Electrical Resistivity Tomography Monitoring of an Infiltration Test in an Agricultural Context within a Miscanthus Parcel

An artificial run-off hydrogeophysical experiment was conducted in cultivated fields (Gembloux (BE) to study the infiltration patterns of water in the transition between an empty beet field and a band where miscanthus has been planted. Such set-ups are designed to mitigate the flooding risk and the erosion from runoff during intense rainfall events. The objective of this experiment was to determine with geophysical methods whether miscanthus enhance water infiltration in addition to blocking mud and slow water flow. The experiment was repeated 3 times, each time next to each other with the same experimental setup: a 1-meter-wide and 6-meter-long long band isolated with plastic boards, 3m is uncovered and the other 3m is covered with miscanthus plant base and roots. The band has a slight inclination, and saline water was poured to create a surface run-off at a rate of 1L/s at the top of the band. A primary longitudinal profile composed of 16 electrodes (0.4m spacing) was used to monitor the infiltration, with measurements taken approximately every 2.5 minutes. Two perpendicular profiles (4.5m long, 0.3m spacing) in each section were used to do background measurements and after the experiment.  

A first analysis has been carried out on apparent resistivity to avoid any inversion bias. Each parcel shows a greater starting mean apparent resistivity in the miscanthus parcels. During the infiltration, apparent resistivities decrease more rapidly in the miscanthus parcel during the first minutes of the experiment and reach a lower value than in the bare parcel. Subsequently, resistivities in both parcels decrease at a slower rate but do not reach a steady state, even after 3 hours of infiltration. Once water injection ceases, resistivities quickly stabilize within a few minutes at a lower value than the starting value but higher than at the end of injection.  

Timelapse inversion revealed a decrease in resistivity in the top 40cm soil after only a few minutes following the start of the water injection. We estimate that the layer below this horizon corresponds to the plough layer, where the higher density and lower permeability of the soil beneath this level doesn’t allow the infiltration at this time scale. However, inversions of perpendicular profiles reveal lateral extension of the resistivity decrease in the parcels without miscanthus, a pattern which is absent in the miscanthus parcel. 

Further data processing will focus on the inversion problem and on the influence of the surface water height during the water injection and pedophysics experiments will allow us to estimate the water content. Along with other inversion parameters, this will help provide a better understanding of the dynamics of infiltration rates in the different parcels.