A copula-based approach for the estimation of wave height records through spatial correlation

Information on the wave climate at a particular location is essential in many areas of coastal engineering from the design of coastal structures to flood risk analysis. It is most commonly obtained either by direct measurements or hindcast from meteorological data. The extended deployment of a wave buoy to directly measure wave conditions and the application of wave transformation models used in hindcasting, including public domain models such as Wavewatch and SWAN, are both expensive. The accuracy of the results given by the latter are also highly sensitive to the quality of the wind data used as input. In this paper a new copula-based approach for predicting the wave height at a given location by exploiting the spatial dependence of the wave height at nearby locations is proposed. By working directly with wave heights, it provides an alternative method to hindcasting from observed or predicted wind fields when limited information on the wave climate at a particular location is available. It is shown to provide predictions of a comparable accuracy to those given by existing numerical models.