Accuracy of United States Tropical Cyclone Landfall Forecasts in the Atlantic Basin (1976–2000)

Abstract

About 13% of all Atlantic basin tropical cyclone forecasts issued from 1976 to 2000 are for landfalls along the United States coastline, and 2% more are for storms forecast to make landfall in the United States but that remain at sea. Landfall position and time forecasts are skillful at all forecast time periods and are more skillful than Atlantic basin track forecasts as a whole, but within 30 h of predicted landfall, timing errors demonstrate an early bias of 1.5–2.5 h. Landfall forecasts are most accurate for storms moving at oblique or normal angles to the coastline and slow-moving storms. During the last quarter century, after adjustment for forecast difficulty, no statistically significant improvement or degradation is noted for landfall position forecasts. Time of landfall forecasts indicate no degradation at any period and significant improvement for the 19–30-h period. The early bias and lack of improvement are consistent with a conservative or “least regret” forecast and warning strategy to account for possible storm accelerations. Landfall timing uncertainty is ~11 h at 24 and 36 h, which suggests that hurricane warnings could be disseminated about 12 h earlier (at 36 h, rather than 24 h, before predicted landfall) without substantial loss of lead time accuracy (although warning areas necessarily would be larger). Reconsideration of National Weather Service Strategic Plan and United States Weather Research Program track forecast goals is recommended in light of these results.