Evaluation of CORDEX Multi-RCM for Indian Subcontinent Using NASA’s RCMES

Abstract

Mean, minimum, and maximum surface air temperature, precipitation and cloudiness are evaluated for their strengths and limitation in the period (1991−2001) from the CORDEX regional climate model hindcast experiment. South Asia region has limited CORDEX data availability, 3 RCMs namely RCA4 (SMHI), REMO (MPI), and REGCM4 (ICT) were available for assessing their climate feature simulation capabilities. Regional climate model evaluation system (RCMES) is a powerful tool for characterizing strength, weakness, and uncertainty involved in the model. In this study, Indian region has been divided into 17 subregions in such a way that they represent spatial climatic variability of the country. These regions are assessed individually as well as collectively with CORDEX data which is available at a resolution of 0.5° for their biases and correlation against the gridded surface station data of climate research unit (CRU TS v. 4.01). Comparative study of simulated precipitation against CRU dataset revealed that systematic model biases do occur across all the models whereas basic climatological features are simulated reasonably well except for few regions including Himalayas as well as regions of Gujarat and Arunanchal Pradesh. The study also revealed that surface air temperature is simulated better in comparison to precipitation and cloudiness, multi-model ensemble mean outperforms individual models. Taylor diagrams are used to enhance studies effectiveness by incorporating correlation coefficient, RMSE and standard deviation among various models.