Application of the Bivariate Frequency Ratio Method for Landslide Susceptibility Mapping of Manthali Municipality, Ramechhap District, Nepal

This study intends to create a landslide susceptibility map for Manthali Municipality in Ramechhap District, Nepal, utilizing the frequency ratio method. The landslide inventory map was developed using Google Earth and Landsat satellite imagery, with a 70% allocation for training data and 30% for testing data. The inventory map has been finalized subsequent to the field verification process. The assessment of landslide distribution encompassed the evaluation of ten variables, including slope, aspect, elevation, geology, proximity to roads and streams, curvature, land cover, topographical wetness index and precipitation patterns. The definitive weight for each factor was obtained from the raster analysis tool in ArcGIS. The resultant map was classified into four distinct susceptibility categories: low, moderate, high, and very high. The result was validated using the area under the curve method. The results demonstrate that the classifications of low, moderate, high, and very high accounted for 40.8%, 38.64%, 17.64%, and 2.92% of the total area, respectively. The area under the receiver operating characteristic curve (AUC) for the success rate was calculated at 78.94%, whereas the prediction rate was evaluated at 74.06%. The results indicate a significant level of predictive accuracy. The designated study area can utilize the generated landslide susceptibility map for effective landslide mitigation, strategic land use planning, and efficient settlement management.