water requirements, assessment of water availability may also require analysis of concurrent water demands (e.g., for irrigation) in adjacent areas that are part of the same regional aquifer system.

The first interpretation provided in the TBR is the cited estimate of 76 to 380 years (Young, 1972) to dewater the aquifer at a pumping rate of 1,000 gallons per minute (3,800 liters per minute) or >5 × 10⁸ gallons per year (> 1.9 × 10⁹ liters per year). Potential problems with this estimate resulting from limitations of the model assumptions are noted in the previous section. The TBR itself questions this interpretation, noting that observed

recovery of water levels in a pumping well following decreases in pumping rate is inconsistent with Young's assumption of no recharge.

The TBR also concludes that the short-term effects of pumping on the regional potentiometric surface are negligible and that the tuff aquifer is experiencing regional recharge. This interpretation is consistent with reported recoveries of water levels during periods of reduced pumping and with the limited drawdowns observed during historic pumping. In discussing the water level monitoring data summarized in Figures 3.3-1 through 3.3-3, the TBR notes that these data suggest that no permanent drawdowns have been caused by ground water withdrawals to date in Jackass Flats. However, without additional information on the relative locations of pumping and observation wells, and without more extensive data on regional pumping rates and other ground water fluxes, it is difficult for the committee to evaluate this conclusion.

The TBR also cites a conclusion from Younker et al. (1992) that preclosure hydrology issues (presumably those related to water supply) can be accommodated by RAT. As in the case of the estimate of dewatering rates, no data or analyses to support this conclusion are included in the TBR. The water supply discussion in Younker et al. (1992) refers to an earlier environmental assessment (DOE, 1986), which cites a production rate of about 1.3 × 10⁸ gallons per year (4.9 × 10⁸ liters per year) for well J-13 over the period 1962-1983. The source of this production rate estimate is not provided. Thordarson (1983), one of the references cited in DOE (1986), reports that J-13 was pumped nearly continuously (at an unspecified rate) for many years during which time decreases in the static water level were less than 1 m.

The final interpretations in this section are drawn from the numerical modeling results of Czarnecki (1991). Maximum water level declines of about 10 feet (3 m) were simulated for “an extreme case” of pumping at 1,390 gallons per minute (5,300 liters per minute) for 10 years. Whether this represents an extreme case depends on the water requirements for the repository and other activities in Jackass Flats, requirements that are not quantified in the TBR. As noted previously, the model simulation results are a function of numerous model assumptions and parameter estimates that are not discussed or evaluated in the TBR.

The interpretations that would be most sensitive to alternative hypotheses are those related to the numerical simulation results of Czarnecki (1991). These results are based on a particular conceptual model of the hydrogeologic system, described by Czarnecki (1985), which assumes that a steep gradient to the north of the site results from a zone of low permeability. Several other conceptual models have been proposed to explain the high-gradient area, but none of these are described in the TBR. They include a model in which a high-permeability zone acts as a drain to connect the tuff aquifer to the underlying carbonate aquifer (Fridrich et al., 1994) and another model in which the anomalously high water levels north of the proposed repository correspond to perched conditions (Ervin et al., 1994). The TBR does not discuss if and how acceptance of any of the alternate conceptual models might affect predictions of drawdowns and sustainability of ground water supplies during repository construction and operation.

The TBR also does not consider the effects of other water demands in the region (e.g., for irrigation) on the availability of ground water for the repository. Extensive pumping in adjacent basins could have significant effects on water levels in the vicinity of Yucca Mountain, particularly if these rates represent a significant portion of the regional discharge. For example, La Camera and Westenburg (1994) estimate that production rates in the Amargosa Desert during 1985-1992 were 20 to 60 times higher than those in Jackass Flats. The maximum estimated production rates cited by La Camera and Westenburg (1994) in the Amargosa Desert [more than 3 × 10⁹ gallons per year (1 × 10¹⁰ liters per year)] are very similar to the estimate of total steadystate discharge in Franklin Lake Playa used in the model by Czarnecki and Wadell (1984), which constitutes more than 60 percent of the estimated steady-state discharge in the region.

If the particular values of drawdown predicted by Czarnecki (1991) are critical to regulatory compliance, DOE will need to refine its understanding of the apparent high-gradient zone through field testing. Appropriate tests would include drilling holes in the high-gradient area for evaluation of lithologic and hydrologic properties. To test the perched water hypothesis, water level measurements and hydraulic tests (pumping, slug, injection, tracer) in multiple holes completed to different depths, or a series of packer tests in a single hole, could yield useful information on vertical gradients and lateral continuity of possible perched intervals. To test the drain model, multiple wells extending into the carbonate aquifer below the tuff would be required.

Conceptual and numerical models of the ground water system would benefit from additional measurements of water levels and aquifer properties throughout the region. New estimates of transmissivity could be obtained from analysis of tests in recently completed boreholes in Midway Valley (Geldon, 1993). Additional data that would be useful to an assessment of competing future demands for ground water would be production histories and projected future pumping rates of wells in nearby areas such as the Amargosa Desert.

Alternatively, if bounding estimates of drawdown are all that are required for regulatory purposes, it may be acceptable to perform a sensitivity study by constructing numerical models consistent with the alternative hypotheses for the high-gradient zone and comparing drawdowns predicted by these models to those of Czarnecki (1991). A simplified analytical approach, evaluating whether drawdowns under alternative conceptual models will be greater than those of Czarnecki (1991), could also suffice.

Drilling of additional holes and field testing are likely to yield the most useful information on the hydrologic system that could help resolve questions related to the nature of the high-gradient area to the north of the proposed repository. For the purposes of bounding the availability of ground water supplies for construction, however, this may not be necessary, as discussed below.

The water supply section of the TBR presents interpretations related to ground water availability, with emphasis on historic pumping rates and on drawdowns predicted by using a nu-

merical model. It provides few data to support these interpretations and does not consider alternative conceptual models of the ground water system that may affect model predictions. The most serious deficiency of the discussion of preclosure water supply issues in the TBR is the lack of a clear statement of the technical questions that must be answered to establish the sufficiency of the water supply for repository construction and operation. The technical questions could be framed in terms of absolute availability of water or in terms of a bounding calculation. If DOE chooses to attempt to quantify the absolute availability of ground water in the Jackass Flats area, testing and comparison of alternative models would be required. The lack of data and constraints for ground water flow models throughout the basin, but particularly in the area of apparent high gradient in the water table north of Yucca Mountain, would be significant limitations to answering this technical question.

It is likely that the regulatory requirements could be satisfied by answering a technical question framed in terms of bounding calculations. A three-step approach that could be taken to address a technical question of this form is outlined below:

1. Determine a reasonable upper-bound estimate of water requirements for repository construction and operation, including peak annual usage and duration. The technical question to be addressed is whether existing water supply wells in the tuff aquifer can supply water at this rate for the specified duration, when other concurrent demands are also considered.

2. Compare historic production rates and the observed drawdowns in existing wells and nearby monitoring wells to the required pumping rates for the bounding estimate of water supply needed for the repository. This comparison may indicate directly

that existing wells have produced water, without excessive drawdowns, at similar rates and for similar durations to those required for the repository. If this is the case, the technical question may be answered without numerical modeling of the system.

2. If the comparison of historic pumping rates to an upper-bound estimate of water requirements does not yield a clear answer to the technical question, numerical models could be employed. If such models are used to address the technical question of water availability, a discussion of model assumptions and calibration data must be included. The possible changes in predicted water level declines that would result from incorporating alternative conceptual models into the simulations would have to be assessed in this case. Especially important to this analysis would be a comparison of results generated by numerical models incorporating the alternative conceptual models that have been proposed to account for the high-gradient area north of the proposed repositoy.