Vulnerability evaluation utilizing AHP and an ensemble model in a few landslide-prone areas of the Western Ghats, India

Landslides are among the most perilous hazards that usually happen in hilly terrains. The loss that ensues during a landslide, especially in highly-populated regions, calls for a vulnerability study. Thus, the purpose of this research is to detect landslide-vulnerable villages in a small part of the Western Ghats, an orographic mountain chain in South India that is proverbially prone to landslides. The study also evaluates the prediction capabilities of analytical hierarchy process (AHP) and ensemble fuzzy-AHP (F-AHP) models. 22 vulnerability indicators (11 physical-environmental and 11 socio-economic) served as the basis for this modeling. These data, derived both from field studies and remotely sensed satellite data, were collated in a geographic information system (GIS) environment, and landslide vulnerability maps were generated. Landslide vulnerability modeling using AHP and F-AHP models found 12.07% and 4.53% of the region, respectively, as very high-vulnerable. The developed landslide vulnerability maps are validated using the receiver operating characteristic (ROC) curve, sensitivity, specificity, Kappa index, mean squared error (MSE), and root mean squared error (RMSE) techniques. Based on the area under the ROC curve (AUC) scores, the landslide vulnerability maps developed utilizing these models were found to be outstanding. With an AUC score of 96.55% (0.96), the ensemble (F-AHP) was found to be more competent than the AHP model, which had an AUC value of 95.14% (0.95). The sensitivity, specificity, Kappa index, MSE, and RMSE values for the F-AHP model are 95.14%, 93.61%, 94.35%, 0.091, and 0.211, respectively, and for the AHP model, they are 92.93%, 94.04%, 92.45%, 0.099, and 0.228. Hence, in this study area, it can be affirmed that F-AHP is the better model for distinguishing vulnerable zones. As per the F-AHP model, Vellarada village is very highly vulnerable, and villages, namely Keezharoor and the western part of Peringamala, Vithura, and Mannoorkara, are highly vulnerable to landslides.