Landslide Susceptibility Mapping Using Frequency Ratio-Based Fuzzy Logic Approach in Lima Puluh Kota Regency

Abstract

One of the non-linear models for determining landslide-prone areas is the fuzzy logic method. This research aims to determine the landslide prone zoning map in Limapuluh Kota Regency, by using fuzzy gamma operations in GIS, from 11 factors: aspect, curvature, elevation, slope, landform, geology, proximity to stream, landcover, rainfall intensity, NDVI and soil. The data used is 149 landslide events in Limapuluh Kota Regency. The data was processed using the factors that cause landslides with a train and test ratio of 60: 40. The research results show that by using γ = 0.975, an area with high landslide characteristics was obtained covering an area of 28,463 ha (8.79%), with a medium area of 76,544 ha (23.64%) and low area of 218,820 (65.57%). By validating using the ROC curve and test data, an AUC value of 0.746 was obtained, which means the accuracy level of the resulting map is high.