Estimating electric permittivity from GPR surface reflection data for water content estimates

Abstract

Estimating surface water content from GPR data requires first an accurate estimate of the electric permittivity that can be attributed to a small depth range below the surface. Only when this is achieved does it make to sense to try and relate the permittivity to water content, which will then be surface water content. For small and large scale investigations, suitable methods are available. For the intermediate field or watershed scale, GPR is the most researched method. We briefly go through the most promising method available now using GPR surface reflection data calibrated by metallic plate reflections. We summarize the assumption underlying the common surface reflection model and extend the model by relaxing some of the assumptions, without the need of performing full waveform inversion. We find that the plane wave reflection coefficient approximation holds quite well over a wide frequency range, while keeping the antenna at sufficient height above the surface that is already necessary due to other assumptions. We find that extending the model to allow for frequency dependence also improves the permittivity estimate quite strongly in the high-frequency regime.