Cordilleran Section - 97th Annual Meeting, and Pacific Section, American Association of Petroleum Geologists (April 9-11, 2001)

Paper No. 0
Presentation Time: 2:30 PM

3-D GEOSPATIAL MODELING OF A DNAPL SOURCE AREA


MADRID, V. M., HALDEN, R. U., DEMIR, Z., VALETT, J. E. and GREGORY, S. D., Environmental Restoration Division, Lawrence Livermore National Lab, 7000 East Ave, Livermore, CA 94550, madrid2@llnl.gov

We have developed a three-dimensional (3-D) geospatial model of a vadose zone source area contaminated with dense, non-aqueous phase liquid (DNAPL) located approximately 300 feet above a water-supply aquifer. The contamination results from multiple releases of volatile organic compounds (VOCs), primarily DNAPL trichloroethylene (TCE), during the 1960s and 1970s. A mixture of TCE (100-30%) and silicone oil (0-70%) was used as a heat exchange fluid in weapons testing at Lawrence Livermore National Laboratory (LLNL) Site 300, a test facility located in the Altamont Hills of central California. Leakage of the heat-exchange fluid from pipelines resulted in contamination of approximately 100,000 cubic meters of soil and sediment at concentrations up to 970 mg/kg. Perched ground water beneath the site contains dissolved TCE at concentrations approaching the solubility limit. Temporal changes in VOC concentrations indicate that intrinsic bioremediation of the TCE occurs as a result of silicone oils, whose fermentation yields hydrogen required for the microbial dechlorination of VOCs. The 3-D geologic model is based on integration of multiple data sets including borehole samples, geophysical logs, and surface outcrops. The integrated data sets are used to characterize the 3-D geometry of high conductivity fluvial channels encased in finer grained overbank deposits. The fine-grained, clay-rich sediments serve as a perching horizon for shallow ground water and may cause pooling of DNAPL. The 3-D-contaminant model is based on more than 300 soil samples collected in 28 boreholes with a maximum depth of 90 feet. Additional control for the contaminant model is provided by passive soil vapor data collected using GORE-Sorbers. Inclusion of the GORE-Sorber data helped minimize extrapolation in order to more accurately define the outer limits of soil contamination. Spatial and temporal modeling of VOCs in ground water is based on two years of monitoring data collected monthly for characterization of VOC concentrations, Eh, pH, dissolved oxygen, and other biogeochemical indicators. This model is used to optimize site cleanup in this complex hydrogeologic setting using a combination of ground water extraction, soil vapor extraction, and intrinsic bioremediation.