Performance of Qualitative and Quantitative Models in Creating Landslide Susceptibility Map of Chaliyar River Basin

Abstract

A Landslide Susceptibility Map (LSM) for Kerala's Chaliyar river basin is what this study aims to provide. Several landslides occurred in this basin as a result of the floods in 2018 and 2019. There were 592 identified landslides. Using ArcGIS 10.7 software, the landslide inventory was taken from the inventories created by the National Remote Sensing Center (NRSC) and Kerala State Disaster Management Authority (KSDMA), and the future incidence of landslides was projected by linking the landslide cause variables. Landslide inventories were split into training and validation groups in this study, with the ratios fixed at 70:30. Two models, including a qualitative one called Weighted Linear Combination (WLC) and a quantitative one (a bivariate statistical model) called Weights of Evidence (WOE) model, were used to evaluate landslide susceptibility. The following factors were employed as causative parameters: Slope, Aspect, Curvature, Relative Relief, TWI, Distance to Road, Distance to Streams, Distance to Lineaments, Land Use Land Cover, Drainage Density, Road Density, Lineament Density, Geomorphology, Soil Texture, and NDVI. The performance of the models was evaluated using the Receiver Operating Characteristic (ROC)'s Area Under the Curve (AUC). In the study, the WLC model yields an AUC success rate accuracy of 59.9%, while the WOE model yields an accuracy of 70.9%. In terms of ratio of landslide validation accuracy, the WOE model outperforms the WLC model by 11%. The anticipated landslide area is included in the landslide susceptibility map, which can be incorporated to lessen the risk of landslides in this research.