TABLE OF CONTENTS This application provides significant wave height estimates for various return periods. Confidence intervals are also provided. The approach developed by Goda (1988) is used to fit five candidate probability distributions to an input array of extreme significant wave heights. Candidate distribution functions are Fisher-Tippett Type I and Weibull with exponents ranging from 0.75 to 2.0. Goodness-of-fit information is provided for identifying the distributions which best match the input data.
General assumptions used in this approach are:
All extreme wave heights come from a single statistical population of storm events. For example, the heights might represent only extratropical storms in the Northern Hemisphere at a site.
Wave height properties for an event are reasonably represented by the significant height.
Extreme wave heights are not limited by any physical factors such as shallow-water depth.
The input array of significant wave heights for extremal analysis is extracted from a longterm data source of measurements, hindcasts, or observations. The reliability of predicted extremes is directly related to the accuracy of available data and the number of years of record. Therefore, the longest high-quality data source available should be used. As a general rule-of-thumb, heights can be extrapolated to return periods up to 3 times the length of record, K. Thus, 20 years of data can be used to estimate significant heights up to a 60-year return period. If return periods longer than 3K are requested,...
