A Statistical Study of Tropical Cyclone Positioning Errors with Economic Applications.

摘要

Hurricane landfall forecasts are based heavily on the latest available motion vector and position of a storm. Inaccuracies in these data are closely related to errors in the time and place of storm landfall. The study uses a Monte Carlo simulation of hurricane positioning errors to determine a statistical relationship between positioning errors and landfall errors. It is shown that for a typical 18-hour landfall forecast, approximately 22% of the landfall position error can be attributed to initial data uncertainties. It is further shown that a 20% increase in the size of a hurricane warning zone can be expected if the currently observed positioning errors are increased an average of 10 nmi. However, a 10 nmi decrease in positioning error yields only an 11% decrease in the size of the warning area. An economic analysis of potential changes in the size of hurricane warning areas is also shown. It is estimated that protection costs for a typical 300 nmi Gulf of Mexico coastal hurricane warning zone total $25.1 million.