Here we assess only AOGCM-related results pertaining to tropical cyclones. For further discussion of results from embedded and mesoscale models regarding possible future changes in tropical cyclone activity, see Chapter 10 (also refer to Box 10.2 for a summary). The ability of global models to accurately represent tropical cyclone phenomena, and their present limitations in this regard, is important for understanding their projection of possible future changes. These capabilities are discussed in detail in Chapter 8.
Some of the global climate models suggest an increase in tropical storm intensities with CO2-induced warming (Krishnamurti et al., 1998), though a limitation of that study is the short two year model run. However, the highest resolution global climate model experiment reported to date (Bengtsson et al., 1996; see Chapter 10) still has a resolution too coarse (about 1°) to simulate the most intense storms or realistically simulate structures such as the hurricane eye.
Indices of tropical cyclone activity (Gray, 1979) summarise the necessary large-scale conditions for tropical cyclone activity from coarse resolution GCMs (Evans and Kempisty, 1998; Royer et al., 1998). The latter study examined large-scale atmospheric and oceanic conditions (vertical shear, vorticity and thermo-dynamic stability), and suggested that only small changes in the tropical cyclone frequencies would occur (up to a 10% increase in the Northern Hemisphere primarily in the north-west Pacific, and up to a 5% decrease in the Southern Hemisphere). Climate change studies to date show a great sensitivity to the measure of convective activity chosen, and depend less on the model produced fields. Additionally, the broad geographic regions of cyclogenesis, and therefore also the regions affected by tropical cyclones, are not expected to change significantly (Henderson-Sellers et al., 1998). This is because results from Holland's (1997) Maximum Potential Intensity model show that even with substantial (1 to 2°C) SST increases in the tropics from global warming, one would also get a correspondingly much bigger warming in the upper troposphere leading to very little change in the moist static stability (Holland, 1997). Another study shows areas of deep convection that can be associated with tropical cyclone formation would not expand with increases in CO2 due to an increase of the SST threshold for occurrence of deep convection (Dutton et al., 2000). Additionally, since tropical storm activity in most basins is modulated by El Niño/La Niña conditions in the tropical Pacific, projections of future regional changes in tropical storm frequencies may depend on accurate projections of future El Niño conditions, an area of considerable uncertainty for climate models (as noted in Section 184.108.40.206).
Other reports in this collection