Author(s)

J. L. Vega & G. Rodríguez

Abstract

This paper presents the probabilistic modelling of the mean wave direction derived from directional spectral analysis of waves recorded by directional buoys. The analysis is performed on the mean wave direction in terms of the climatic season, the sea state severity and the period of the dominant waves. The usefulness of the von Mises theoretical models to describe the empirical kernel density estimates is examined. It is observed that the single von Mises theoretical model results are useful to fit the observed distribution only for moderate and severe sea states while the mixture of two von Mises distributions enhances significantly the degree of fitness. Keywords: wave modelling, mean wave direction, kernel density estimation, circular variables, von Mises distribution, von Mises mixtures.

Keywords

wave modelling, mean wave direction, kernel density estimation, circular variables, von Mises distribution, von Mises mixtures