Abstract
The Nilgiris district in Tamil Nadu, India, ranks as one of the most susceptible to landslides in the Western Ghats Mountain ranges due to the rugged terrain, fragile geology and high monsoonal rainfall. The terrain in the study area presents considerable elevation ranges between 91 m and 2634 m above mean sea level. The slope gradients attain as high as 89.8°. This study attempts to evaluate the susceptibility to landslides through the combined approach of the Analytical Hierarchy Process (AHP) and Frequency Ratio techniques. Various factors inSluencing landslides were integrated into the model. These include elevation, slope, aspect, curvature, Hill shade, geology, geomorphology, lithology, lineament density and their distances from roads, rivers and lineaments. Land use/land cover information derived from Landsat-8 satellite images and annual rainfall between 2370 mm and 2850 mm were also used as parameters. According to the AHP model results, the most inSluencing factor was slope with a weight of approximately 0.36, followed by geology with a weight of approximately 0.28. Rainfall and lineament density also showed considerable inSluence. Five susceptibility zones were identiSied as very low, low, moderate, high and very high. Approximately 27- 29% of the district area falls within the high to very high susceptibility zones, particularly around Udhagamandalam and Coonoor. Gudalur shows relatively low susceptibility. The model results were validated through the Receiver Operating Characteristic (ROC) curve. The Area under the curve (AUC) values for the AHP model were found to be 0.81, while the FR model results showed a high accuracy with a value of 0.87.
Keywords: Analytical Hierarchy Process, Geospatial analysis, Hazard mapping, Landslides Susceptibility, Nilgiris district, ROC-AUC validation