Probability Distributions for Hurricane Effects

Abstract

This investigation treats indirect development of probability functions for hurricane effects for specified time intervals. Methods outlined here provide a way of obtaining magnitude-recurrence interval relationships for hurricane effects. Methods are suitable for sites where insufficient data does not permit getting effect magnitude-recurrence interval information directly from historic records. A stochastic model for hurricane occurrences around sites is developed; a periodic Poisson distribution best described hurricanes on the Texas Coast. Data imply that Texas hurricanes tend to follow a cyclic trend with a period of about 33 yr. A hurricane occurrence model is then combined with distribution for hurricane wind, given a hurricane occurrence. If hurricane occurrences follow a periodic Poisson law, then exceedances of given magnitudes of hurricane effects are also shown to obey a periodic Poisson law.