CHANGES IN SUMMER PRECIPITACION IN SOUTHEASTERN SOUTH AMERICA BY THE WCRP/CMIP5 MODELS

DÍAZ LEANDRO; VERA CAROLINA

Montevideo

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

Unidad de Ciencias de la Atmósfera, Facultad de Ciencias, Universidad de la República

The climate changes observed in the last decades has raised concern among policy and decision makers about the importance of the knowledge and prediction of climate. In particular southeastern South America (SESA) has experimented large positive precipitation trends during the 20th century that impacted many socio-economic sectors.There is a concern if this trend is due to the climate anthropogenic forcing and whether it will continue in the future or not. The fifth phase or the World Climate Research Program-Coupled Model Intercomparison Project (WCRP-CMIP5) Experiment currently provides the most comprehensive scenarios for studies of present and future climates. However, an exhaustive evaluation of the results given by these models in representing the present climate is needed in order to assess their abilities to estimate the main and relevant features of the regional climate and its variability. The goal is then to assess the skill of the WCRP/CMIP5 climate simulations in representing the mean, variability and trends of theaustral summer precipitation in South America during the 20th century. In addition, an exploration of the precipitation trends in southeastern South America (SESA) and its possible causes is also made.Historical simulations of 33 WCRP/CMIP5 coupled general circulation models (CGCM) are considered in order to make a preliminary evaluation of the representation of the austral summer (December-January-February, DJF) rainfall mean and variability in South America. Simulated climatological mean rainfall computed over the last part of the 20 th century shows for most of the WCRP/CMIP5 models similar biases, when compared to observations, than those shown by the previous version of the models (WCRP/CMIP3). Model errors are still largest over northern coast of South America and in less extent over SESA. Nevertheless, an improvement of the representation of the interannual variability is found and in particular of its first leading pattern of precipitation (EOF1). The pattern, identified from an Empirical Orthogonal Function analysis, is characterized by a dipole-like structure with centers of action of opposite sign located over SESA and the South Atlantic Convergence Zone (SACZ), which is simulated by all models considered.DJF rainfall trends were computed over the 1902-2005 period. Most of the models show a significant positive trend over SESA in agreement with observations. However the simulated trend is much weaker. Moreover, the trend seems to be at least partially associated with a frequency increase of EOF1 positive events, which are characterized by precipitation increase (decrease) in SESA (SACZ). Precipitation trend from the multi-model ensemble mean is significantly different from inter-model dispersion over central-eastern Argentina. However, uncertainties related to internal variability (represented by model member dispersion) are large. Preliminary studies reveal that the anthropogenic forcing (associated to greenhouse gas concentration increase) seems to at least partially explain the positive rainfall trend.