Inferring aquifer storage parameters using satellite and in situ measurements: Estimation under uncertainty

We present a robust optimization method for estimating aquifer storage parameters (specific yield or storativity) using the Gravity Recovery and Climate Experiment (GRACE) data, in situ well level observations, and other ancillary information. Uncertainty inherent in the remotely sensed and in situ time series can adversely affect the parameter estimation process and, in the worse case, make the solution completely meaningless. Our estimation problem is formulated to directly minimize the negative impact of data uncertainty by incorporating bounds on data variations. We demonstrate our method for the interconnected Edwards-Trinity Plateau and Pecos Valley aquifers in central Texas. The study area is divided into multiple zones based on the geology and monitor well coverage. Our estimated aquifer storage parameters are consistent with previous results obtained from pumping tests and model calibration, demonstrating the potential of using GRACE data for validating regional groundwater model parameters.