Nearshore wave prediction by coupling a wave model and statistical methods

An onshore wave prediction model is proposed by coupling a wave model and statistical methods: principal component analysis and Kalman filter. The wave model computes the offshore wave field, and the statistical methods evaluate the onshore wave climate from the computed offshore wave spectra. The wave climate changing from offshore to onshore, due among others to the localized coastal geometry, is considered implicitly by the statistical methods. Incompleteness of modeling the physical processes and uncertainties of the models' input data cause the limitation of wind wave prediction in the nearshore with numerical models. However, these problems can be avoided by statistical models that predict the waves with observation data without considering physical wave processes. In the model proposed here, the long-term fluctuations, e.g. seasonal fluctuations, are also considered. Field observation data at the Japan Sea, where the wave climate is quite different in summer and winter due to monsoons, are used to probe the applicability of this model. The model predicts appropriate wave heights throughout a year. A sensitivity analysis of the model parameters is performed for discussing the characteristics of the model.