Model-based automatic dense velocity analysis of GPR field data for the estimation of soil properties

Causse (2004 J. Geophys. Eng.1 28–45) and Causse and Hokstad (2005 US Patent no 6839658 B2) have proposed a model-based approach to build accurate travel time approximations that take into account the vertical velocity heterogeneities in the earth. This new approach can improve the estimation of kinematic parameters from reflection data, e.g. the stacking velocity. We use these new equations to perform an automatic velocity analysis of reflection data, and apply the method to two ground-penetrating radar data sets acquired at the same agricultural site at two different dates. We obtain profiles of radar velocities, which are used to estimate the water content and the porosity in the soil. These estimates show a good correlation with independent measurements of the water content and the porosity.