The SWAN model does not use physical wave equations, but statistical (empirical) relationships between cause - wind and effect - waves. Or between waves in deep water and statistics on transforming waves in shallow water, currents.... This model is averaged over the entire spectrum of emerging waves (from ripples (small period) to swells (long period)). In addition, the model works with phase-averaged waves. Actually with wave energy. What's the catch? Engineering solutions in hydraulic engineering construction, for example, are aimed at preventing the destruction of an object. Answer the question yourself which waves are dangerous: extreme height or height averaged over periods. I'll give you an example. With an average energy (height) of a wave packet (according to the SWAN model), for example, 3 meters, you have no risks. Realistically, there can only be one swell wave (long wave) in this package that will destroy your object. This is a shortcoming of statistical methods. You already know about it. I recommend using wave equations. attached presentations on this topic, unfortunately in Russian).tely in Russian).