Resumen
The bivariate probability distribution of significant wave heights and mean wave periods has an indispensable guiding role in the implementation of offshore engineering, which has attracted great attention. This work gives a new bivariate method to describe the bivariate distribution of significant wave height and mean wave period at the NanJi, BeiShuang, and XiaoMaiDao stations from 2018 to 2020. A mixed lognormal distribution is used for univariate probability analysis of wave data, and the method of connecting two mixed lognormal distributions with copula functions is applied to construct bivariate distribution. The results show that compared with Weibull and lognormal distributions, the mixed lognormal distribution shows good performance in fitting marginal distributions. In the bivariate probability analysis, the conditional model overestimates the probability of lower wave heights, and the bivariate function model has a poor fitting effect in the region with larger periods. In contrast, the copula model based on mixed lognormal distribution is more suited to describe the joint distribution of significant wave height and mean wave period.