Probability distributions for Texas Gulf coast hurricane effects of engineering interest.

This investigation treats the indirect development of probability functions for hurricane effects for specified time intervals. The methods suggested in this dissertation provide a way of obtaining hurricane effect magnitude-recurrence interval relationships. The methods are suitable for sites where insufficient data does not permit getting the effect magnitude-recurrence interval information directly. The hurricane effects considered are the maximum values for individual storms of wind gusts, significant wave heights, and storm surge elevations. The method is illustrated for two Texas sites, one in 300 feet of water, and one on the Gulf coastline. A simulation procedure is used to derive distributions for the hurricane effects, given that a hurricane occurs. This derived distribution technique combines probabilistic information on storm paths with probabilistic models for the hurricane wind, wave, and storm surge fields. A stochastic model for hurricane occurrences around the sites is developed; a periodic Poisson distribution best describes hurricane occurrences on the Texas coast. The data implies that Texas hurricanes tend to follow a cyclic trend with a period of about 33 years. The hurricane occurrence model is then combined with the distributions for hurricane effects, given a hurricane occurrence. If hurricane occurrences follow a periodic Poisson law, then the exceedances of given magnitudes of hurricane effects are shown to also obey a periodic Poisson law. This investigation indicates that hurricane design criteria for Texas offshore and coastal structures should be modified according to the stage of the hurricane cycle at construction.