Advancing landslide early warning systems through saturation monitoring and prediction

Abstract

Landslides most commonly occur in the rainy season, resulting in damage to infrastructure and human lives. An early prediction framework for landslides would clearly help to mitigate damage. In this work, a prediction framework for shallow landslide initiation was developed and validated with a real case study. To assess the reliability of the prediction framework, back-analysis of a landslide that occurred in Malin village, Maharashtra, India on July 2014 was performed. Relations of landslide stability with soil saturation were established through a physically based approach using GeoStudio software. A leaky barrel algorithm was developed for the study location to monitor the effect of rainfall' on soil saturation. Simulation results of landslide stability were compared with the rainfall–soil saturation algorithm based on the leaky barrel. The presented framework was found to have good predictability of shallow landslide occurrence and is therefore recommended for real-time monitoring of landslide-prone locations.