Hydrogeophysics is attracting increasing interest in the environmental sciences community as it provides a variety of non-intrusive techniques to study near surface processes. Among the available geophysical methods, I have mainly worked with geoelectrical methods, which include electrical resistivity tomography (ERT), spectral induced polarization (SIP), and self-potential (SP). When used together, ERT and SP methods are of interest in hydrology and environmental sciences because of their non-invasive nature and their sensitivities to soil structures and flow and transport processes in the subsurface. ERT estimates the subsurface electrical conductivity distribution, which is mainly driven by porosity, clay content, and pore water salinity. The SP signal, on the other hand, can be induced by different sources including water fluxes. The latter interesting contribution, the so-called streaming potential, results from the presence of an electrical double layer at the mineral-pore water interface. When water flows through the pore space, it gives rise to a streaming current and a resulting measurable electrical voltage between non-polarizable electrodes placed at different locations. Electrical geophysical methods are now well understood in water saturated porous media, but the modeling of streaming currents or electrical conductivity under partial saturation is still an area of active research.
In order to evaluate the ability of ERT and SP methods to characterize structures and flow and transport in the vadose zone, we conducted a field-based monitoring of the vertical processes using ERT and SP during different hydrologic events. The investigations were carried out at the Voulund agricultural test site that is part of the HOBE hydrological observatory, located in the Skjern river catchment (Denmark) in the middle of a cultivated area. The site has been instrumented since 2010 to monitor suction, water content and temperature down to a depth of 3 m, together with meteorological variables and repeated geophysical surveys (cross borehole electrical resistivity tomography and ground penetrating radar). In July 2011, we installed 15 non-polarizable electrodes at 10 depths within the vadose zone (from 0.25 to 3.10 m) and a reference electrode below the water table (7.30 m). More than 2 years of data acquired every 5 minutes are now available for various hydrologic events, such as natural infiltration, water table rises, and a high salinity tracer test. I have developed a fully coupled numerical scheme to simulate water fluxes, ionic transport, electrical conductivity and SP signal during different hydrological events. With this approach, I considered the most recent models to link electrical conductivity changes and SP generation to the relevant hydrological parameters. To compare simulated and measured geophysical results, I conducted a Markov-chain-Monte-Carlo (McMC) inversion of suction and water content data to obtain the medium van Genuchten parameters. The first results compare fairly well to the measured geophysical data (ERT and SP). These initial results will serve as a starting point for a detailed assessment of the value of ERT and SP data in vadose zone hydrology, particularly as a tool for in situ monitoring of water flux.