Maximum covariance analysis of the sea surface backscatter signal models

Abstract

In this work we study the applicability of the maximum covariance analysis (MCA) for the analysis of matrices characterizing the spatiotemporal models of sea surface backscatter signals for different types of sea waves. The method is based on the singular value decomposition of the covariance matrix describing the relationship between two spatiotemporal matrices. The dependence of the obtained correlation coefficients on the degree of sea roughness, as well as on the ratio of the heights of wind waves and rogue waves are determined. The statistical characteristics of the obtained correlation coefficients of the sea surface backscatter signals are analysed. Our results indicate that the MCA method, at least from the modelling perspective, could be applicable to the classification of the sea surface from its backscatter signal characteristics, including an early detection and analysis of the rogue waves onset and development.