Daily extreme temperature and precipitation compound events in different datasets in southern South America.

OLMO, MATIAS; BETTOLLI MARIA LAURA

New York

Workshop; Workshop on Correlated Extreme Events.; 2019

Columbia University

Southeastern South America (SESA) covers central-northeastern Argentina,Uruguay and southern Brazil (20−40° S, 45−70° W). It is a highly populated regionwith large urban settlements, where socio-economic activities are mainly based onrainfed agriculture and cattle rising and, therefore, vulnerable to extremeprecipitation events. These events are commonly associated with a temperatureextreme event occurring together. In addition, both types of extremes have becomemore frequent in the region during the recent period. However, the understandingof the behavior of extreme temperature and precipitation compound events and itsimpacts in SESA is still very limited. Long records of high-quality and high-resolution observational datasets are necessary to perform climate studies ofspatial and temporal variability of extremes. In some areas of SESA, the density ofmeteorological stations may be very low, and/or their temporal coverage may alsobe limited. Therefore, the characterization and study of these extreme compoundevents over the region should consider as much available information as possible.Taking into account these drawbacks, the aim of this work is to compare dailyextreme temperature and precipitation compound events in two different datasets.To this end, daily gridded temperature and precipitation data from the CPC globaldataset and information from meteorological stations of Argentina, Brazil andUruguay were evaluated in the period 1979-2015. During the warm season, it wasfound that the probability of occurrence of heavy precipitation events (precipitationhigher than the 75th percentile) significantly increased during or following theoccurrence of a warm night (minimum temperature higher than the 90th percentile),mainly in southern SESA. This probability decreased during a cold night (minimumtemperature lower than the 10th percentile). CPC was able to represent the spatialand temporal distribution of the different compound events analyzed whencompared with station data, although the probabilities were slightly overestimated.