Evaluation of extreme precipitation climate indices and their projected changes for Brazil: From CMIP3 to CMIP6

Extreme events usually cause numerous economic and social losses, especially in vulnerable countries, such as Brazil. Understanding whether the evolution of Earth System Models (ESMs) from Coupled Model Intercom-parison Project (CMIP) improves the representation of extreme events and investigating their future change is fundamental because device policies of adaptation and mitigation to climate change generally consider the re-sults of the most recent generation of ESMs. This study analyzes the performance of a subset of 40 ESMs from CMIP3, CMIP5, and CMIP6 in simulating eight extreme precipitation climate indices over Brazil during 1981-2000 and also estimates their projected changes for the middle (2046-2065) and far future (2081-2 100) under the worst-case scenario for each CMIP generation. Results reveal that CDD are the most challenging precipitation index to be simulated, while the best ones were PRCPTOT and R20mm. The model performance shows that CMIP3 has the best skill for Northeast Brazil, CMIP5 for Center-West, and CMIP6 for North, Southeast and South regions. Thus, at least for Brazil, the evolution of the ESMs from CMIP did not reflect a substantial improvement in the representation of precipitation climate extremes over all Brazilian regions. In addition, all the models across CMIP generations have difficulty in simulating the observed trends. This indicate that im-provements are still needed in CMIP models. Despite the relative low performance in the historical climate, the climate projections indicate a consensus signal among most of precipitation climate extremes and CMIP gen-erations, which increase its reliability. Overall, the extreme precipitation events are projected to be more severe, frequent, and long-lasting in all Brazilian regions, with the more pronounce changes expected in heavy rainfall and severe droughts in the central northern portion of Brazil and in the southern sector.