DISPOSICIONES ADICIONALES

One of the important world-wide assessments of changes in observed daily rainfall extremes have been made by Alexander et al. (2006), who presented a comprehensive global picture of trends in extreme rainfall indices for the period of 1951-2003 using results from a number of workshops held in data-sparse regions and high-quality station data supplied by numerous scientists worldwide. Their results showed widespread

rainfall changes with significant increases in extreme rainfall throughout the 20th

century, but the changes are much less spatially coherent compared to temperature changes and the change in annual and/or seasonal rainfall (Groisman et al., 2005). Global rainfall projection data suggest more spatially variable results. For example, Semenov and Bengtsson (2002) analysed data from a coupled atmosphere-ocean general

trends in mean and extreme rainfall over continental areas. Predicted rainfall trend patterns in the period of 2000-2099 followed the tendency obtained for 1900-1999 (observed) but with significantly increased magnitudes. On the other hand, negative trends were found in the number of wet days over most of the land areas except high latitudes in the Northern Hemisphere, which suggests more and more intense rainfall in shorter interval of time. Lau and Wu (2007) did a similar study of analysing global data sets from the Global Precipitation Climatology Project (GPCP) and the Climate Prediction Center Merged Analysis Product (CMAP). Their research revealed that there was a significant shift in the probability distribution functions of tropical rainfall during the period of 1979-2003. This shift featured a positive trend in the occurrence of heavy (top 10% by rain amount) and light (bottom 5%) rain events in the tropics during 1979- 2003 and a negative trend in moderate (25-75%) rain events.

Some researchers looked at the frequency and/or severity of global climatic extremes changed during 20th century based on selective indicators derived from daily totals of rainfall. Frich et al. (2002) used such 10 selective indicators during the interval from 1946 to 1999 and reported that extreme rainfall has more mixed patterns of change but significant increases have been seen in the extreme amount derived from wet spells and number of heavy rainfall events. Tebaldi et al. (2006) also examined 10 indicators of extremes derived from an ensemble of 9 GCMs that contributed to the Fourth Assessment Report of the IPCC (IPCC-AR4). They observed that there is a trend towards a world characterized by intensified rainfall, with a greater frequency of heavy- rainfall and high-quantile events, although with substantial geographical variability. Particularly the high latitudes of the northern hemisphere showed the most coherent regional patterns of significant positive changes in the intensity of wet events, as also reported by Sillmann and Roeckner (2008) who calculated the indices for temperature and rainfall extremes on the basis of global climate model ECHAM5/MPI-OM simulations of the twentieth century and SRES A1B and B1 emission scenarios for the twenty-first century. Tebaldi et al. (2006) speculated that the changes in intense rainfall in the northern hemisphere are related to the greater moisture-holding capacity of the warmer air contributing to greater moisture convergence, as well as a pole-ward shift of the storm tracks.

To examine the impact of increasing atmospheric CO2 on high and low extremes of

global monthly-to-annual rainfall, Raisanen (2005) performed twenty model experiments. The author reported that the extremes correlate well with the changes in the long-term mean rainfall. Wet extremes were predicted to be more severe especially where the mean rainfall increased, and dry extremes were predicted to be severe where the mean rainfall decreased. The changes in frequency of extremes were found to be much larger than the changes in their magnitude. Through a similar experiment on equilibrium changes in daily extreme rainfall and temperature events in response to

doubled atmospheric CO2, Barnett et al. (2006) also demonstrated that increase in CO2

lead to changes in the shape of the daily distributions for both temperature and rainfall, but the effect of these changes on the relative frequency of extreme events is generally larger for rainfall.

While global datasets usually refer to increase in rainfall extremes, regional studies tend to give different results since changes in total and extreme rainfall vary depending upon geographic location (Bell et al., 2004). Extreme events may have major societal, economical and environmental impacts at regional level (Easterling et al., 2000). At the local level and over short periods many of these effects are expected to be more adverse (Cruz et al., 2007). Most analyses of long-term climate observations of rainfall have focused on the change in global annual and seasonal average values. However, extremes at the very local level often cause the biggest impacts and are a medium for alarm about climate change at regional levels because extremes considered in one part of a region might not be quite meaningful across the other part (Manton et al., 2001). The fact that engineering works need to be designed for extreme conditions also requires that special attention is paid to singular values at locals. Regional climate extremes occur when climate variability imposes large anomalies on the average state of the climate system. Hence, a detailed understanding of the spatial and temporal variability of the extremes at the very local level is essential.

Studies related to rainfall extremes are discussed separately for the different continents in the globe, as follows.