Evaluation of Rainfall Variability Over Major River Basins of India Using Multi Temporal Satellite Data Products

Abstract

High resolution remotely sensed rainfall data is indispensable to accurately monitor rainfall variability at a river basin scale under climate change. The aim of the study is to assess the change in rainfall climatology and determine the performance of three GPM rainfall products (IMERG Final_Run, GSMaP_Gauge and recently developed GSMaP_ISRO) against in situ observation over major Indian River basins for the monsoon season rainfall during 2000–2020. The analysis provides valuable insights into the issues in the rainfall products and their performance under varying rainfall intensity (low, moderate and heavy) and orography. Results indicate that mean monsoon rainfall is better represented in GSMaP_ISRO than IMERG Final_Run and GSMaP_Gauge estimates. GSMaP_ISRO outperformed IMERG Final_Run and GSMaP_Gauge with smaller root-mean-square error and higher correlation coefficient. It was observed that the performance of GPM rainfall is influenced by rainfall intensity and terrain height of basins. In particular, for high rainfall occurrence, the Brahmaputra and Barak basins in northeast India exhibited large uncertainties in IMERG Final_Run and GSMaP_Gauge products. The statistical evaluation of different rainfall scores (POD, FAR and CSI) suggested that the GSMaP_ISRO rainfall has significant skill over Indian river basins. A significant improvement is observed over Brahmaputra (RMSD = 34.46, CC = 0.92, FAR = 0.26), Barak and others (RMSD = 65.75, CC = 0.95, FAR = 0.29) and Cauvery basin (RMSD = 23.1, CC = 0.8, FAR = 0.15) for GSMaP_ISRO estimates. These findings provide valuable information on the accuracy of GPM rainfall necessary for hydro-meteorological applications.