On the Cauchy problem for gravity water waves (Q464827)

The authors consider the Cauchy problem for the system of water waves without surface tension in an arbitrary dimension. They study the regularity threshold for the well-posedness of such an initial-boundary value problem. They show that the threshold is the Lipschitz condition in terms of the free water surface, despite of some opposite conjectures. One denotes the time-space variables by \((t,x,y)\in{\mathbb R}\times{\mathbb R}^d \times{\mathbb R}\), where \(x\) (resp. \(y\)) corresponds to the horizontal (resp. vertical) spatial direction. There is a fixed container \({\mathcal O}\subset{\mathbb R}^d \times{\mathbb R}\) and the fluid domain \(\Omega\) is given by \(\Omega =\{(t,x,y) \in{\mathbb R}\times{\mathcal O};\;y<\eta (t,x)\}\). Here \(y=\eta (t,x)\) is the free boundary of the fluid moving in the container. The authors assume that the fluid is inviscid, incompressible, and irrotational. The unknown functions are \(\eta (t,x)\) and \(\psi (t,x)=\phi (t,x,\eta (t,x))\), where \(\phi (t,x,y)\) is the potential of the velocity. These functions must satisfy the Dirichlet-Neumann equation, and one studies the corresponding initial-boundary value problem. The result is as follows: Let \(d\geq 1,s>(2+d)/2\), and consider \(\eta (0,x),\psi (0,x)\in H^{(2s+1)/2} ({\mathbb R}^d )\). One assumes that \({\mathcal O}\) contains a fixed strip arround the free surface, and that the Taylor coefficient \(-(\partial_y P)(t,x,\eta ((t,x))\) is positive. Then there exists \(T>0\) such that the Cauchy problem has a unique solution \((\eta ,\psi)\in C^0 ([0, T]; H^{(2s+1)/2} ({\mathbb R}^d )\times H^{(2s+1)/2} ({\mathbb R}^d ))\). Note that in view of the Sobolev embedding theorem, the initial surface turns out to be only of \(C^{(3+\varepsilon)/2}\) class for some \(\varepsilon >0\), and consequently may have an unbounded curvature. The analysis is based on a para-differential study of the water waves system. In the present case the domain belongs merely to \(C^{3/2}\), and a micro-local description of the Dirichlet-Neumann equation is necessary.