Authors: D Ramakrishnan T N Singh A K Verma Akshay Gulati K C Tiwari
Publish Date: 2012/09/09
Volume: 65, Issue: 1, Pages: 315-330
Abstract
This paper mainly presents a case study of landslide vulnerability zonation along TawaghatMangti route corridor in Kumaon Himalaya India An attempt is made to predict landslide susceptibility using backpropagation neural network BPNN and propose a suitable model for that zone which can be successfully implemented for the prevention of slides Various landslide affecting parameters such as lithology slope aspect structure geotechnical properties land use landslide inventory and distance from recorded epicenter are used to model the landslide susceptibility The database on the above parameters derived from satellite imageries topographic maps and field work are integrated in the GIS to generate an information layer Database of this information layer is used to train test and validate the BPNN model A threelayered BPNN with an input layer two hidden layers and one output layer is found to be optimal The developed model demonstrates a promising result and the prediction accuracy has been found to be 80 in the field
Keywords: