Daily extreme rainfall and temperature events over South America as represented by four regional models and observations

TENCER, BÁRBARA; ROBLEDO, FEDERICO ARIEL; ARMELLE RECA REMEDIO; ANNA SÖRENSSON; DANIELA JACOB; LAURENT LI; CLAUDIO MENENDEZ; OLGA C. PENALBA; RUSTICUCCI, MATILDE; ENRIQUE SANCHEZ; PATRICK SAMUELSSON; HERVE LE TREUT

Foz do Iguazu

Conferencia; The Meeting of the Americas; 2010

American Geophysical Union

The impact of climate variability on the environment and the economic activities mainly depends on changes in the frequency of occurrence of extreme events. Relatively small changes in the mean can be related to substantial changes in the frequency of extreme events. The objective of this work is to study daily extreme rainfall and temperature events over South America as represented by four regional models and new observational datasets. Within the framework of the EU FP6 project CLARIS, four regional modelling groups have coordinated simulations of South American present climate (1992-2000). The models domains cover the entire continent and are driven at the lateral boundaries by ERA40 reanalysis with a horizontal resolution of 50 km. Daily rainfall data used in this study were provided by different national institutions in Argentina, Uruguay, Brazil, Bolivia and Paraguay. The distribution of gauges is relatively homogeneous, except in certain areas in western Argentina and some areas of Brazil and Uruguay. The observational rainfall data are distributed across the La Plata Basin and southern South America regions (70W-40W and 40S-15S). We define a rain day as one on which the rainfall is greater than 0.1 mm and extreme rainfall event is considered when the daily rainfall is greater than the 75th, 90th and 95th daily percentiles threshold. For each subregion selected, the different daily percentiles are calculated on annual and seasonal time scales for each meteorological station and for each grid point and model. Due to the nature of climate model, representing the rainfall of an area determined by the model resolution (in this case 50x50km), we expect the percentiles of the station data to be higher than the percentiles of the four models. However, the idea behind this comparison is to verify if the models can capture qualitatively the differences between different regions surrounding the La Plata Basin. The daily temperature data used in this study belongs to an observational gridded dataset with the same resolution as the models that was developed within the framework of the EU FP7 project CLARIS LPB. The daily gridded dataset of surface minimum and maximum temperature covers the region 70W-45W and 40S-20S. For each variable, a warm extreme is considered when the daily temperature is greater than the 75th, 90th, and 95th percentile and a cold extreme when is lower than the 25th, 10th, and 5th percentile. For each grid point, the difference between daily percentiles calculated from model and observational data is computed on a monthly basis. For minimum temperature, all RCMs tend to overestimate warm extremes and underestimate cold extremes. For maximum temperature extremes, results are not so consistent between models. In future studies, the results presented here will be compared with simulation outputs from updated model versions driven by ERA-INTERIM.