Bivariate modelling of wave height and period

This paper examines the magnitude of the errors associated with the estimation of bivariate distributions of wave height and period based only upon significant wave height (H_s), mean time period (T_m) and bandwidth (e) parameters. The errors arise from the sensitivity of H_s, T_m and e to the characteristics of the measurement equipment; from the method of the calculation of the parameters; and from the difficulties of describing real wave height/period distributions using existing theoretical models. The authors conclude that the most reliable estimates of sea-state parameters are those based on the raw data rather than on those derived from spectral estimates. However, ideally the raw wave data itself should be stored rather than just the summary statistics. Additionally, a purely empirical model which has not previously been used to describe individual wave height/period combinations is presented as a possible aid to the estimation of their joint bivariate distribution.