Cluster analysis of explicitly and downscaled simulated North Atlantic tropical cyclone tracks

The response of tropical cyclone (TC) activity to climate change is a question of major interest. In order to address this crucial issue, several types of models have been developed in the past, such as Global Climate Models (GCMs). However, the horizontal resolution of those models usually leads to some difficulties in resolving the inner core of TCs and then to properly simulate TC activity. In order to avoid this problem, alternative tools have been developed such as downscaled models. Here, we utilize the one developed by Emanuel (2005). This technique uses tracks that are initiated by randomly seeding large areas of the tropics with weak vortices. Then the survival of the tracks is based on large-scale environmental conditions produced by GCMs in our case. Here we compare the statistics of TC tracks simulated explicitly in nine GCMs to the results of the downscaling technique driven by four of the same GCMs in the present and future climates over the North Atlantic basin. Simulated tracks are objectively separated into four groups using a cluster technique (Kossin et al. 2010). The four clusters form zonal and meridional separations of tracks as shown in Figure 1. The meridional separation largely captures the separation between hybrid or baroclinic storms (clusters 1 and 2) and deep tropical systems (clusters 3 and 4), while the zonal separation segregates Gulf of Mexico and Cape Verde storms. This study first shows that, except for the seasonality, the downscaled simulations better capture the general characteristics of the clusters (mean duration of the tracks, intensity…) compared with the explicit simulations. In the second part, we use three different scenarios to examine the possible future changes of the clusters from the downscaled simulations. We explored the role of a warming of the SST, an increase in carbon dioxide and a combination of both ones. The results show that the response to each scenario is highly varying depending on the simulation examined. References - Kossin, J. P., S. J. Camargo, and M. Sitkowski, 2010: Climate modulation of North Atlantic hurricane tracks. Journal of Climate, 23, 3057-3076, DOI: 10.1175/2010JCLI3497.1. - Emanuel, K., 2005: Climate and Tropical Cyclone activity: A new downscaling approach. Journal of Climate, 19, 4797-4802.