Modelling trivariate distribution of directional ocean data in the Barents Sea seasonal ice zone

The Arctic Ocean has proved to possess a wealth of natural oil and gas resources. Although the ice cover complicates marine industrial activities and transportation, exploratory drilling and processing facilities operation is possible in the region where the sea ice is only present for a portion of the year. This study focuses on the statistical characteristics of directional wave climate in the seasonal ice zone of the Barents Sea. The joint distributions of significant wave height, mean wave period, and mean wave direction were constructed using a mixture trivariate distribution model. The total ocean data were classified into four groups based on the relative weights of the energy content of wind wave and swell fields. The joint distribution for each wave component was constructed by the vine copula model, where the marginal distributions can be well described by the mixture univariate distribution models, the bivariate distributions of angular-linear variables were fitted by the Johnson-Wehrly model adequately, and the bivariate distribution of the linear-linear variable can be accurately modelled by a product of two Gaussian copulas. The results suggested that the mixture trivariate distribution for the total sea state constructed by a linear combination of the four submodels is capable of providing reasonable description for the directional wave data. To generate trivariate samples for model evaluation, an improved pseudorandom number generation algorithm was applied. Furthermore, the environmental contours associated with return periods of 1-, 5-, 20-, and 100-year in different directions were calculated and used to determine the extreme tensions of a semi-submersible platform model based on the global maxima approach.