Seasonal prediction of US landfalling hurricane wind energy from 1 August

Hurricanes rank historically above earthquakes and floods as the major geophysical cause of property damage in the United States. The annual mean damage bill and its standard deviation for hurricanes striking the continental US 1950-2002 is US $ 4.8 billion and US $ 7.7 billion respectively. Skillful seasonal prediction of US landfalling hurricane activity would benefit business, government and society by forewarning of damage and disruption. However, significant seasonal landfalling skill has not been reported to date. This contrasts with the demonstrated significant skill for the seasonal prediction of North Atlantic hurricane activity from 1 August. Here we show that seasonal US landfalling hurricane wind energy 1950-2002 is predictable from the 1 August start of the main Atlantic hurricane season with significant (p < 0.01) and useful skill. Predictability arises from a largescale pattern of North Atlantic tropospheric wind variability in July which establishes persistent steering winds that either favour or hinder US hurricane landfall. Predictions from this model are linked significantly (p ~ 0.01) to US hurricane economic and insured losses 1950-2002, and offer application to business.