LAUSR

Abstract The short-term distribution of wave periods is very important for ocean and coastal engineering applications. At present, the vast majority of research studies are confined to single-wave systems. Most available theoretical distributions of wave periods, which are based on the narrowband approximation, are inapplicable to actual sea states. This study focuses on the probability distribution of individual wave periods in combined sea states with two parametric mixture distribution models. The expectation-maximisation (EM) algorithm is used to calculate the maximum likelihood estimates of the mixture models. Further, the mixture distributions are compared with other two models: a theoretical and a parametric model. In situ-measured data with two-peaked spectra and simulated data obtained with the six-parameter Ochi–Hubble model allow for a thorough assessment of the distribution models. The patterns of the distributions of wave periods in nine types of mixed sea states are considered and discussed. According to the results, the theoretical distribution model is unsuitable for the description of the distributions in mixed sea states; in particular, when the patterns exhibit bimodal characters. By contrast, despite having a higher calculation complexity, the mixture distribution models provide an improved performance for all combined-sea state cases.