How well are RCMs representing extreme indices in the SAM region?

Regional climate models (RCMs) serve as a crucial tool for understanding climate and its application in vulnerability, impacts, and adaptation. In South America (SAM), there has been significant growth in the study of RCMs, mainly from a climatological perspective. Nevertheless, their capacity for representing extremes needs to be addressed. In the CORDEX SAM region, we analyze the behavior of three RCMs driven by different Global Climate Models: RegCM4-7, REMO2015, and Eta. The main purpose of this research is to contrast RCM representativeness in the SAM region for different AR6 sub-regions in four climate extreme indices, CDD and RX5day for precipitation and TXx and TNn for temperature. First, we evaluated observational uncertainty between surface observations and two observational products: CPC and AgERA5. We compared RCMs spatial and interannual variability in the historical period (1981-2005) and trends in a second step. Our results show that most RCMs have several limitations in representing the interannual variability of extremes indices and their trends in almost all AR6 regions. Not surprisingly, the most deficient performance of RCMs in representing extremes are in regions with scarce surface observations in the center and north of South America. Overall, the trends present several differences between the two observational products all over SAM, especially in the signal of change that diverges between RCMs and the different ensembles.