Estimating hydrologic parameters from water table dynamics using coupled hydrologic and ground-penetrating radar inversion

During water table drawdown the differential relationship between hydraulic conductivity and the distance above the water saturated zone (SZ) leads to delay in drainage and stretching of the transition zone. The ground-penetrating radar (GPR) reflection from the SZ depends both on the width and shape of the transition zone, and dynamic changes of the saturation curve alter the GPR reflection amplitude, frequency, and phase. We have implemented a method to invert directly for van Genuchten-Mualem parameters from time-lapse GPR data acquired during a pumping test. The forward model consists of simulating the dynamic water saturation response to drawdown, integrated with a model of GPR response based on the full analytic reflectivity solution. We use a gradient based, multi-parameter method to optimize for an effective GPR source function and three van Genuchten-Mualem parameters Ks, n, and α, where Ks is the saturated hydraulic conductivity and n and α are fit parameters that control the shape of the soil-moisture retention curve. For field data acquired at the Boise Hydrogeophysical Research Site, this approach provided a hydrologic parameter set that substantially improved the fit to in-situ soil moisture curves over an initial “best-guess” model.