An intercomparison of model-simulated in extreme rainfall and temperature events during the last half of the twentieth century. Part 1: mean values and variability

MATILDE RUSTICUCCI · JOSÉ MARENGO · OLGA PENALBA · MADELEINE RENOM

CLIMATIC CHANGE

SPRINGER

Año: 2010 vol. 98 p. 493 - 508

0165-0009

In this study we examine the performance of eight of the IPCC AR4global coupled climate models used in the WCRP CMIP3 Multimodel Dataset, aswell as their ensemble mean, in simulating annual indices of extreme temperatureand precipitation climate events in South America. In this first part we focus oncomparing observed and modeled mean values and interannual variability. Twoextreme temperature indices based on minimum temperature (warm nights andfrost days) and three indices of extreme precipitation (R95t, R10 and consecutivedry days), obtained both from meteorological stations during 1961?2000 and modeloutputs, were compared. The number of warm nights are better represented bymodels than the FD. The interannual variability pattern is also in good agreementwith the observed values. For precipitation, the index that is best represented bythe models is the R95t, which relates the extreme precipitation to local climate. Themaximum of dryness observed over the central Argentinian Andes or the extensivedry season of the Amazon region could not be represented by any model.