Landslides are one of the major geo hazards responsible for the huge loss of resources worldwide. Since time immemorial, Nilgiris, a hilly district of Tamil Nadu, has been repeatedly ravaged by landslides. With an aim to develop landslide early warning systems for Nilgiris, the paper develops different models to assess landslide occurrence risk based on daily rainfall forecasts and rainfall thresholds. The paper employs Artificial Neural Networks to predict one day advance rainfall intensity and then assesses the risk of landslide occurrence by comparing it with rainfall thresholds. The data set comprises of daily recorded rainfall intensities at 14 rain gauge stations located in and around Coonoor. The results obtained and sensitivity analysis performed establishes the efficiency and adequacy of rainfall data as a supplement to different meteorological parameters and suitability of artificial neural networks in forecasting rainfall and hence evaluating the risk of landslide occurrence.
|Journal||Data powered by Typeset2014 International Conference on Advances in Computing, Communications and Informatics (ICACCI)|
|Publisher||Data powered by TypesetIEEE|