The objective of this study is to justify the use of the Hilbert Transform, i.e., the Orbital Method, on analysis of wave statistics obtained from field data and simulated data. The results were also compared to those analyzed by the zero-up-crossing method and the modified zero-up-crossing method (Thompsom, 1984) it showed that wave statistical characteristics calculated with the orbital discretion criterion are more consistent in different spectral width conditions than those with the other two methods. The results may suggest that the orbital method is a more proper wave discretion criterion than the (modified) zero-up-crossing method.
For scientific research or engineering design, wave statistics is very important. However, in consideration of the wave randomness, the proper way to identify individual wave in a time series for deriving wave statistics is still in question. The most used method is the Zero- Up Crossing criterion (ZUC). It is a simple method, but some noise and spurious waves (i.e. false waves) will be treated as real waves by the ZUC and make the wave statistics deviate from the "true" value. To avoid the influence of false waves, many methods have been proposed. For example, Thompsom(1984) proposed that any wave height lower than one standard deviation of wave time series should be eliminated. There are several more methods introduced to eliminate false waves, but no one is the accepted standard rule. Wave characteristics can also be expressed as spectrum parameters by spectrum analysis in the frequency domain. Although the wave properties are represented in different ways, they could be related to each other based on theoretical considerations. Gim6nez et al. (1994) and Yang et al. (1995) used the orbital method as a wave discretion criterion to calculate the statistics of simulated wave data.
