Deriving hydrological parameters from VRP data: accounting for uncertainties in inverted velocities and petrophysical models

Abstract

Today, vertical radar profiling (VRP) is mainly used to derive 1D velocity models in the vicinity of a borehole by inverting direct arrival traveltimes. In hydrological applications, the resulting velocity models are often used to estimate hydrological material properties such as soil water content or porosity. However, uncertainties in the inverted velocity and in the employed petrophysical models are typically ignored. We present a workflow to appraise uncertainty and nonuniqueness issues inherent to VRP traveltime inversion and to quantify the influence of these issues on the following petrophysical translation. Our strategy relies on an efficient global inversion approach, which results in an ensemble of velocity models explaining the data equally well. For estimating water content and porosity, we use the entire ensemble of velocity models which results in an ensemble of possible petrophysical parameter distributions. In addition, this Monte-Carlo procedure allows us to investigate the impact of uncertainties in the employed petrophysical model by using a variety of appropriate translations. Using synthetic VRP data and a field example, we demonstrate the applicability and the potential of the proposed workflow.