Estimating Flow Parameters Using Ground-penetrating Radar and Hydrological Data during Transient Flow in the Vadose Zone

Methods for determining the parameters necessary for modeling fluid flow and contaminant transport in the shallow subsurface are in great demand. Soil properties such as permeability, porosity, and water retention are typically estimated through the inversion of hydrological data (e.g., measurements of capillary pressure and water saturation). However, ill-posedness and non-uniqueness commonly arise in such inverse problems making their solutions elusive. Incorporating additional types of data, such as from geophysical methods, may greatly improve the success of inverse modeling. In particular, groundpenetrating radar (GPR) has proven sensitive to subsurface fluid flow processes. In the present work, an inverse technique is presented in which permeability distributions are generated conditional to timelapsed GPR measurements and hydrological data collected during a transient flow experiment. Specifically, a modified pilot point framework has been implemented in iTOUGH2 allowing for the generation of permeability distributions that preserve point measurements and spatial correlation patterns while reproducing geophysical and hydrological measurements. Through a numerical example, we examine the performance of this method and the benefit of including synthetic GPR data while inverting for fluid flow parameters in the vadose zone. Our hypothesis is that within the inversion framework that we describe, our ability to predict flow across control planes greatly improves with the use of both transient hydrological measurements and geophysical measurements (GPR-derived estimates of water saturation, in particular).