DaleSpam said:
I am unconvinced mathematically. As far as I know a monochromatic sinusoidal plane wave is a solution, so the problem can be formulated as a classical wave equation even if there exists an alternative formulation as Laplace's equation. Furthermore dispersive and nonlinear waves are still solutions to the classical wave equation where c is a function of wavelength or amplitude.
Regarding dispersive effects, a nonlinear wave equation can be written for amplitude effects
c(u)^2 \nabla^2 u = \frac{\partial^2 u}{\partial t^2},
but how would you formulate this to include frequency-dependent dispersion? You could prescribe a phase speed for every Fourier component
c = \frac{\omega}{\kappa}
but how would you determine c as a function of x, t?
By contrast, the dispersion of deep water waves already captured by the
linearized system which arises from the Laplace equation, and as a result you get the dispersion relationship right out of it. You don't need to introduce nonlinearity to describe this behavior.
It might be also be helpful to realize the "wave" propagates not just along the surface, but also through the body of water. That is, the stuff between the surface and the bottom matters (which is why the dispersion is depth-dependent).
In the limit of shallow water, the dispersive effects go away, and you do in fact recover the 2-d wave equation. So if you are to argue from this standpoint that all water waves are just shallow water waves (and hence obey the wave equation) with some corrections, then I suppose you could make a case for that. But they would
not be small corrections in many cases..
I see no reason to trade a clear definition for a sloppy one.
I hope it's getting clearer rather than sloppier!