Pre-Season Prediction of Landfalling Tropical Cyclone Frequency

Abstract: It is widely believed that although the conditions that produce hurricane landfalls are well known, it is currently not possible to confidently predict the number or intensity of landfalling hurricanes at extended ranges (http://www.cpc.ncep.noaa.gov/products/ outlooks/hurricane.html). Nevertheless, there have been several reports on the predictability of landfalling tropical cyclones at extended ranges. In this study, statistical models are developed for pre-season prediction of landfalling tropical cyclone frequencies both along the United States East Coast and the mainland coast of China. The predictors for landfalling Atlantic hurricanes include Nino 3 sea surface temperature anomalies, Atlantic sea surface temperature dipole mode index, and the North Atlantic Oscillation index. The statistical models predict three categories of Atlantic hurricane activity: 1) annual frequency of landfall tropical cyclones that maintained hurricane strength at landfall; 2) annual frequency of landfall tropical cyclones that reached hurricane strength at or prior to landfall; and 3) annual frequency of hurricanes that traveled past west of 75W. The predictors for landfalling West Pacific typhoons include Tibetan Plateau snow cover index, Nino 3 sea surface temperature anomalies, and Arctic Oscillation Index. Comparisons of observed and pre-season (April 1 and June 1) predictions of landfalling tropical cyclone frequencies demonstrate measurable skill for both West Pacific typhoons and Atlantic hurricanes.