Developing an ANN Model in R-Studio for Identifying Landslide Triggering Parameters with the Help of GIS and Remote Sensing Data

The Himalayan mountain ranges are very vulnerable to regular geohazards such as earthquakes, landslides, cloudbursts, flash floods, and more due to their neo tectonically dynamic nature. The rock mass of the Himalayan slopes is severely broken jointed, and sheared. This has an impact on the rocks’ strength, which in turn has a significant impact on the stability feature of slopes that have been made worse by human activity. Incidents involving landslides cause grave financial losses, deaths, and permanent harm to the surrounding environment. This study uses the ANN methodology to cover the landslide inventory mapping method. In this case, we have used nine different landslide-affecting factors: slope, aspect, curvature, NDVI, lithology, soil types, precipitation, LULC, and distance from drainage. The landslide susceptible maps (LSMs) were produced by calculating the relationship of landslide-influencing elements and historical landslide placements using the artificial neural network (ANN) approach. Weight of each element is calculated using this model, through this we can conclude that which parameter plays important role in triggering landslide. The LSMs were examined and validated using data from the most recent landslide inventories. This model has an accuracy of 77.7%. Precipitation has the highest weightage, which means that it plays the important role in landslide occurrence other than that slope and NDVI are also important parameters to consider in landslide susceptibility mapping. The great precision attained with the ANN model indicates a reliable standard for the zonation of landslide susceptibility in the future. To prevent landslides and to reduce disaster risk, this map may be helpful for development planning and management, which will aid in regional development planning and assist lower the chance of tragedy.