Using a High-Resolution Global Climate Model to Simulate Extreme Extratropical Cyclones

Abstract:

Extreme coastal storms devastate heavily populated areas around the world. Our understanding of exposure to extreme storms is limited due to the short duration of the observational record, which causes difficulty in assessing their true probability of occurrence. Global climate models provide a means of simulating a much larger sample of extreme events, allowing for better resolution of the tail of the distribution. Both tropical and extratropical cyclones (ETCs) occur over the northwestern Atlantic Ocean, and the risks associated with ETCs can be just as severe as those associated with tropical storms (e.g. high winds, storm surge). Therefore, we examine the ability of a high-resolution coupled atmosphere-ocean general circulation model (GFDL FLOR) to realistically simulate extreme ETCs in the northwestern Atlantic Ocean. We analyze similarities between results from a long (i.e. multi-century) FLOR simulation and several atmospheric reanalysis products. After considering differences in spatial and temporal resolution, results indicate that atmospheric measures of ETC intensity are comparable to those diagnosed from reanalyses. The full 1500-year simulation provides a higher frequency of the strongest intensity measures over the northwestern Atlantic Ocean compared with reanalyses. This illustrates that the larger number of realizations in the simulation provides a better opportunity to sample the tail of the ETC distribution. We further investigate the realism of simulated ETCs by using a tracking algorithm to conduct quantitative comparisons of feature, track, cyclogenesis, and cyclolysis densities of simulated ETC subsamples with storms from recent history (using reanalyses).