Climate Change and Climatic Variability in Latin America and the Caribbean
347
Against this backdrop, the array of models that yield robust projections of the global rise in mean temperature yields much less robust estimates of regional climate change and variability (IPCC 2007). To the extent that assessments of future climate change and variability can be made, they are the following: • All of Central and South America is very likely to warm during this century. The annual mean warming is likely to be similar to the global mean warming in southern South America but larger than the global mean warming in the rest of the area. • Annual precipitation is likely to decrease in most of Central America, with the relatively dry boreal spring becoming drier. Annual precipitation is likely to decrease in the southern Andes, with relative precipitation changes being largest in summer. A caveat at the local scale is that changes in atmospheric circulation may induce large local variability in precipitation changes in mountainous areas. Precipitation is likely to increase in Tierra del Fuego during winter and in southeastern South America during summer. • It is uncertain how annual and seasonal mean rainfall will change over northern South America, including the Amazon forest. In some regions, qualitative consistency is seen among the simulations (rainfall increasing in Ecuador and northern Peru, and decreasing at the northern tip of the continent and in southern Northeast Brazil). • The systematic errors in simulating current mean tropical climate and its variability, and the large differences among models in future changes in El Niño amplitude, preclude a conclusive assessment of the regional changes over large areas of Central and South America. Most models are poor at reproducing the regional precipitation patterns in their control experiments and have a small signal-to-noise ratio, in particular over most of Amazonia. The high and sharp Andes Mountains are unresolved in low-resolution models, affecting the assessment over much of the continent. As with all landmasses, the feedbacks from land use and land cover change are not well accommodated and lend some degree of uncertainty. The potential for abrupt changes in biogeochemical systems in Amazonia remains as a source of uncertainty. Large differences in the projected climate sensitivities in the climate models incorporating these processes and a lack of understanding of