The Central America and the Caribbean (CAC) region is affected by several phenomena, including the ITCZ (Section 22.214.171.124), NAMS (Section 126.96.36.199), ENSO (Section 14.4) and TCs (Section 14.6.1; Table 14.3; also Gamble and Curtis, 2008). The annual cycle results from air–sea interactions over the Western Hemisphere warm pool in the tropical eastern north Pacific and the Intra Americas Seas (Amador et al., 2006; Wang et al., 2007). The Caribbean Low Level Jet is a key element of the region’s summer climate (Cook and Vizy, 2010) and is controlled by the size and intensity of the Western Hemisphere warm pool (Wang et al., 2008b). It is also modulated by SST gradients between the eastern equatorial Pacific and tropical Atlantic (Taylor et al., 2011d). ENSO is the main driver of climate variability, with El Niño being associated with dry conditions and La Niña with wet conditions (Karmalkar et al., 2011). Other teleconnection patterns, such as the NAO (Section 14.5.1) and the strength of boreal winter convection over the Amazon, influence trade winds over the Tropical North Atlantic and can combine with ENSO to modulate the summer Western Hemisphere warm pool (e.g., Enfield et al., 2006). Table 14.3 summarizes the main phenomena and their relevance to climate change over the CAC.
Because inter-decadal climate variations can be large in the CAC region, precipitation trends must be interpreted carefully. From 1950 to 2003, negative trends were seen in several data sets in the Caribbean region and parts of Central America (Neelin et al., 2006). However, regarding secular trends (1901–2005), this signal was identified only in the Caribbean region (Trenberth et al., 2007b). Prolonged dry or wet periods are related to decadal variability of the adjacent Pacific and Atlantic (Mendoza et al., 2007; Seager et al., 2009; Mendez and Magaña, 2010), and the intensity of easterlies over the region. For instance, increased easterly surface winds over Puerto Rico from 1950 to 2000 disrupted a pattern of inland moisture convergence, leading to a dramatic precipitation decrease (Comarazamy and Gonzalez, 2011).
Table 14.2 provides an overall assessment of GCM quality for simulations of temperature, precipitation and main phenomena in the CAC sub-regions. Annual cycles of temperature and precipitation are well simulated by CMIP5 models, though precipitation from June to October is underestimated (Figure 9.38). Regional models also simulate temperature and precipitation climatologies, and the magnitude and annual cycle of the Caribbean Low-Level Jet reasonably well (Campbell et al., 2010; Taylor et al., 2013).
CMIP3 models generally projected a precipitation reduction over much of the Caribbean region, consistent with the observed negative trend since 1950 (Neelin et al., 2006; Rauscher et al., 2008). The subtropics are generally expected to dry as global climate warms (Held and Soden, 2006), but in both CMIP3 and CMIP5 models the CAC region shows the greatest drying. Future drying may also be related to strengthening of the Caribbean Low-Level Jet (Taylor et al., 2013) and subsidence over the Caribbean region associated with warmer SSTs in the tropical Pacific than Atlantic (Taylor et al., 2011d). A high-resolution regional Ocean GCM using a CMIP3 ensemble for boundary conditions confirms that the Intra American Seas circulation weakens by similar rate as the reduction in Atlantic Meridional Overturning (Liu et al., 2012c). This weakening causes the Gulf of Mexico to warm less than other oceans. Downscaling experiments for the region have shown a mid-21st century warming between 2°C and 3°C (Vergara et al., 2007; Rauscher et al., 2008; Karmalkar et al., 2011). Precipitation decreases over most of the CAC region, similar to the signal in driving global models (Campbell et al., 2010; Hall et al., 2012). However, only a few downscaling studies took into account key elements of the region’s climate, such as easterly wave activity, TCs, or interannual variability mechanisms linked to ENSO (Karmalkar et al., 2011).
By century’s end, CMIP5 models project greatest warming in the CAC region in JJA. Warming is projected to be larger over Central America than the Caribbean in summer and winter (Figures AI.24, AI.2, Table 14.1). From October to March, ensemble mean projections indicate precipitation decrease in northern Central America, including Mexico. In the Caribbean precipitation is projected to decrease in the south (consistent with the observed trends) but to increase in the north (Figure AI.26). From April to September, the projected zone of precipitation reduction expands over the entire CAC region, and this signal is generally larger than the models’ estimates of natural variability (Figure AI.27). Precipitation changes projected by CMIP3, CMIP5 and a high-resolution model show a similar reduction in parts of Mexico and the southern Caribbean in DJFM, and in Central America and the Caribbean in JJAS (Figure 14.19). The CMIP5 ensemble shows greater agreement in the DJFM precipitation increase in the northern Caribbean sector than CMIP3. These projected changes are also reflected in Table 14.1. Figures AI.26, AI.27 and Figure 14.19 suggest an intensification and southward displacement of the East Pacific ITCZ, which can contribute to drying in southern Central America (Karmalkar et al., 2011).
ENSO will continue to influence CAC climate, but changes in ENSO frequency or intensity remain uncertain (Section 14.4). Projected drier conditions may also be related to decreased frequency of TCs, though the associated rainfall rate of these systems are higher in future projections (Section 14.6.1).
In summary, owing to model agreement on projections and the degree of consistency with observed trends, it is likely warm-season precipitation will decrease in the Caribbean region, over the coming century. However, there is only medium confidence that Central America will experience a decrease in precipitation.