ARTIFICIAL NEURAL NETWORKS IN THE FORECASTING OF WAVE PARAMETERS

Assessing the present status of the spectral wind wave modeling, Liu et al. (2002) compared four prediction models representing different levels of sophistication all based on the concept of a wave energy spectrum. It was concluded that the models performed in a similar way reflecting the general trend and patterns presented in observations. The differences between the results of computations were of the same order of magnitude as the differences between model results and observations. There is, therefore, a certain need to explore alternative approaches.