A novel approach for the computation of soil slope stability in distributed hydrologic applications

Abstract

The identification of potential source areas for landslides at the basin scale is a fundamental step for risk assessment in mountain regions. In the literature, many contributions deal with this task by using the Infinite Slope (IS) assumption. However, there is evidence that the widely used IS assumption becomes increasingly inadequate with growing space resolution of the elevation model, contributing to the overprediction of unstable areas that affect this model. In this paper, we propose an original improvement to the stability calculation by presenting a modification of Janbu's method of slices, specifically devised for basin-scale applications. The method is coupled with a hydrological model to provide transient soil saturation during a rainfall event. We show that the proposed improvement is systematically better than the local IS prediction for a set of elementary test cases where a rigorous stability analysis is available, confirming the predictive capabilities of the proposed approach. Finally, we show the applicability of the model at the watershed scale by considering its application to a well-known test case in the literature, i.e. the Mettman Ridge site, Oregon.