Resumo:
Climate change is the greatest challenge due to threats in the natural and social systems. For adaptation actions planning to be effective, it is necessary to analyze the impacts on climate variables, especially in potentially vulnerable regions, such as South America. Additionally, as a new generation of general circulation models (GCMs) becomes available, it is mandatory to reevaluate the performance and reliability of this new set in representing the behavior of the climate system. Therefore, the present study analyzed and compared simulations and projections of climate change over South America using ten GCMs from Coupled Model Intercomparison Project, using fifth and sixth phases (CMIP5 and CMIP6). The projections (PF; 2081-2100) were compared to the present period (PP; 1986-2005), for four radiative forcing scenarios: RCP4.5 and RCP8.5 (Representative Concentration Pathways) and SSP2-4.5 and SSP5-8.5 (Shared Socioeconomic Pathways). Furthermore, this study analyzed space-time trends and the contribution rate of climate variables that influence evapotranspiration in South America, considering the PP and the SSP5-8.5 future scenario of CMIP6. For this, a set of daily data of mean air temperature at 2 m above the surface (Tmean), relative air humidity (RH), global solar radiation (Rs), soil moisture at 10 cm from the surface (SM) and precipitation (P) was used of ten GCMs from both sets of CMIP. To obtain the daily evapotranspiration (ET) the Turc method was used. In addition, the aforementioned variables were validated with data from the Climate Prediction Center (CPC) and ERA5 reanalysis. The CMIP6 GCMs exhibited better performance compared to the CMIP5 GCMs in representing the analyzed climate variables over South America, especially for Tmean and ET. The GCMs projected a positive trend in ET, which was accompanied by an integrated effect of a positive trend in Tmean and Rs, and a negative trend in RH, SM and P, which indicated drier climate conditions for the end of the 21st century mainly in the Amazon region, north of Bolivia and southern Peru. Despite the spatial distribution of the contribution rate having differences in regional patterns, the dominant climate variable that contributed to the positive trend of ET was Tmean, mainly for the future period, indicating that this contribution was not weakened by the contribution of other climate variables. In general, these results help in evaluating the variables that influence the change in ET in South America.
