Precipitation variability in South America from IPCC-AR4 models. Part II: Influence of SH circulation leading patterns

31. VERA, C. S.; SILVESTRI, G. E.; LIEBMANN, B.; GONZALEZ, P.

Foz do Iguacu, Brasil

Conferencia; 8th AMS International Conference on Southern Hemisphere Meteorology and Oceanography; 2006

American Meteorological Society

The ability of the climate model simulations performed for the IPCC Fourth Assessment Report (AR4) in reproducing the leading patterns of atmospheric circulation in the Southern Hemisphere (SH) and their influence on precipitation variability in South America are discussed here, using the same subset of the climate simulations of the 20th century (20c3m) described in Part I. CMAP precipitation dataset and NCEP-NCAR reanalyses were used to describe the observed patterns. The leading patterns of circulation in the SH were identified through an analysis of the Empirical Orthogonal Functions (EOFs) for the 500-hPa geopotential height anomalies over the SH, southward of 20°S. Regression and correlation maps were computed for observations and model simulations. For those models that an ensemble of runs are available, the maps were first computed per individual run and then averaged over all runs available for each model. Results show that models are able to reproduce some of the features of the leading modes of SH circulation interannual variability (particularly those associated with the 1st leading pattern known as the Antarctic Oscillation, AAO). Although the simulated anomalies exhibit different amplitude and are somewhat misplaced than those observed. Furthermore, it was found that the ability of the models in representing the 2nd and 3rd SH leading modes (known as PSA1 and PSA2 respectively) is affected by the way that models reproduce ENSO features and the mean circulation along the SH subpolar regions. Observations show that the AAO is negatively correlated with precipitation anomalies over southeastern South America (SESA); PSA1 is related with the typical ENSO-induced precipitation anomaly pattern with negative anomalies in the tropical region and positive ones in SESA, while PSA2 correlates with positive precipitation anomalies further south. Results show, however that models have serious deficiencies to reproduce the observed influence of the three leading patterns of SH circulation onto precipitation variability in South America. Preliminary results show that UKMO, GFDL and MPI are the models that better depict the main features of the SH circulation anomalies associated with precipitation variability in SESA. Nevertheless, it will also be tested the ability of the multi-model ensemble those climate features.