Regional climate projections of daily extreme temperatures in Argentina applying statistical downscaling to CMIP5 and CMIP6 models

Abstract

Abstract Argentina is a wide country with a variety of climates, where an increase in mean and extreme temperatures is currently on-going, demanding regional climate information to design and implement effective strategies for climate change adaptation. In this regard, the use of Empirical Statistical Downscaling (ESD) procedures can help providing tailored climate information. In this work, a set of ESD models were tested to generate plausible regional climate projections for daily maximum and minimum temperatures (Tx, Tn) in Argentina. ESD models were applied to an ensemble of CMIP5 and CMIP6 global circulation models (GCMs) to downscale historical and future worst-case scenarios. The plausibility of the ESD projections was analysed by comparing them with their driving GCMs and with CORDEX regional climate models (RCMs). Generally, all ESD models added value during the historical period, in mean values as well as in extreme indices, especially for Tx. The climate projections depicted an extended signal of warming (both in the mean and in the frequency of extremes), consistent between all simulations (GCMs, RCMs and ESD) and strongest over northern Argentina. ESD models showed potential to produce plausible projections, although, depending on the technique considered (for Tx) and the predictor configurations (for Tn), differences in the change rates were identified. Nevertheless, the uncertainty in future changes was considerably reduced by RCMs and ESD when compared to their driving GCMs. Overall, this study evidences the potential of ESD in a climate change context and contributes to the assessment of the uncertainty on the future Argentine climate.