1893, 1938, and 2012 (Sandy): Quantifying the role of Fujiwhara interaction on U.S. tropical cyclone landfall risk

Motivated by the historical cases of 1893 (New York City), 1938 (New England) and the 2012 case of Hurricane Sandy, a climatology of the occurrences of Atlantic vortex-vortex interaction (Fujiwhara 1921; Lander and Holland 1993) involving at least a single TC is presented. The first half presents the interaction associated with multiple TCs (e.g. 1893), while the second half of the talk presents interaction associated with a single TC and a non-tropical closed low (e.g. 1938 or Sandy).

The spatial distribution of the Fujiwhara occurrences is analyzed, including not only the simultaneous occurrence of multiple vortices within critical threshold distances, but also quantification of the future track impact of that interaction (assuming the future tracks are irrevocably changed by that interaction). While the absolute frequency of multiple TC interaction in the Atlantic is maximized near the maximum of overall TC occurrence (NE and E of the Bahamas), the relative frequency of vortex interaction has a much broader distribution with local maxima extending to the U.S. East Coast. The duration of TC interaction has a climatological plateau from 6hr to 2 days with rapidly decreasing occurrence from 3 days to a maximum of approximately one week. The PDF of TC motion shifts from a bimodal distribution (expectedly toward WNW and NE for overall Atlantic TC existence) to a unimodal distribution (toward the N for interacting TCs). Since the Fujiwhara interaction should alone lead to no significant shift in TC motion (due to opposing motions for each TC), one possible explanation for the overall shift in motion is due to an enhanced Beta-drift resulting from the larger cyclonic gyre encompassing interacting TCs. The regional track influence of TC-TC interaction is variable, with a slightly enhanced landfall threat along the U.S. East Coast (particularly the NE) and a slightly decreased landfall threat along the U.S. Gulf Coast. This regionality results from the mean climatological location of the opposing TC.

The talk concludes with a similar examination of the spatiotemporal climatology of TC-nonTC vortex interaction (e.g. 1938 and Sandy) and example WRF-ARW simulations illustrating the mean results shown above. The results of this analysis suggest that capturing sufficiently the interaction of multiple vortices may be important for the development of stochastic event sets that seek to accurately quantify landfall risk.