The accurate assessment of groundwater and its management requires obtaining reliable estimates of hydraulic conductivity (K) and specific storage (Ss). A large number of empirical, laboratory and field techniques have been developed over the last several decades. However, research suggests that Hydraulic Tomography (HT) yields the most accurate hydraulic parameter estimates that can then be used to build robust groundwater flow models. The majority of algorithms used for HT analysis has relied on geostatistics, however, a number of studies have shown that smooth K and Ss estimates are obtained when the inversion begins with homogeneous hydraulic parameter estimates and when data densities are not high. These smooth estimates are not visually appealing from a geological standpoint. One could overcome this by integrating geological data that are typically available through outcrops and borehole logs.
Here, we examine the usefulness of geological data for HT analysis in unconsolidated deposits by: (1) comparing “traditionally” calibrated geological models to highly parameterized geostatistical inverse models and (2) using geological models as prior information for the geostatistical inversion approach. The investigation has been conducted with laboratory sandbox experiments, at a small-scale field site on the University of Waterloo campus consisting of highly heterogeneous glaciofluvial deposits and using data obtained from a municipal well field.
Results reveal that the calibration of groundwater models built primarily with geological data, yields mixed results in terms of model performance, perhaps reflecting the uncertainties in geological structures along the vertical direction and between boreholes. The geostatistical inversion approach without the explicit reliance of geological data yields improved model performance over traditional geological models when data densities are high, although the resulting K and Ss distributions may not be geologically appealing. On the other hand, when geological data are fused with the geostatistical inversion approach, the resulting K and Ss estimates are more visually appealing from a geological standpoint, and that model performance is most robust. Overall, our results suggest the joint use of both geological and pumping test data for HT analysis when accurate geological data are available.