Estimation of mixed soil water content by impedance inversion of GPR data

In the vadose zone, soil has become an object of research due to its importance for environmental issues. Description and estimation of the mixed soil water content or dielectric parameter is the essential condition and the key to improving soil investigation with GPR detection. In this paper, first of all, a way to describe 3D random media which the preferred orientation of the multi-scale inhomogeneity is proposed and the importance of reducing the numerical errors with tapering function is stated. Then, we apply the FDTD method to simulate the GPR signal response of random model and use S-transform to test the simulation accuracy. For the complex random soil media, conventional method likes transmission wave method provide model parameter estimation of limited resolution only. Here, we apply a novel reflection amplitude inversion workflow for GPR data which is capable of resolving the subsurface dielectric permittivity and related water content distribution with markedly improved resolution. The synthetic results demonstrate that this method has extensive applicability in complex mixed random soil media detection and physics parameters estimation.