Characterization and Prediction of Subsurface Pneumatic Pressure Variations at Yucca Mountain, Nevada

Characterization and Prediction of Subsurface Pneumatic Pressure Variations at Yucca Mountain, Nevada C. Fredrik Ahlers, Stefan Finsterle, Gudmundur S. Bodvarsson Earth Sciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720 Yucca Mountain, Nevada is being investigated as the proposed site for geologic disposal of high level nuclear waste. A massive data collection effort for characterization of the unsaturated zone is being carried out at the site. The USGS is monitoring the subsurface pressure variations due to barometric pumping in several boreholes. Numerical models are used to simulate the observed subsurface pressure variations. Data inversion is used to characterize the unsaturated system and estimate the pneumatic diffusivity of important geologic features. Blind predictions of subsurface response and subsequent comparison to recorded data have built confidence in the models of Yucca Mountain. Introduction Barometric pumping of the subsurface was first described by Buckingham (1904), and he provided an analytical solution for gas pressure as a function of depth for the idealized conditions of a single homogeneous layer bounded above by a sinusoidally varying pressure boundary condition and bounded below by an impermeable boundary. Actual field measurements of subsurface response to barometric pumping have been used to estimate hydrologic properties by several researchers. Stallman (1967) and Stallman and Weeks (1969) describe estimation of permeability assuming a single layer bounded below by an impermeable boundary. Rozsa et al. (1975) describe estimation of diffusivity assuming a semi-infinite model. Weeks (1978) describes a numerical technique for estimation of permeability of a multilayered system and results from several field tests. Shan (1995) describes an analytical solution to the multilayered problem and its application to field data from one site. This paper will present the results of subsurface pneumatic pressure data inversion using ITOUGH2 (Finsterle, 1997a), an automated inversion program based on the T O U G H 2 (Pruess, 1991) multiphase numerical simulator. In situ measurements of surface and subsurface pneumatic pressure, gathered by the USGS, are used together with geologic based numerical models of Yucca Mountain to estimate the pneumatic diffusivity of several layers. Site Description Yucca Mountain, Nevada, located approximately 120 km northwest of Las Vegas, is the proposed location for a geologic repository designed to permanently store high level radioactive waste. Many surface based boreholes have been drilled in the Yucca Mountain area and several of these have been instrumented to monitor in situ air pressure. A 9 km long tunnel, the Exploratory Studies Facility, ESF, has been excavated to nearly 300 m below the surface in the unsaturated zone. The unsaturated zone at Yucca Mountain is up to 700 m thick. The volcanic rocks that make up the unsaturated zone are alternating layers of welded and nonwelded ash flow and air fall tuffs. Figure 1 shows a cross section with the hydrogeologic units of Montazer and Wilson (1984) labeled. These units are based mainly on degree of welding. From the surface down, these units are the Tiva Canyon welded (TCw), the Paintbrush nonwelded (PTn), and the Topopah Spring welded (TSw). The TCw and TSw are highly fractured and very permeable to air. The PTn is less fractured and much more porous than the overlying or underlying units making it an impediment to gas flow and thus an important layer to characterize. Sublayering of the PTn using the nomenclature of Buesch et al. (1995) is also shown on Figure 1. The Yucca Mountain and the Pah Canyon Tuffs, Tpy and Tpp, respectively, are partially welded while the other PTn sublayers are nonwelded, bedded tuffs. Very little pneumatic pressure data is available to characterize rocks below the TSw. Data Subsurface gas (pneumatic) pressure data have been collected at twelve boreholes on the Yucca Mountain site by several principal investigators and organizations. For this paper, data from two boreholes, NRG-6 and NRG-7a, are analyzed. Data from NRG-7a, shown in Figure 2, are characteristic of a typical data set. The pressure at the land surface shows the largest variation (and the lowest pressure because it is at the highest elevation). With increasing depth below the surface, the pressure signal shows increasing amplitude