Having reviewed the factors affecting tropical cyclones during the present climate, we next consider the projections of climate change relevant to the Caribbean as agreed in the most recent review by the Intergovernmental Panel on Climate Change (IPCC, 2007a and b). Utilizing these combined resources, we develop projections of Caribbean tropical cyclone changes due to global warming. In the final section of the paper we will summarize these changes in the context of health impacts.
IPCC climate change projections relevant to the Caribbean
Here we review the IPCC (2007a) projections of regional climate changes relevant to the Caribbean. The IPCC expresses confidence in their projections and employs terminology that can be equated to probability of occurrence using the following scale: Virtually certain > 99%, Extremely likely >95%,Very likely (VL) > 90%, Likely (L) > 66%, More likely than not (medium confidence, M) > 50%, Unlikely (U) < 33%, Very unlikely (VU) < 10%, Extremely unlikely < 5% (IPCC, 2007b).
Caribbean islands are very likely to warm during this century (1.33.4 C, IPCC 2007a Table 11.1), but this warming is likely to be smaller than the global annual mean temperature. Summer rainfall in the Caribbean is likely to decrease in the vicinity of the Greater Antilles but changes elsewhere (and in winter) are uncertain (the median decrease is 12 %, but model scenarios range from a 58 % decrease to 19 % increase; IPCC, 2007a; Table 11.1); these precipitation changes may translate to a 20-40 % increase in dry summer (JJA) and fall (SON) seasons.
Finally, on average sea levels are likely to rise during this century in the Caribbean Sea; this rise will likely not be geographically uniform but large deviations among model predictions make estimates across the Caribbean uncertain. Sea level rise is of general concern as, in addition to the direct loss of coastal lands, rising sea levels shrink the available fresh water lens for island nations, thereby reducing potable water supplies. In the context of tropical cyclones, sea level rise can lead to storm surge inundation into previously safe regions. This can also affect freshwater flooding, since runoff from the rains drains into the rivers, which ultimately empty into the sea; storm surge “piles up” water at the mouths of the rivers, preventing the rivers from draining and causing the waters to spread out into freshwater floods.
Projections of changes to Caribbean tropical cyclones due to climate warming
The IPCC has made conclusions on tropical cyclone changes due to global warming (summarized in Table 2), but confines these conclusions to general statements on global trends (IPCC, 2007a). Here we will combine our previous discussion on factors affecting tropical cyclones with detailed results from Global Climate Model (GCM) projections of features of the future climate important to tropical cyclones. Over thirty separate modeling groups contributed to reports on global climate change (IPCC, 2007a and b), providing a resource for our exploration.
Our decadal analyses were suggestive of a direct link between ENSO and Caribbean tropical cyclone activity (more Caribbean storms in an ENSO phase on average), so we ask what happens to ENSO in a warmer climate? Unfortunately, the answer is not clear. The news on ENSO from the GCM is mixed: many models project more ENSO events, and even go as far as suggesting that the climate may be more “ENSO-like” on average, but concerns with details of the representation of clouds in the models require caution using these results.
Why not just count the number of “tropical cyclones” predicted in each GCM? This is because while the GCMs are generally reliable at predicting larger weather features (such as monsoons and winter storms), they can vary widely on projecting trends for smaller weather systems: many models show the global number of tropical cyclones decreasing, but a reasonable fraction still projects an increase in global numbers of tropical storms (IPCC, 2007a). The discrepancy between model results becomes larger as the region of interest becomes smaller and smaller. The skill of GCM has improved through time, including for tropical cyclones. For now though, we need an indirect approach for using GCMs as tools to help infer future tropical cyclone activity.
Failing in the “direct approach” and the ENSO path, an investigation of changes in the necessary conditions for tropical cyclone formation can provide yet another way to infer likely future changes in tropical cyclones. Increasing ocean temperatures with global warming suggest more thunderstorm activity in the tropics (Ibid.), providing the necessary fuel for tropical cyclones. The global monsoons maintain similar locations and seasonality, so the regions of tropical cyclone formation are sustained and we continue to expect storms to form off West Africa and to track west towards the Americas.
Given that these favorable environments will persist, the question for inferring a future tropical cyclone climate in the North Atlantic is: what will happen to vertical wind shear, the SAL and other negative factors affecting tropical cyclones? Recent studies suggest that vertical wind shear will become stronger (remember that we linked it to the monsoons and ocean temperatures?). Stronger vertical wind shear inhibits intensity in tropical cyclones, and this includes suppressing their formation altogether (Vecchi and Soden, 2007; WMO, 2006; Knutson et al., 2010). However, in some ways the jury is still out: while the evidence for a decrease in tropical cyclones globally with a warmer climate is beginning to gain traction (Knutson et al., 2008), the current thinking is that the storms that do occur may be somewhat more intense (WMO, 2006; IPCC, 2007a; Emanuel, 2006; Knutson et al., 2010).
So what does this say about the future of tropical cyclones in the Caribbean? The most likely scenario is that North Atlantic tropical cyclone numbers will decrease in time, and so will the number of storms impacting the Caribbean. However, if (i) our suggestions of a direct link between ENSO and (especially eastern) Caribbean tropical cyclone activity are correct and (ii) the climate becomes more “ENSO-like” then the Caribbean might experience more tropical cyclones than present. This increase in Caribbean storm numbers would likely be larger than in the North Atlantic as a whole since ENSO tends to reduce storm numbers over most of the North Atlantic. Any changes in Caribbean storm numbers could differ between the western and eastern Caribbean, especially if the relative phasing of the NAO and ENSO shifts again. Changes in tropical cyclone tracks will also affect rainfall in the Caribbean since one of the peak rainfall periods is in the hurricane season (Giannini et al., 2000).
Finally, what about intensity? The local ocean temperatures in the Caribbean are not expected to increase as much as the global average (IPCC, 2007a and b) and the peak intensities possible relate to ocean temperatures if everything else is favorable. Thus, theories for the maximum possible tropical cyclone intensity lead us to expect that the most intense tropical cyclones could become even more intense (Emanuel, 2006). However, the intensity of an individual tropical cyclone is the result of the interactions of that storm with other weather systems as well as the ocean (Evans, 1993). The other weather systems change the tropical cyclone environment, such as the vertical wind shear (see Section 2). Vecchi and Soden (2007) explored the changes in vertical wind shear across the North Atlantic simulated by the climate models used in the IPCC (2007a and b). They found that many more models simulated increases in vertical wind shear than decreases. This means that the most likely effect of climate change on the storm environment is to provide an even stronger inhibition to tropical cyclone intensity than in the present climate. Thus, while the most intense hurricanes could become more intense, it is not possible to ascertain the future storm intensities in response to regional environmental changes.