Does the independent validation of different models using significantly different physics indicate that wave prediction models are insensitive generally to the assumed physics?
[We find that this is not the case. For example, it is impossible to tune a first-generation DP model to agree with a second-generation CH or CD model for both fetch-limited and duration-limited wave growth, as also pointed out by Resio and Vincent (1979).]
What are the differences in practice between the alternative spectral representations adopted in the second-generation models of type CH (hybrid) and CD (discrete)?
[In the present stage of development we find little real distinction between the two. To exploit fully the potential advantage of the more detailed, but more time-consuming, CD models, generalised parameterizations of the nonlinear transfer are needed.]
Can critical elements in the model physics be identified which are responsible for strong divergences between the predictions of different models?
[The case studies were designed to reveal these. Major sources of uncertainty were found to be associated with the response to changing wind directions and the modeling of the transition region between windsea and swell. However, surprisingly large discrepancies were also found in the basic fetch- and duration-limited growth curves for a uniform wind field.]
Having identified the most critical weaknesses of present wave models, what should be done to improve our understanding and modeling of the wave spectral energy balance?
[A number of recommendations are given in the individual case studies and are collected in Chapter 13. Suggestions are made for the analysis of selected sets of existing data, the collection of new observations, and the development of improved numerical parameterizations.]