Three-dimensional Geostatistical Inversion with Multiple Joint Data

Estimating the hydraulic conductivity of an aquifer is an important task of groundwater inverse modeling. Typically, the number of available measurement locations is limited. Using different types of measurements helps improving the estimate because different types of measurements have different sensitivity pattern. In this study, we consider the most frequently used types of measurements such as direct conductivity measurements, head measurements obtained in hydraulic tomography, thermal signals, and geoelectrical potentials, and use them for geostatistical inversion. In the literature, there is a lack of concrete comparisons of estimates based on different kinds of joint measurements. It is uncertain which type of measurements are well suitable for which kind of different structures of aquifers. We want to fill this gap using a series of synthetic scenario tests to illustrate the differences of the estimates with combinations of different types of measurements. This allows us to determine what information is carried in each kind of data. Furthermore, it tells which combination of joint data is well suitable for a significantly improved estimate. In our artificial test scenarios, we are easily able to quantify the goodness of the estimate by comparing it to the generated "true" conductivity field.