| To improve the accuracy of wind wave and swell components separation of one-dimensional wave spectrum, the paper proposes a separation method for wind wave and swell components based on one-dimensional wave spectrum data from the National Data Buoy Center and the wave spectrometer. The method consists of two main parts, the parameter optimization part of the spectral model and the discriminant part of the decision function, which corresponds to the separation of the wave components in the wave spectra and the identification of wind waves and swell, respectively. Firstly, the spectral models developed by the North Sea Joint Wave Program are utilized to fit one-dimensional wave spectra for wave component separation. Each spectral model represents a wave component. The spectral model parameters are iteratively optimized by a double-simulated annealing algorithm. Secondly, based on four buoys data in December 2022, a decision function on wave period and wind speed is built by a Support Vector Machine to identify each wave component as wind waves or swell. Finally, spectral models belonging to the same wave component are merged to obtain a wind wave component and a swell component. The method is tested using data from two additional buoys in June 2022. The results show that the method in the paper is in better agreement with the two-dimensional spectral wind waves-swell separation method compared to the modified wave steepness method currently used by the United States National Data Buoy Center. In addition, the method can also be utilized on the wind wavesswell separation of one-dimensional wave spectrum data from the wave spectrometer. |
