| Popis: |
Landslide susceptibility mapping provides the probability of sliding of the landmass based on the intrinsic characteristics of the area. Various models were applied to get such pixel wise probability as a function of weightage assigned to individual dependent variable influencing the event. This study was conducted on a part of Darjeeling Himalaya within West Bengal, India, where, effort was made to run a comparative analysis among the output susceptible zones generated by few well known models like AHP, Information value method and logistic regression. For the purpose, elevation, slope, aspect, geology, land use, lineament density, some topographic indices were taken as factors influencing the hazard. Digital elevation model, Landsat-8 Operational Land Imager satellite image, and other secondary data sources were taken to generate the thematic rasters for these dependent variables. By means of weightage calculation and integration, pixels with highest value imprinting the highest probability to slide were detected individually in each model used. Thus, the magnitude of applicability of individual method was assessed through validation with reference to the existing landslide occurrences. Such a comparative analysis can assure the reliability of such models and certainly help to manage this natural hazard with a better accuracy. |
