On the reliability and uncertainties of a regional climate simulation.

PESSACG NATALIA; SOLMAN SILVINA ALICIA

Buenos Aires, Argentina

Workshop; CLARIS-LPB Kick-off Meeting; 2008

CLARIS-LPB / CIMA/ UBA

Regional climate models (RCMs) represent at present an important tool for studying the behavior of regional climate in a large range of temporal and spatial scales. RCMs allow simulating the regional climate and performing climate change projections. Besides, RCMs can be used for studying climate processes, in particular it is possible to explore the regional climate response to different local or remote forcings. However, RCMs have different uncertainty sources mainly associated with the interbal variability of the model, the model configuration, physics and initial and boundary conditions. With the aim of decreasing the uncertainties related to the model deficiencies and in order to find the optinmal configuration of the RCM to represent the climatic conditions over South America, we performed a series of sensitivity experiments. We evaluated the best choice of several parameterizations, mainly concerned with convection a boundary layer physics. The experiment were performed for climatic conditions associated with extreme events, in terms of anomalous precipitation over La Plata basin region. We found that the MRF Planetary Boundary Layer scheme performs better than any other PBL scheme.  The results performed using the Kain-Fritsh convection shceme and the Grell convection shceme showed that ovetrall South America, the Grell scheme reproduces better the extreme conditions over the La Plata basin region and over the South Atlantic Concvergence zone region, though it strongly underestimates precipitation over tropical regions.