How well CMIP6 models simulate key boundary conditions affecting South American climate? Insights for regional modeling efforts

Abstract

South America is a large continent with a wide diversity of weather and climate features, including tropical, subtropicaland extratropical regimes interacting within a complex landscape. Simulations by global climate models, as well as theirdownscaling through regional circulation models or statistical methods, are important tools, particularly when assessing theimpacts of climate change in the continent. This work evaluates 57 models of the sixth phase of the Coupled Model IntercomparisonProject (CMIP6) in their simulation of various spatial patterns and circulation features over South America. Ourevaluation aims to provide useful input for the selection of climate models to force regional simulations in South America.Therefore, we focus on spatial fields that are relevant for regional simulations, such as horizontal winds, sea level pressure,sea surface temperature (SST), and moisture and energy fluxes across the domain boundaries. Additionally, we evaluateddifferent circulation features influencing the regional climate of South America that have not been widely evaluated in thesemodels. Several indices are studied to assess the main low-level and upper-level continental-scale circulation patterns, theregional Walker and Hadley cells, the subtropical highs and the boundary SST patterns. Our results show that no singlemodel performs best across all evaluated features, highlighting the importance of in-depth model evaluation for the regionconcerning the features of interest.