Estimate of hydrodynamic parameters with a coupled hydrogeophysical inversion using GPR surveys

We develop a methodology to estimate soil hydrodynamic parameters from a water infiltration experiment monitored with a GPR (Ground Penetrating Radar). Such an experiment, carried out on both controlled and natural site, consists in applying a water charge in a tank on the soil surface. During the water infiltration, the water layer thickness above the soil surface in the tank and the GPR response on the infiltration water front are monitored. The infiltration experiment is then modelled numerically using hydrogeological parameters which describe the constitutive relationships between water content, pressure and hydraulic conductivity. In that goal, we use the WAMOS-1D code which combines the 1D Richards equation and the Mualem – van Genuchten model. From the hydrogeological models outputs and petrophysical relationships, corresponding GPR velocity models are created to generate the resulting GPR signals. Then, an inversion algorithm couples both the hydrogeological and the geophysical models to seek the optimal hydrodynamic parameters. The inverse problem objective function is calculated from the estimated arrival time of the GPR waves reflected by the water infiltration front and compared to the measured ones. Preliminary inversion tests explore the hydrodynamic parameters space using synthetic data. First results show that the saturated hydraulic conductivity parameter can be estimated. Further tests are performed to improve both our experimental set-up and methodology and allow an estimation of the other hydrodynamic parameters. An emerging idea is to complete the objective function by analyzing the arrival time corresponding to additional reflectors to the water infiltration front.