Identifying discrete geologic structures that produce anomalous hydraulic response: An inverse modeling approach

[1] Subsurface hydraulic response is controlled by discrete geologic structures in a variety of aquifer settings. In this work, we analyze data from the well-studied alluvial fan system that underlies the Lawrence Livermore National Laboratory. There, subsurface flow and transport behavior is strongly influenced by discrete channel deposits, which are embedded within less permeable floodplain deposits. We model the distribution of these deposits using a multiple-point geostatistical approach that enables simulation of geologically plausible connected permeable bodies consistent with a training image. By coupling the geostatistical model with a dynamic flow model in a simulation inversion framework, we identify specific channel structures that are consistent with transient aquifer test data at six observation wells. We perform shortest-path analysis to characterize the geometry of the simulated channels and to locate critical high-permeability conduits that control the system's response to pumping. In one area, all successful inverse solutions identify the same set of stacked channel deposits, a geologic feature that provides an explanation for anomalous response behavior observed during the aquifer test.