Wave hindcasting in Anzali, Caspian Sea: a hybrid approach

Wave characteristics are necessary for many coastal operations and marine structures design. Different methods such as empirical methods, numerical models and soft computing have been developed for this purpose. In this study, one of the well-known numerical models, SWAN model, has been used for the prediction of wave parameters in Anzali, Caspian Sea. Wind field is one of the most important factors in the wave modeling. In this study, QuikSCAT wind field was investigated for wave hindcasting in Anzali using SWAN model. For this purpose, the gaps in the QuikSCAT measurements data were filled using spatial interpolations. Furthermore, the Anzali buoy recorded data were used to calibrate and validate the model. In order to select the appropriate intervals of calibration and verification, the statistical similarities between corresponding data in both periods were considered. Hence, the buoy was located in deep water the whitecapping dissipation coefficient was more influential than friction and breaking effects on the results. Finally, the numerically modeled wave characteristics were improved using linear transfer function and artificial neural network (ANN). The results indicated that the wave characteristics modification using ANN yields more accurate results in comparison with linear transfer function.