INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. (2009) Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/joc.1863 Future change of temperature and precipitation extremes in South America as derived from the PRECIS regional climate modeling system J. A. Marengo, a * R. Jones, b L. M. Alves a and M. C. Valverde a a CPTEC/INPE, Rodovia Dutra Km, 40 12630-000 Cachoeira Paulista, S˜ ao Paulo, Brazil b UK Met Office Hadley Centre (Reading Unit), Meteorology Building, University of Reading, Reading, RG6 6BB, UK ABSTRACT: Using the PRECIS regional climate modeling system this study analyses the distribution of extremes of temperature and precipitation in South America in the recent past (1961–1990) and in a future (2071–2100) climate under the IPCC SRES A2 and B2 emissions scenarios. The results show that for the present climate the model simulates well the spatial distribution of extreme temperature and rainfall events when compared with observations, with temperature the more realistic. The observations over the region are far from comprehensive which compromises the assessment of model quality. In all the future climate scenarios considered all parts of the region would experience significant and often different changes in rainfall and temperature extremes. In the future, the occurrence of warm nights is projected to be more frequent in the entire tropical South America while the occurrence of cold night events is likely to decrease. Significant changes in rainfall extremes and dry spells are also projected. These include increased intensity of extreme precipitation events over most of Southeastern South America and western Amazonia consistent with projected increasing trends in total rainfall in these regions. In Northeast Brazil and eastern Amazonia smaller or no changes are seen in projected rainfall intensity though significant changes are seen in the frequency of consecutive dry days. Copyright  2009 Royal Meteorological Society KEY WORDS downscaling; climate change; vulnerability Received 3 August 2007; Revised 5 January 2009; Accepted 11 January 2009 1. Introduction Large departures from a mean climate state (hereafter ‘extreme events’) occur on scales ranging from days to millennia though in this paper we will focus short-term extreme (weather related) events given their potential for significant impacts. Extreme climate events are also an integral aspect of climate variability, i.e. defining its envelope as well as being important contributors to the mean and other measures such as its standard devia- tion. This explains the difficultly in viewing short-term extreme events as a manifestation of either ‘weather’ or ‘climate’. From a climatological point of view, short-term extremes are events whose probability of occurrence is conditioned by larger scale, longer lasting patterns which dynamically favour or inhibit the short-term events. One of the most important questions regarding short- term extreme events is whether their occurrence is increasing or decreasing over time; i.e. whether the shape of the distribution, which defines if these events occur more or less frequently, is changing significantly. Statistically this can be thought of as resulting from changes of the variable’s mean and/or other parameters of * Correspondence to: J. A. Marengo, CPTEC/INPE, Rodovia Dutra Km, 40 12630-000 Cachoeira Paulista, S˜ ao Paulo, Brazil. E-mail: marengo@cptec.inpe.br its distribution or climatologically as a result of changes in the frequency and intensity of the associated weather events themselves and/or the occurrence of the large-scale patterns conditioning these events. Thus a sequential analysis of time series can determine the existence of trends without giving any information on the underlying causes (even in the absence of abrupt changes which may have contributed). Cold spells, heat and cold waves, intense rainfall, floods, dry spells among other extreme events affect South America in all seasons, and their impacts vary according to the sector. Heavy or extreme precipita- tion events have important effects on society, since flash floods associated with intense, but often brief, rainfall events may be the most destructive of extreme events. Over many areas the frequency of heavy precipita- tion events has increased, consistent with warming, and widespread changes in extreme temperatures have been observed over the last 50 years (Intergovernmental Panel on Climate Change IPCC, 2007a; Marengo et al., 2007). Cold days, cold nights and frost have become less fre- quent; while hot days, hot nights, and heat waves have become more frequent (Vincent et al., 2005; Alexan- der et al., 2006; Caesar et al., 2006; Haylock et al., 2006; Tebaldi et al., 2006). Such changes in extremes have impacts on human activities such as agriculture, Copyright  2009 Royal Meteorological Society