Constraining GPR data inversion using hydrodynamic laws for noninvasive soil hydraulic and electric property determination

Abstract

We constrain full-wave inversion of time-lapse radar data using hydrodynamic modeling to simultaneously identify the shallow subsurface hydraulic properties and continuous vertical electric profiles. Radar data are acquired in the frequency domain using a vector network analyzer combined with an off-ground monostatic antenna. This permits to accurately filter antenna effects and to derive Green functions from which the inversion is initiated. In order to demonstrate that enough information is contained in the radar data so as to ensure unique estimates, hydrodynamic events were simulated for three different textured soils, namely, coarse, medium, and fine. The corresponding time-lapse radar data were subsequently computed and inverted to find back key soil hydraulic parameters, i.e., α, n, and Ks in Mualem-van Genuchten's model. For the three scenarios considered, the three hydraulic parameters were exactly retrieved, and hence, the corresponding time-dependent electric profiles as well. Provided that the soil-specific relations between the soil water content and its electric properties and the hydrodynamic initial and boundary conditions are known, the proposed method appears to be promising for proximal mapping of the shallow subsurface hydraulic properties and monitoring of the water dynamics at the field scale.