2005 Salt Lake City Annual Meeting (October 16–19, 2005)

Paper No. 16
Presentation Time: 8:00 AM-12:00 PM


RUST, G.L., Department of Geological Sciences, Michigan State University, 206 Natural Science Building, East Lansing, MI 48824-1115, WEISSMANN, Gary Stephen, Department of Geological Sciences, Michigan State University, 206 Natural Science Building, East Lansing, MI 448824-1115 and MIKESELL, Leslie R., Geological Sciences, Michigan State Univ, 206 Natural Science Building, East Lansing, MI 48824-1115, rustginn@msu.edu

Although accurate groundwater modeling is greatly dependent on the amount of hydrogeologic heterogeneities captured, optimal methods for modeling these heterogeneities are still under investigation. Geologically-based stratigraphic reconstruction of the depositional system, in combination with transition probability geostatistics, allows for characterization and modeling of facies distributions and enhances the accuracy of hydrogeologic flow models. Therefore, generating correlations of stratigraphic markers and determining regional interconnectedness is a critical step toward developing reasonable models.

Multiple correlation strategies were utilized to reconstruct depositional sequences in the alluvial fan deposits at the Lawrence Livermore National Laboratory (LLNL), California. LLNL resides in a transpressional basin, the Livermore valley, located approximately 45 miles east of San Francisco. The alluvial sediments originate from the northeast within the Altamont Hills. Visual description of approximately 1,737 meters of core from 54 boreholes along with geophysical well logs and driller's logs provide information on subsurface facies distributions. An additional 19 wells lack core but facies were interpreted from geophysical well logs. Preliminary research conducted north of the study site suggests a minimum range of correlation of approximately 15 meters for most facies. A more regional provenance study of gravel deposits was also utilized to differentiate between three large-scale fan sequences encompassing the total depth of the study region. Finally, modern alluvial fan analogs were employed to help explain observed inconsistencies of lateral bed distributions.

Four geologically-derived hydrofacies (gravel, sand, silty-sand, and paleosol) compose the stratigraphy of the site. Stratigraphic columns illustrate thicknesses and depth relationships of these markers between wells. Cross-sections of the stratigraphy are utilized to identify the maximum extent of correlation for this portion of the alluvial fan. The correlated stratigraphy is used to (1) guide Markov chain model development, (2) provide soft conditioning for realizations, and (3) guide selection of reasonable geostatistical realizations of aquifer heterogeneity.