Estimation of Hurricane Wind Speed Probabilities: Application to New York City and Other Coastal Locations

Abstract

A procedure is presented for estimating parametric probabilistic models of hurricane wind speeds from existing information on state-of-the-art estimates of wind speeds with various mean recurrence intervals (MRIs). Such models may be needed, for example, for the estimation of hurricane wind speeds with long MRIs required for the performance-based design of structures susceptible of experiencing nonlinear behavior. First, the procedure is applied to the case where that information is obtained from ASCE 7-10 wind maps, and examples are provided of its application to a number of coastal mileposts on the Gulf and Atlantic coasts. Next, the procedure is applied by using, in addition to the ASCE 7-10 information, hurricane wind speeds with 1,000,000- and 10,000,000-year MRIs estimated in a 2011 Nuclear Regulatory Commission (NRC) report. It is then argued that ASCE 7-10 Standard basic wind speeds for New York City are not conservative with respect to their counterparts specified in the standard for other U.S. hurricane-prone locations. Finally, it is shown that, for the randomly selected cases examined here, best-fitting extreme value distributions of hurricane wind speeds typically have finite upper tails of the reverse Weibull type, rather than infinite upper tails of the Gumbel type. This result, if confirmed by additional studies, may help to change the still widely held belief that extreme wind speeds are modeled appropriately only by the Gumbel distribution.