A new method for modelling the space variability of significant wave heights (Q2463682)

The paper is devoted to stochastic modelling of significant wave heights \(H_s\) (\(s\) being the spatial location) in the North Atlantics based on data of the TOPEX-Poseidon satellite, which measures \(H_s\) along one-dimensional tracks. In the model \(X_s=\log H_s\) is considered as a sum of two independent Gaussian processes which describe a long scale dependence connected with the storms and waves propagations and a medium scale dependence which could be considered as a correlated noise. Estimates of the model parameters are based on the moments of total variation of \(X_s\). The fitted model is used to compute the distribution of the global maximum of \(H_s\).