Incorporating preferential flow structures into hydrogeologic models

Accounting for preferential flow behavior is a major challenge when developing subsurface models.  Highly conductive geologic structures (e.g., faults, conduits, channel deposits) strongly influence subsurface dynamics by allowing for the rapid transport of fluids and solutes.  Most field data sets contain limited information on the locations of structures; thus, spatial modeling methods are needed to incorporate these features into hydrogeologic models.  This talk will describe the use of pattern-based geostatistics to model structural heterogeneity in alluvial aquifer systems.  Case studies are presented for two high-profile sites: the Lawrence Livermore National Laboratory site in northern California and the Columbus Air Force Base site in Mississippi.  At both sites, preferential groundwater flow and transport occur within coarse-grained deposits at the base of paleochannels.  The pattern-based simulation approach allows for a realistic representation of channel geometry and connectivity.  Specific channel locations are identified by conditioning the geostatistical simulations to static geologic observations (core sampling) and dynamic hydraulic test data (aquifer and tracer tests).