Bivariate maximum entropy distribution of significant wave height and peak period

A bivariate maximum entropy distribution of significant wave heights and the corresponding peak periods is proposed. Two approaches are used to construct the bivariate maximum entropy distribution, the first of which is to adjust equivalent normal bivariate distributions to data transformed from the univariate maximum entropy distributions of wave height and period. The other method is the conditional approach of adjusting one maximum entropy marginal distribution to wave height and several log-normal distributions to the wave periods conditional on the wave height. The bivariate maximum entropy distribution is compared with other bivariate approaches using data from one location in the North Atlantic and a simulated sample by Monte Carlo method.

► A bivariate maximum entropy distribution of wave heights and peak periods is proposed.
► The equivalent bivariate MED is to adjust equivalent bivariate normal distribution to actual data.
► The conditional approach is constructed by marginal MED and conditional Lognormal distributions.