13.3 Assumptions about future trends

13.3.1 Climate

13.3.1.1 Climate-change scenarios

Even though climate-change scenarios can be generated by several methods (IPCC, 2001), the use of GCM outputs based on the Special Report on Emissions Scenarios (SRES: Naki?enovi? and Swart, 2000) is the adopted method for the Fourth Assessment Report (AR4). Projections of average temperature and rainfall anomalies throughout the current century derived from a number of GCMs are available at the IPCC Data Distribution Centre (IPCC DDC, 2003; http://www.ipcc-data.org//) at a typical model resolution of 300 km, and for two different greenhouse-gas (GHG) emissions scenarios (A2 and B2). Additionally, Chapter 11 of the Working Group I Fourth Assessment Report (Christensen et al., 2007) presents regional projections for many parts of the world. Table 13.4 indicates ranges of temperature and precipitation changes for sub-regions of Latin America for several time-slices (2020, 2040, 2080), obtained from seven GCMs and the four main SRES emissions scenarios.

For 2020, temperature changes range from a warming of 0.4°C to 1.8°C, and for 2080, of 1.0°C to 7.5°C. The highest values of warming are projected to occur over tropical South America (referred to as Amazonia in Table 13.4). The case for precipitation changes is more complex, since regional climate projections show a much higher degree of uncertainty. For central and tropical South America, they range from a reduction of 20% to 40% to an increase of 5% to 10% for 2080. Uncertainty is even larger for southern South America in both winter and summer seasons, although the percentage change in precipitation is somewhat smaller than that for tropical Latin America. Analyses of these scenarios reveal larger differences in temperature and rainfall changes among models than among emissions scenarios for the same model. As expected, the main source of uncertainty for regional climate change scenarios is that associated with different projections from different GCMs. The analysis is much more complicated for rainfall changes. Different climate models show rather distinct patterns, even with almost opposite projections. In summary, the current GCMs do not produce projections of changes in the hydrological cycle at regional scales with confidence. In particular the uncertainty of projections of precipitation remain high (e.g., Boulanger et al., 2006a, b, for climate-change scenarios for South America using ten GCMs). That is a great limiting factor to the practical use of such projections for guiding active adaptation or mitigation policies.

GCM-derived scenarios are commonly downscaled using statistical or dynamical approaches to generate region- or site-specific scenarios. These approaches are described in detail in Chapter 11 of the Working Group I Fourth Assessment Report (Christensen et al., 2007). There have been a number of such exercises for South America using an array of GCM scenarios (HADCM3, ECHAM4, GFDL, CSIRO, CCC, etc.), usually for SRES emissions scenarios A2 and B2: for southern South America (Bidegain and Camilloni, 2004; Nuñez et al., 2005; Solman et al., 2005a, b), Brazil (Marengo, 2004), Colombia (Eslava and Pabón, 2001; Pabón et al., 2001) and Mexico (Conde and Eakin, 2003). Downscaled scenarios may reveal smaller-scale phenomena associated with topographical features or mesoscale meteorological systems and land-use changes, but in general the uncertainty associated with using different GCMs as input is a dominant presence in the downscaled scenarios (Marengo and Ambrizzi, 2006).