Abstract
Artificial Neural Networks (ANNs) are widely used in the field of wave forecasting as data-based soft-computing techniques that do not require prior knowledge regarding the nature of the relationships between the forecasted waves and the controlling physical mechanisms. Among ANN-techniques is the Nonlinear Auto-Regressive Network with eXogenous inputs (NARX), based on which two models were developed in this study to predict the significant wave heights in Eastern Central Red Sea for the next 3, 6, 12 and 24 h. The two NARX-based models differ only by the inclusion of the variance between wind and wave directions in one model and not in the other. Both models have shown the ability to efficiently predict the significant wave heights up to 12 hours in advance. However, the outperformance of the model that included the difference between wind and wave directions indicated the significance of the inclusion of such an input term.