Extreme cold events over southern South America: comparison between the most extreme CMIP3 and CMIP5 climate scenarios using the GFDL model

MULLER G. V.; REPINALDO C. R.; ARANEO D.; ANDRADE K. M.

Montevideo

Conferencia; WCRP Conference for Latin America and the Caribbean: Developing, linking and applying climate knowledge; 2014

WCRP

In order to study the atmospheric circulation associated to extreme cold events in the present climate and future scenarios for three regions of southeast South America (Paraguay/Brazil, Uruguay and Wet Pampa of Argentina), we analyze the ability of GFDL model for simulating the observational period. This analysis is performed with daily data from the NCEP/NCAR reanalysis 1 of the period May to September, and GFDL outputs for CMIP3 (GFDL-CM2.0) and CMIP5 (GFDL-CM3). We analyze the reanalysis and the model outputs for the period 1961-1990, representing the present climate, and the period 2081-2100 representing the future scenarios A2 (CMIP3) and RCP 8.5 (CMIP5). The analyzed fields are: sea level pressure and anomaly, temperature and anomaly and wind (direction and speed) at 850 hPa and 250 hPa, for the five most extreme events with temperature below 0°C at 850 hPa. In the CMIP3 experiment the model is capable of simulating the temperature and its anomaly over Uruguay, although it overestimates the incursion of the 0°C isotherm, intensifying negative temperature anomalies over Paraguay/Brazil and the Wet Pampa. The model simulates the post-frontal anticyclone, although narrower in comparison to reanalysis, intensifying the anomalous interaction over the continent. It also simulates well the wind direction and magnitude in Paraguay/Brazil and Uruguay, intensifying the magnitude over Uruguay. The model represents well position of the subtropical jet over the three areas, although over Paraguay/Brazil it is at lower latitudes. For the future scenarios, the model projects shorter incursions of the 0°C isotherm and weaker negative temperature anomalies at 850 hPa. The model also intensifies the post frontal anticyclone, suggesting that future cold air incursions would have less latitudinal range, although they may be more persistent. For CMIP5, the model is able to simulate well the temperature and anomaly fields in comparison with the reanalysis, although it overestimates the intensity of the temperature anomaly over Paraguay/Brazil and the Wet Pampa. The model composites of SLP and anomaly overestimate both the post-frontal anticyclone and the anticyclonic anomaly over Paraguay/Brazil and Uruguay, representing them well over the Wet Pampa. For the future scenario, the model projects both the incursion of the 0°C isotherm and the negative temperature anomalies similar to the climatological simulation. The model also projects longitudinally elongated post-frontal anticyclones, and anomalous interactions that would be closer to the continent. This suggests that future cold air incursions will not be very different from the present ones, although they would be more continental.