A new third-generation spectral wave model is presented for prediction of the wave climates in offshore and coastal areas. The essential topic of the present paper is new numerical techniques for solution of the governing equations. The discretisation in the geographical space and the spectral space is based on cell-centred finite volume technique. In the geographical space, an unstructured mesh is applied. Due to the high degree of flexibility, unstructured meshes are very efficient to deal with problems of different characteristic scales. The time integration is performed using a fractional step approach, where an efficient multisequence explicit method is applied for the propagation process. The model is verified by comparison with observations for two real field cases.

From an empirical fit to measurements from RASEX together with other previously measured data sets they find the coefficients A=1.89 and B=1.59. One of the main problems in these results was the conflicting, apparent trend of decreasing Charnock parameter with inverse wave age in the RASEX data set taken alone (see also Taylor and Yelland (2001)). The problem with this kind of scaling is that the two quantities z ch and wave age, between which a functional relationship is proposed, are not independent of each other. This can lead to self-correlation problems, i.e. the functional relationship might be distorted or even determined by the common scaling variable. A theoretical analysis of the self-correlation problem has been presented by Hicks (1978) and specifically for the question of wave age dependent Charnock parameter by Smith et al. (1992). The latter group concluded that self-correlation had an influence on the results of the HEXOS data. JHVL98 generalised that this will always be the case for a

No abstract

No abstract

No abstract