Abstact

STATISTICAL DOWNSCALING OF PRECIPITAION DATA OVER YAMUNA RIVER BASIN

Global Circulation Models (GCMs) are a major tool for projection of how climate will change under different emission scenarios. The GCM model gives the output with coarse resolution. Downscaling is a technique for obtaining fine resolution data from coarser resolution data of GCM. Downscaling involves two approaches (i) dynamic and (ii) statistical downscaling and in this study we aim to investigate and evaluate promising statistical downscaling technique. In this study we use Artificial Neural Network (ANN), Support Vector Machine (SVM) and Genetic Programing (GP) for statistical downscaling and compare the results generated from the above methods at six different locations of Yamuna river basin. The model is calibrated and validated using the National Centres for Environmental Predication (NCEP) reanalysis data and future projection is done using the predictor of CanCM4 GCM model produced by Canadian centre for climate change for AR5 scenarios.