Development and Assessment of GIS-Based Landslide Susceptibility Mapping Models Using ANN, Fuzzy-AHP, and MCDA in Darjeeling Himalayas, West Bengal, India

Landslides, a natural hazard, can endanger human lives and gravely affect the environment. A landslide susceptibility map is required for managing, planning, and mitigating landslides to reduce damage. Various approaches are used to map landslide susceptibility, with varying degrees of efficacy depending on the methodology utilized in the research. An analytical hierarchy process (AHP), a fuzzy-AHP, and an artificial neural network (ANN) are utilized in the current study to construct maps of landslide susceptibility for a part of Darjeeling and Kurseong in West Bengal, India. On a landslide inventory map, 114 landslide sites were randomly split into training and testing with a 70:30 ratio. Slope, aspect, profile curvature, drainage density, lineament density, geomorphology, soil texture, land use and land cover, lithology, and rainfall were used as model inputs. The area under the curve (AUC) was used to examine the models. When tested for validation, the ANN prediction model performed best, with an AUC of 88.1%. AUC values for fuzzy-AHP and AHP are 86.1% and 85.4%, respectively. According to the statistics, the northeast and eastern portions of the study area are the most vulnerable. This map might help development in the area by preventing human and economic losses.

Automated summary

Landslide susceptibility maps are important for managing, planning, and mitigating landslides. In this study, an analytical hierarchy process (AHP), a fuzzy-AHP, and an artificial neural network (ANN) were used to construct landslide susceptibility maps for a region in West Bengal, India. The ANN model performed the best in terms of accuracy.