There is still considerable uncertainty concerning the change in tropical cyclone (TC) properties in response to global warming and climate change produced by the increase of greenhouse gas (GHG) forcing (Henderson-Sellers et al. 1998; Walsh 2004; Pielke et al. 2005). The tools appropriate for studying this question are global, or regional, coupled atmosphere-ocean climate models that can be applied in climate simulations for present and future conditions. There are two complementary aspects that need to be addressed, namely changes in tropical cyclone frequencies,
J.B. Elsner and T.H. Jagger (eds.), Hurricanes and Climate Change, 213
doi: 10.1007/978-0-387-09410-6, © Springer Science + Business Media, LLC 2009
that depend on their genesis and trajectories, and changes in tropical cyclone maximum intensity. In this paper only the question relevant to TC genesis in global coupled simulations will be considered. The question of TC intensity requires different methodologies applied to high resolution nested regional models (Knutson and Tuleya 2004; Stowasser et al. 2007).
The methods that have been used to study the genesis of tropical cyclones, in global climate simulations, can be divided in two categories:
• The direct approach by tracking methods
• The indirect approach by means of a cyclogenesis potential based on large-scale environment factors.
The direct approach consists simply in identifying TC-like vortices in sequences of daily (or sub-daily) large-scale fields, and tracking them in order to construct their trajectories. Though conceptually simple, the application of this method requires in practice consideration of different rather complex technical issues. For example, the choice of the detection thresholds needs to be adapted to the model resolution (Walsh et al., 2007), and criteria need to be defined for checking the continuity of the constructed trajectories. In order to yield convincing results this method requires a rather high resolution in the models, so that the simulated tropical vortices can have a reasonable resemblance to observed tropical cyclones. Though it has been applied initially to models with low resolution (Manabe et al. 1970; Broccoli and Manabe 1990, Krishnamurti et al. 1998; Tsutsui 2002), it is only with a grid reaching about 100 km resolution that the vortices start to look somewhat more realistically like TCs, rather than like large-scale lows, as was shown in the pioneering work of Bengtsson et al. (1982, 1995,1996). However the application of this tracking method to the analysis of climate change has still been rather limited (Haarsma et al. 1996; Sugi et al. 2002; McDonald et al. 2005; Chauvin et al. 2006, Oouchi et al. 2006) since only a very small number of high-resolution simulations have been available, and this has precluded the assessment of model uncertainties.
Therefore it may appear preferable to try to assess the simulations by means of indirect methods allowing an estimation of cyclone properties by a combination of large-scale environment factors that have been shown physically relevant to TC formation or development. The most well known of such indices is the Yearly Genesis Parameter (YGP) developed by Gray (1968, 1975). The potential interest of the application of such indices to the analysis of climate simulations is that one can hope that the results should be less dependent on the model resolution than the direct approach, since the index depends on large-scale environmental properties, generally averaged over time. As some of the indices rely on time-averages properties over a month or a season, their computation should be much less costly in computer resources, and much easier to apply in practice than the tracking method that requires analysis of high-frequency (daily or subdaily) maps.
The indirect approach can thus be applied efficiently to a large set of different model simulations, which can be useful to assess the differences in their results and thus the uncertainties in the response (Ryan et al. 1992; Watterson et al. 1995; Tsutsui and Kasahara 1996). The initial application of the YGP to the analysis of climate simulations has given a high increase of the YGP in the case of doubled CO2 simulations (Ryan et al. 1992). Royer et al. (1998) have confirmed the large increase of YGP in the case of doubled CO2 simulations. However they have shown that most of the increase comes from the thermal potential and its dependence on a fixed 26°C SST threshold. They defined a modified index in which the thermal potential is replaced by convective precipitation, and have shown that this modified Convective YGP (CYGP) was able to reproduce tropical cyclogenesis distribution. The CYGP is much less sensitive to surface temperature and provides a more modest and more credible change in cyclogenesis for the doubled-CO2 case.
McDonald et al. (2005) have compared tracked cyclones in high resolution simulations and have found that the CYGP was able to provide a good agreement with the genesis of tropical cyclones. Similar results were found by Chauvin et al. (2006) with higher resolution simulations for the change of the CYGP in future climate simulations. However Camargo et al. (2007) found less good agreement between the actual TC track density or TC numbers and the Genesis Potential Index defined by Emanuel and Nolan. The agreement of the CYGP with the results of cyclone tracking shows that the CYGP could be a useful diagnostic for analyzing cyclogenesis in climate simulations.
The aim of this paper is to illustrate the potential of the CYGP by applying it to a subset of IPCC climate simulations for the past climate, and for an A2 scenario for the future. The methodology is described in section 2, and the models in section 3. The results of applying the CYGP to model simulations for the current climate are given in section 4. The results for scenario A2 at the end of the 21st century are presented in section 5. Finally some conclusions and perspectives are discussed in section 6.
Was this article helpful?