Full-wave inversion of near-field ground penetrating radar data for hydrogeophysical characterization of soil

Accurate quantification of soil moisture is crucial for hydrology, meteorology, and environment. Recently, GPR has become a popular technique to characterize soil moisture. Although there have been intensive studies on applications of GPR for soil moisture estimation, significant efforts are still needed to improve its accuracy. In addition, in many applications, knowledge of the soil moisture profile is essential, which has remained a major challenge for many years. For increasing the accuracy of soil moisture estimation, we improved the GPR modeling and petrophysical relationships. For the GPR model, we further developed the near-field antenna model introduced by Lambot and Andre (2013). We successfully calibrated and validated the model using both numerical and laboratory experiments. Then, we improved the petrophysical relationship by coupling full-wave GPR inversion, dielectric mixing model and Debye's equation to account for the frequency dependence of electric properties and directly estimate soil moisture. For estimating the soil hydraulic properties and moisture profile, the GPR model was integrated into the hydrodynamic model in an assimilation framework. The hydrodynamic model was employed to simulate the spatiotemporal dynamics of soil moisture in the unsaturated zone, while the GPR model and petrophysical relationships were used to link the soil moisture profile with GPR data. The assimilation was performed using the Maximum Likelihood Ensemble Filter algorithm. Instead of using the surface soil moisture only, the approach allows us to use the information of the whole soil moisture profile for the assimilation. This thesis opens a new development and application avenue for digital soil mapping and soil water resources monitoring.