Water content dynamics at plot scale – comparison of time-lapse electrical resistivity tomography monitoring and pore pressure modelling

Abstract

Physically-based dynamic modelling of shallow landslide susceptibility rests on several assumptions and simplifications. However, the applicability of physically-based models is only rarely tested in the field at the appropriate scale. This paper presents results of a spray irrigation experiment conducted on a plot of 100m2 on an Alpine slope susceptible to shallow landsliding. Infiltrating precipitation applied at a constant rate (27.5mm/h for 5.3h) was monitored by means of 2D time-lapse electrical resistivity tomography, combined with time-domain reflectometry sensors installed at various depths. In addition, regolith characteristics were assessed by dynamic cone penetration tests using a light-weight cone penetrometer. The spray irrigation experiment resulted in a vertically progressing wetting front to a depth of 80–100cm. Below that, the unconsolidated material was already saturated by rainfall in the previous days. The observed mean resistivity reduction attributed to infiltrating water during irrigation was scaled to pressure head. Mean variations in pore pressure were reproduced by a linear diffusion model also used in physically-based dynamic landslide susceptibility modelling. Sensitive parameters (hydraulic conductivity and specific storage) were tested over selected value ranges and calibrated. Calibrated parameter values are within published and experimentally derived ranges. The results of the comparison of observations and model results suggest that the model is capable of reproducing mean changes of pore pressure at a suitable scale for physically-based modelling of shallow landslide susceptibility. However, small-scale variations in pore pressure that may facilitate the triggering of shallow landslides are not captured by the model.