Binomial-Bivariate Log-normal Compound Model and Its Application in Extreme Sea Condition Prediction

Extreme value analysis is a necessary method for predicting the probability of marine disasters and calculating the conditions of marine engineering. However, the rationality of extreme value analysis results can easily be affected by the lack of sample data. Although the peaks over threshold (POT) method and the process sampling method based on the compound extreme value distribution (CEVD) theory are effective methods to expand samples, they still rely on long-term sea condition data. In order to construct an extreme value model which uses short-term sea condition data instead of traditional annual extreme value series, a Binomial-Bivariate Log-normal Compound Extreme Value Distribution (BBLCED) model which is based on short-term data is established by combination of POT and CEVD. This model not only considers the frequency of extreme sea conditions, but also can reflect the correlation between different sea condition elements and reduce the requirement for the data series length. The model has been applied to the calculation of design wave elements in a certain area of the Yellow Sea and the results indicate that the BBLCED model has a good fitting to the short-term sea condition sequence. There is not much different from the probability prediction results obtained by using the long-term data, indicating that the model can reasonably reflect the probability distribution characteristics of the extreme sea conditions. This model can provide a reliable basis for the coastal engineering design under the condition of lack of marine hydrological data, and it is also suitable for the extreme value prediction calculation in the field of disaster prevention and mitigation.