Existence of non-trivial non-concentrated compactly supported stationary solutions of the 2D Euler equation with finite energy (Q6947005)

This paper addresses some really interesting questions concerning rigidity/flexibility properties of stationary solutions of two-dimensional incompressible Euler equations in vorticity form\N\[\N\begin{cases} \N\partial_t\omega+u\cdot\nabla\omega=0 &\text{in }\mathbb{R}^2\times\mathbb{R}_+\,,\\\Nu(\cdot,t)=-\nabla^{\perp}(-\Delta)^{-1}\omega(\cdot,t) &\text{in }\mathbb{R}^2\,,\\\N\omega(\cdot,0)=\omega_0 &\text{in }\mathbb{R}^2\,, \N\end{cases}\N\]\Nwhere \(\nabla^\perp:=(-\partial_{x_2},\partial_{x_1})\), \(u=u(x,t)=(u_1,u_2)\) is the velocity field and \(\omega=\partial_{x_1}u_2-\partial_{x_2}u_1\) is the vorticity. \N\NStationary solutions of Euler equations are a significant building block in many research directions aimed at better understanding some key mechanisms in fluid dynamics. They come into play in order to understand turbulence, both in 2D and 3D, where turbulent flows are realized as superposition of Beltrami flows, namely particular stationary solutions of 3D Euler equations whose curl is proportional to themselves. In numerical simulation of 2D Navier-Stokes equations with small viscosity, the Navier Stokes solutions are approximated as a superposition of steady eddies of constant vorticity that solve the 2D Euler equations. In the context of convex integration, Beltrami flows and Mikado flows are stationary solutions of the Euler equations that are used within an iteration scheme in order to build weak solutions, by generating fast oscillating perturbations. \N\NThe authors mainly work in the \textit{patch} setting, where the vorticity \(\omega(\cdot,t)=\sum\Theta_i 1_{D_i(t)}\) is a sum of indicator functions of bounded sets \(D_i(t)\) moving with the fluid and \(\Theta_i\in\mathbb{R}\). Nevertheless some of their results translate into the smooth setting too (where \(\omega(\cdot,t)\) is smooth and compactly supported in space). \N\NIn [Duke Math. J. 170, No. 13, 2957--3038 (2021; Zbl 1478.35013)], the authors of the present paper together with \textit{Y. Yao} were able to show that stationary solutions of the 2D Euler equations (in both the patch and smooth settings), for which \(\omega\ge 0\) and is compactly supported, have to be radial. \N\NIn this paper, the necessity of the assumption \(\omega\ge 0\) is addressed. The main results are provided by Theorem 3.1, pertaining the existence of stationary solutions with infinite kinetic energy, and Theorem 4.1, pertaining the existence of stationary solutions with finite kinetic energy. \N\NThe proof of such results employs bifurcation theory, trying to find a perturbation of trivial solutions (radial vorticity). A very natural procedure is to frame it by applying a Crandall-Rabinowitz bifurcation theorem from [\textit{M. G. Crandall} and \textit{P. H. Rabinowitz}, J. Funct. Anal., 8, 321--340 (1971; Zbl 0219.46015)] to seek non-trivial zeros of an appropriate nonlinear functional in a suitable functional setting. \N\NAs a consequence of Theorem 3.1, the authors deduce the existence of non-radial, sign-changing stationary vortex patch solutions with analytic boundary to the \(2D\) Euler equations with infinite kinetic energy, see Theorem 1.2 (Corollary 3.2) and Remark 1.3. \N\NConcerning the finite kinetic energy case, the second main Theorem 4.1 immediately implies the existence of non-radial stationary vortex patches with finite kinetic energy, see Theorem 1.4 (Corollary 4.5, Theorem 4.27). The proof relies on several clever observations and the application of the Nash-Moser iteration procedure. In particular, a key new observation is that finite energy, stationary solutions with simply-connected vorticity have compactly supported velocity. \N\NThe article starts with an introduction, where the authors set up the problem they are going to deal with. In particular, in Subsection 1.1, they summarize some of the history of stationary solutions, mostly focusing on the rigidity (only trivial solutions exist) vs flexibility (i.e. existence of non-trivial solutions) and, in Subsection 1.2, they describe the structure of the proofs of the main results. Chapter 2 is devoted to introduce the proper functional framework where the problem will be then solved. Chapter 3 contains the proof of the first main Theorem 3.1. As a by-product, the existence Theorem 1.2, in the easiest case where the finite energy hypothesis is dropped, immediately follows. Chapter 4 gives the proof of the second main Theorem 4.1 and (as an immediate consequence) the one of Theorem 1.4 on the existence of non-radial stationary vortex patches with finite kinetic energy, relying upon the Nash-Moser scheme. A final Appendix collects some technical results used throughout the text, as well as basic bifurcation theory and some integrals needed for the spectral study.