Monitoring drywell infiltration dynamics using time-lapse electrical resistivity tomography

Abstract

Drywell infiltration initiates as water is injected into the drywell. Subsequently, the water level in the drywell builds up the driving head of water flow into the subsurface via the surface area of the drywell. Drywell infiltration is a function of the surrounding media’s hydraulic properties, the drywell’s geometry, and the injection rate. The drywell infiltration capacity property determines the water volume that can infiltrate the subsurface for different injection rates without overspilling. This property can be evaluated under controlled infiltration experiments where water levels in the drywell are continuously monitored during injection. However, no available method exists for revealing spatiotemporal information on the subsurface flow mechanisms, including flow patterns and residual time at the vadose zone. Conventional methods for monitoring the subsurface are intrusive, expensive, and can provide limited information, especially on the spatial extent. Unlike conventional monitoring techniques, electrical resistivity tomography (ERT) can provide continuous, non-invasive information on the subsurface in an easy-to-apply and efficient manner. We examine the ERT applicability to monitor water dynamics at the deep vadose zone (at depths of 20–40 m), induced by drywell infiltration. For that purpose, electrodes were installed at the surface, in the perforated section of the drywell, and at a wetwell, located 5 m from the dry one. Time-lapse ERT surveys were conducted during a controlled drywell infiltration experiment, including the borehole and surface electrodes. The results show that the relative changes in the electrical conductivity can describe water dynamics during the infiltration experiment. Saturation maps translated from the electrical tomograms using calibrated petrophysical relations preserved the water mass balance injected into the well up to $\sim$250 min after the injection started. Modeling this experiment with a semi-analytical solution, assuming a sharp-wetting front interface, agreed with the wetting front location from the time-lapse electrical tomograms and with the water levels measured in the drywell.