Downscaling of climatic data using statistical techniques in coastal saurashtra, Gujarat

Abstract

Downscaling of Global Circulation Models (GCMs) outputs is important to assess regional climate change impacts. Coarse resolution of GCMs presents averaged weather conditions. GCMs outputs cannot be used directly and are usually downscaled to transfer GCM-derived climate outputs into station-based values. Precipitation and temperature are two important climate variables that are used as inputs in hydrologic models in many water resources studies. In this study, a statistical downscaling model (SDSM) is utilized for generating daily future climatic outputs in a semi-arid coastal river basin in the Saurashtra region of Gujarat. Predictors from Hadley Centre Coupled Model version 3 (HadCM3) experiment are used for simulation of climatic data under two IPCC SRES scenarios A2 and B2. The model is calibrated and validated using daily station precipitation and temperature data and gridded data obtained from the National Centre for Environmental Prediction (NCEP) re-analysis data set. Model accuracy is evaluated through standard indices. The results show that the downscaling model is capable of reproducing the daily temperature series at local scale in the study region with much higher degree as compared to precipitation.