Landslide susceptibility mapping using analytical hierarchy process (AHP) in Dabong, Kelantan, Malaysia

Abstract

Identifying the potential landslide threat in the respective area is critical for sustainable development. Landslide is one of the main hazardous events in Kelantan annually, especially during the monsoon season. This study aims to analyse the landslide probability in the study area by generating a landslide susceptibility map. The map was produced by using ArcGIS software. Two different sources of the satellite image, Sentinel-2 and ALOS PALSAR, were used to produce ten thematic maps. The ten thematic maps were slope, soil moisture, lithology, precipitation, distance to roads, distance to drainages, normalised difference vegetation index (NDVI), land use land cover (LULC), lineaments, and aspect. Each parameter was assigned with a different weightage. The weightage was determined by using the AHP method, referring to the previous study. Each thematic layer was then overlayed in ArcGIS to produce a landslide susceptibility map. The susceptibility zone was divided into five classes, the very low zone, low zone, moderate zone, high zone, and very high zone. The highest percentage for landslide susceptibility is low (23.12%), followed by high, moderate, very high and very low, with respective percentages of 22.24%, 20.03%, 17.31% and 17.30%. The accuracy of the model was evaluated using the area under the curve (AUC). The AUC analysis has indicated the value of 0.684, equalling 68.4%, which means that the landslide susceptibility map's accuracy is 68.4%.