Evaluating the aptitude of global climate models from CMIP5 and CMIP6 in capturing the historical observations of monsoon rainfall over Sudan from 1946 to 2005

Sudan is one of the most vulnerable countries to climate change in the African Sahel region. Given the urgent need to understand the potential behaviour of the future climate, this study assessed the skill of 32 global climate models (GCMs) from the fifth (CMIP5) and the sixth (CMIP6) phases of the Coupled Model Intercomparison Project (CMIP) in simulating summer monsoon precipitation over Sudan for the period 1946–2005. The models were ranked based on the Skills Score, estimated using three statistical metrics, correlation coefficient, index of agreement, and normalized root mean square error. The best performing models and their ensembles were used for long-term trend quantification, empirical orthogonal function (EOF) analysis, and assessment of precipitation associations with atmospheric circulation. GISS-E2-H, IPSL-CM5A-MR, and MPI-ESM-LR of CMIP5 and BCC-CSM2-MR, INM-CM4-8, and MPI-ESM1-2-LR of CMIP6 showed the best performance in Sudan. Instead of August's unimodal peak rainfall, the GCMs, especially the CMIP5 GCMs, showed an unrealistic rainfall peak in May and a shift of the August rainfall peak to September. Only the CMIP6 multi-model ensemble reliably reproduced the unimodal rainfall property. All models of both phases failed to reconstruct the dominant mode of monsoon precipitation as determined using EOF analysis, particularly in the country's southern parts. The models also failed to reproduce the long-term rainfall trends in different climate zones. Only MPI-ESM-LR (CMIP5) and MPI-ESM1-2-LR (CMIP6) were successful in simulating monsoon rainfall associations with sea surface temperature, El Niño-Southern Oscillation teleconnection, and winds components at 850 hPa