CMIP6 Model Evaluation for Mean and Extreme Precipitation Over India
Prabha Kushwaha,Vivek Kumar Pandey,Prashant Kumar,Divya Sardana
DOI: https://doi.org/10.1007/s00024-023-03409-5
IF: 2.641
2024-01-12
Pure and Applied Geophysics
Abstract:Extreme precipitation is typical in the Indian subcontinent, and these occurrences might cause human fatalities, property damage, and the environment. Understanding regional extreme precipitation in the global Coupled Model Intercomparison Project Phase 6 (CMIP6) models is challenging. The present study evaluates the performance of 20 CMIP6 models for daily precipitation across India from 1980 to 2014 (35 years) during the Indian Summer Monsoon (ISM) season (June-July–August-September (JJAS)) and ranked 20 models based on four metrics. In this analysis, the extreme precipitation over India is determined by using Generalized Extreme Value (GEV) distribution. The observation data is collected from the Indian Meteorological Department (IMD) over India during JJAS. The performance of CMIP6 models is determined by using four different model evaluation metrics as root mean square error (RMSE), standard deviation (SD), correlation coefficient (CC), and interannual variability score (IVS). A total rank is estimated based on the four skill metrics and the resulting top ten models, i.e., AWI-ESM-1–1-LR, BCC-ESM1, IPSL-CM6A-LR, MPI-ESM1-1–2-HAM, EC-Earth3-Veg, IITM-ESM, INM-CM4-8, GISS-E2-1-G, MIROC6, and NESM3 are found for mean precipitation. In extreme precipitation, the ten best-performing models are given as MIROC6, EC-Earth3-CC, CMCC-CM2-SR5, EC-Earth3-Veg, CMCC-CM2-HR4, BCC-ESM1, FGOALS-g3, INM-CM5-0, CanESM5, and INM-CM4-8. However, a strong uncertainty in CMIP6 models has been observed for extreme precipitation as compared to mean patterns over India. Overall, CMIP6 models perform well for mean precipitation and have a strong bias for extreme precipitation obtained from GEV over India during the ISM season.
geochemistry & geophysics