Ground penetrating radar forward and backward modeling for layered snow

Abstract

Forward and Backward Ground Penetrating Radar (GPR) modeling can provide the information that is useful for geometric and electromagnetic interpretation of layered materials. Given the depth and dielectric constant of snow, one can simulate the GPR signal. This is called forward modeling. On the other hand, given the GPR signal, the determination of depth and dielectric constant is called backward modeling. In this thesis, we briefly discuss a computational GPR forward modeling approach, based on two pieces of software with different principles. The software GPRMAX is able to model snow layers with the Finite Difference Time Domain method; this method is able to distinguish and resolve two consecutive layers with less thickness ([delta]d = 10cm). The software MATHCAD can be utilized to obtain modeling and imaging result using Transmission Line theory; this method is able to distinguish and resolve two consecutive layers with relative small difference in dielectric constant ([lamda epsilon] = 0.01). Then the comparisons of GPR system outputs and modeling techniques give an evaluation of the working performance of the Ground Penetrating Radar system. Secondly, for backward modeling, we used the layer stripping method on the field data as well as forward modeling results. Furthermore, the relative permittivity can be evaluated by the inverse process of transmission line theory with transmission line forward modeling results.