Development of Phase-Resolved Real-Time Wave Forecasting With Unidirectional and Multidirectional Seas

Abstract

Abstract The development of phase-resolved real-time wave forecasting is outlined. This framework is an enhancement over previous work in that the algorithm of real-time wave prediction is extended into multidirectional seas by including the wave measurements and components in direction. However, the computations with multidirectional seas become much more numerically expensive, and hence it is often not possible to accomplish a real-time system of nonlinear ocean wave prediction. Accordingly, we suggest an improved assimilation procedure in the process of wave reconstruction, which is proven to alleviate the computational costs and establish the numerical stability of the Lagrangian approach. In addition, given an observation zone recorded by an optical sensor mounted on a fixed offshore structure, we provide a spatio-temporal prediction zone where it is suitable to obtain the prediction of the wave field by evolving the reconstructed wave information in time and space. In order to validate the phase-resolved wave forecasting, we conducted a tank-scale experimental campaign with unidirectional seas (long-crested waves) and multidirectional seas (short-crested waves). Through the comparison of model performance against the laboratory data between unidirectional and multidirectional seas, it is confirmed that the directional wave components are necessarily considered to increase model accuracy in the multidirectional case as in the unidirectional case.