||The mechanism of freak wave is not clear. Previous studies have suggested the following mechanisms: typhoon weather system, wave groupiness, nonlinear evolution, northeast monsoon, and local topography. There are two research methods for freak wave studies. The first is based on statistics of measured data, and the second is numerical simulation based on mathematical models. In this study, the statistical approach is applied on field measurements at Hualien, Suao and Keelung.|
The analysis shows the significant wave height Hs at Hualien is as large as 0.52 m, while that at Keelung is only 0.21 m. The significant wave period based on the definition of Nair et al. is larger than the result of traditional method by 2 s for all three stations. Usually, the probabilities of freak wave based on either zero up-crossing or zero down-crossing are very close, and the probability for freak waves higher than 1 m is higher during a typhoon. In Keelung, however, the probability for freak waves higher than 1 m in typhoon is only 90% of the long term average, and the freak wave probability calculated by zero up-crossing is twice as that obtained by zero down-crossing. Zero up-crossing also gives larger probability, 1.8 times of that by zero down-crossing, in Hualien if significant wave height is small while significant wave period is large.
Further analyses have been done by dividing field data into groups of different significant wave height, groupiness factor, and run length. Freak wave probability at Keelung has its maximum if significant wave height is between 1 m and 1.5 m, while the peak probability value at Hualien occurs at 1.5 m<Hs <2 m, and that at Suao occurs at 2 m<Hs <2.5 m. At all three stations, freak wave probability increase with the groupiness factor. At both Hualien and Suao, the number of freak wave incidences has its maximum when the run length is three, but most incidences occurred in Keelung when the run length is five.