Medium-Range Forecast of Nearshore Wave Based on Global Wave Ensemble System Using Artificial Neural Network

The workability of vessels largely depends on wave conditions. Installation of offshore wind turbines, which usually takes several days, requires a medium-range wave forecast with a lead time of around 10 days to schedule the constructions. Accurate prediction of wave conditions is important for ensuring safety and reducing the cost of such offshore constructions. This study aims to develop a high-resolution medium-range wave forecast method for nearshore areas based on the prediction of a global wave ensemble system (WENS) using artificial neural network (NN). A feed forward neural network is constructed to predict nearshore waves using WENS, which provides probabilistic information on global ocean waves with a lead time of 264 h. It is found that using the forecast data of the closest grid to the target site is sufficient to provide a reliable prediction. Prediction accuracy can be improved by using the average of all 51 ensemble members as input of NN. Relatively high accuracy of medium-range forecast was achieved indicating potential applications in scheduling offshore construction works.