Geostatistical modeling of pore velocity

A significant part of evaluating a geologic formation as a hazardous waste repository involves the modeling of contaminant transport in the surrounding media in the event the repository is breached. The commonly used contaminant transport models are deterministic. However, the spatial variability of hydrologic field parameters introduces uncertainties into contaminant transport predictions. This paper discusses the application of geostatistical techniques to the modeling of spatially varying hydrologic field parameters required as input to contaminant transport analyses. Kriging estimation techniques were applied to Hanford Reservation field data to calculate media conductivity and the ground-water potential gradients. These quantities were statistically combined to estimate the ground-water pore velocity and to characterize the pore velocity estimation error. Combining geostatistical modeling techniques with product error propagation techniques results in an effective stochastic characterization of ground-water pore velocity, a hydrologic parameter required for contaminant transport analyses