Performance of CMIP6 models over South America

Assessing the performance of global climate models in the present is essential to attribute confidence to their future projections. This work aims to evaluate the ability of 28 'Sixth Phase of the Coupled Models Intercomparison Project' (CMIP6) models to represent the South American (SA) climate during the reference period (1995-2014). Using the top 7 selected models that best represent the SA climate (CMIP6-SA), it will be possible to obtain better simulations than CMIP6. The made assessment compares the seasonal austral summer and winter climatologies simulated by the CMIP6 models to the Global Precipitation Climatology Project (GPCP) dataset and the European Center for Medium-Range Weather Forecasts Reanalysis 5 (ERA5) Reanalysis. Also, the performance of the 28 models was objectively evaluated through Taylor Diagrams, using their precipitation monthly time series (1980-2014) and temperature monthly time series (1950-2014). The Top7-CMIP6-SA precipitation annual cycle is also evaluated concerning GPCP and the ensemble of all 28 models used in this study. The results show that, at lower levels, most models offer a good performance to represent the South Atlantic and South Pacific Subtropical Anticyclones, the Intertropical Convergence Zone (ICTZ), and the South Atlantic Convergence Zone, except for AWI-ESM-1-1-LR, BCC-ESM1, and IITM-ESM, which do not provide a good representation of these systems. At upper levels, most models overestimate the magnitude of the Subtropical and Polar Jets. Most models can adequately represent the position of the Bolivian High and the Northeast Brazilian Trough, except AWI-ESM-1-1-LR, CAS-ESM2-0, CNRM-ESM2-1, FGOALS-f3-L, GISS-E2-1-G, IITM-ESM, INM-CM5-0, IPSL-CM6A-LR-INCA, MPI-ESM-1-2-HAM e NESM3. Regarding the annual precipitation cycle, the ensemble of the Top7-CMIP6-SA presents a better performance at some AR6 regions compared with the 28 models' ensemble, especially over the north of SA.