Development of a regional landslide prediction model for Kerio Valley in Kenya

Abstract

Occurrences of rainfall induced landslides coincide in time with the period of high persistence of heavy wet spells and in spatial locations where vegetative cover has been denuded. A landslide prediction model was formulated and calibrated by simulation of previous landslide occurrences and then applied in scenario analyses to establish the effects of rainfall persistence and land use changes on slope stability in the western escarpment of the Kerio Valley in Kenya. Long term daily rainfall input was obtained using a rainfall generator. HYDRUS 1D model was applied in soil moisture analysis. An EXCEL spreadsheet infinite slope model was developed for model calibration. ILWIS GIS model was used to map spatial distribution of slope stability. The model successfully modelled the occurrence of previous landslide events. Results indicate that rainfall totals and wet spell persistence contribute to increase in unstable areas. Assessment of slope stability under different land uses showed that if the entire study area were completely forested, then the unstable and critical areas would be 0.05% and 1.81% respectively, but would be 0.69% and 5.35% respectively if it were completely under agriculture.