Statistical analysis of long-term sea states

A joint probability density function of significant wave height, H-5, and average zero-crossing period, (T) over bar(o), is developed as a product of the generalized gamma distribution for the marginal density function f(H-5), and the log-normal distribution for the conditional probability density function f((T) over bar(o) \ H-5). Based on this probability density function, methods for estimating the severest sea condition as well as the domain of all sea conditions expected to occur in the long-term (50 and 100 years, for example) are developed. The joint probability distribution is applied for analysis of ten data sets of long-term wave statistics accumulated by the National Data Buoy Center. It is found that the newly developed joint density function represents all data sets with sufficient accuracy, and that the contour curve of 1 x 10(-5) computed by the joint probability density function covers almost all data points in each contingency table of long-term wave statistics.