Assessment of South-West and North-East Monsoon over Indian continent using CMIP5 models
Abstract ID#: 33156
This study compares the present-day skill of 36 state-of-the-art coupled global climate models (GCM) taking part in Coupled Model Intercomparison Project 5 (CMIP5) in simulating NEM and SWM of India. Our analysis uses the simulations from historical runs of CMIP5 models during 1979-2005. The performance of 36 coupled models for the present-day climate is assessed using Taylor diagram, based on the consensus in simulating climatology and interannual variability of NEM and SWM rainfall. The Global Precipitation Climatology Project (GPCP) monthly precipitation dataset is used as reference data. Our analysis reveals large model spread in simulating the observed climate statistics for NEM and SWM rainfall. Considering this large model spread, the selection of best performing models are performed using the following criteria: (1) the pattern correlation in seasonal climatology is above 0.6 (2) the normalized standard deviation of seasonal climatology is between 0.75 and 1.25 (3) the pattern correlation in the interannual standard deviation is above 0.5 and (4) the normalized spatial standard deviation of interannual standard deviation is between 0.75 and 1.25. Out of 36 models, 5 models (BNU-ESM, IPSL-CM5A-MR, GFDL-ESM2G, CNRM-CM5, MRI-CGCM3) are qualified as optimally best performing models in simulating NEM. In contrast for SWM, none of the CMIP5 models satisfies the prescribed criteria, indicating that the models show better skill in simulating NEM compared to SWM. The process studies focusing on systematic errors at regional scale are performed to understand the differences in model skills. Our study has huge implications for useful regional projections as downscaling of the local precipitation is subjected to this inherent uncertainty.