GIS-BASED STATISTICAL ANALYSIS USING FIVE DIFFERENT METHODS FOR LANDSLIDE VULNERABILITY MAPPING AND COMPARISON OF THEIR PERFORMANCE IN KANGRA, HIMACHAL PRADESH, INDIA

Abstract

Landslides are among the most devastating elements on the globe, causing many deaths and financial losses every year. The investigation of variables causing the incidence of landslides in an area and the zoning of the consequent damages will undoubtedly play an important role in minimizing such events. Kangra, the district of Himachal Pradesh, India has experienced frequent landslides in the past decades making it important for the study. The landslide vulnerability zonation map has been formulated by applying 5 non - identical statistical models namely, frequency ratio (FR), Shannon’s entropy (SE), information value (IV), Weight of evidence (WoE), and certainty factor (CF) models. We have considered a total of 200 landslide points at different locations in Kangra for our research purpose. Landslide inventory is created by dividing the total landslide points into training data (80%) and testing data (20%). There can be a lot of triggering factors for landslides but we have taken 13 factors into consideration namely, elevation, slope, curvature, aspect, SPI, TWI, rainfall, distance to roads, rivers & lineaments, LULC, geology, and lithology. The vulnerability map is then created using the 5 models and a comparison is done between them. Then, the map created by the above-mentioned 5 models is validated with the help of a testing data set using the AUC (Area under curve) of the ROC (Receiver Operating Characteristics) curve. The AUC comes out to be 0.863, 0.844, 0.83, 0.88, and 0.88 for FR, SE, IV, WoE, and CF model respectively. These maps can be further utilized by the government in mitigation planning.