WorksMathematics
Optimal local gradient estimate for gravity water waves and its application to the Stokes conjecture
Abstract
In 1880, G. Stokes [Mathematical and Physical Papers, Vol. I] studied a two-dimensional gravity water wave and conjectured that the wave crest forms a symmetric corner of angle 2pi/3 at the stagnation point. The conjecture was first proved by Amick, Fraenkel, and Toland [Acta Math. 148 (1982)] and, independently, by Plotnikov [Dinamika Sploshn. Sredy 57 (1982)] via the Nekrasov integral equation. Their proofs rely on structural assumptions, including the isolation of stagnation points and the symmetry and monotonicity of the wave profile. In 2011, Varvaruca and Weiss [Acta Math. 206 (2011)] introduced a novel geometric approach and gave a new proof of the conjecture without relying on the structural assumptions mentioned above, under a local gradient bound for the two-dimensional stream function. In this paper, we establish an optimal local gradient estimate for general variational solutions of the free boundary problem. Our approach adapts the Hopf differential method of De Philippis, Hirsch, and Nahon [arXiv:2609.12180v2] to the gravitational setting, and combines it with their abstract approximation theorem to turn the holomorphic approximation of the Hopf differential into the local gradient bounds. An important new observation is that the weighted Cauchy-Beurling correction, combined with vanishing on the lower half plane, supplies the additional height decay needed to obtain the local gradient bounds. As a consequence, we remove the a priori growth hypothesis from the geometric analysis in [Acta Math. 206 (2011)]. For general variational solutions, we establish the Stokes corner or full-density blow-up alternative. Under the stated topological and interface hypotheses, we obtain the Stokes corner and two free boundary branches that are C^1 up to the stagnation point.