Generic small-scale creation in the two-dimensional Euler equation

1. Problem Statement and Context

The paper addresses the Cauchy problem for the two-dimensional incompressible Euler equations. While it is a classical result that these equations are globally well-posed for smooth initial data (meaning solutions exist and remain smooth for all time), a long-standing conjecture by Yudovich (1974) posits that for a "substantial set" of initial data, the vorticity gradients grow without bound over infinite time, leading to a loss of regularity (turbulence).

Specifically, the authors investigate whether generic smooth solutions exhibit small-scale creation, characterized by the unbounded growth of high-order Sobolev norms ($H^s$) or Hölder norms ($C^k$) as $t \to \infty$. Previous results on unbounded growth were limited to:

• Singular initial data (e.g., vortex patches with corners).
• Domains with boundaries where growth is driven by boundary effects (stagnation points).
• Specific symmetry classes.

The authors aim to prove that this growth occurs generically (in the sense of Baire category) for smooth initial data in a domain without boundaries (the torus $\mathbb{T}^2$) and without symmetry constraints, driven solely by internal nonlinear dynamics.

2. Main Results

The paper establishes two primary theorems:

• Theorem 1.1 (Torus Case): For any real number $s > 4$, the set of initial vorticities $\omega_0 \in H^s_0(\mathbb{T}^2)$ such that the corresponding solution satisfies
  $$\limsup_{t \to +\infty} \|\omega(t, \cdot)\|_{H^s(\mathbb{T}^2)} = +\infty$$
  is a dense $G_\delta$ set (a countable intersection of open dense sets) in $H^s_0(\mathbb{T}^2)$. This confirms that generic smooth solutions lose regularity in infinite time.

• Corollary 1.2 (Smoothness): The set of initial data in $C^\infty_0(\mathbb{T}^2)$ for which there exists an integer $k$ such that $\limsup_{t \to \infty} \|\omega(t)\|_{C^k} = +\infty$ is also a dense $G_\delta$ set.

• Theorem 1.3 (Whole Space Case): For solutions on $\mathbb{R}^2$ with zero mean and finite $L^1$ norm, the set of initial data leading to unbounded $H^s$ growth (for $s > 1$) is a dense $G_\delta$ set. The growth rate is bounded below by $t^s$.

3. Methodology and Proof Strategy

The proof combines paradifferential calculus, Lagrangian analysis, and topological arguments (Baire category and fixed point theorems). The strategy is divided into five key steps:

Step 1: Paralinearization of the Lagrangian Flow

Following Shnirelman's framework, the authors introduce a vector field $X(t) := T(D\Phi_t)^{-1}\Phi_t$, where $\Phi_t$ is the flow map and $T$ denotes the paraproduct.

• They derive a closed evolution equation for $X$:
  $$\partial_t X + \nabla^\perp Q T\nabla\omega_0 \cdot X = \nabla^\perp Q \omega_0 - R$$
  where $Q$ is a parametrix for the paradifferential Laplace–Beltrami operator and $R$ is a smoothing remainder.
• Key Insight: This equation reveals that the growth of $X$ is driven linearly by the initial vorticity gradient, suggesting that small-scale creation is inevitable for finite regularity data.

Step 2: Microlocal Analysis of Hölderian Cusps

To quantify the growth, the authors analyze the interaction between the flow and specific singular profiles.

• They define a dense set of "unstable" initial data characterized by a slowly decaying Fourier tail and a non-degenerate global minimum.
• They construct test functions based on Hölderian cusps of the form $H(x) \sim |x|^{2\alpha}$.
• Using symbol freezing techniques, they show that the paradifferential operator acts on these cusps essentially as a constant-coefficient multiplier, allowing for sharp lower bounds on the growth of the inner product $\langle \text{curl } X, P_\epsilon (g-m)^\alpha \rangle$.

Step 3: A Dichotomy (Proposition 8.1)

The authors establish a critical alternative for any initial data in the unstable class:

1. Either the vorticity $\omega$ blows up in $H^s$;
2. Or the inverse flow map $\phi_t = \Phi_t^{-1}$ blows up in the Lipschitz norm ($L^\infty$).

• Proof Technique: They use a Schauder fixed point argument. Assuming stability (boundedness of both vorticity and inverse flow), they construct a perturbation map that forces a contradiction regarding the spectral support of the perturbation.

Step 4: Lagrangian Stretching Implies Instability (Proposition 9.1)

If the inverse flow blows up in the Lipschitz norm (Lagrangian stretching), the authors show this drives Eulerian growth of the vorticity.

• Mechanism: Adapting an argument by Koch, they construct a high-frequency wave packet perturbation.
• Scaling: By choosing a perturbation with amplitude $\delta^{s-1}$ and frequency $1/\delta$, and exploiting the stretching factor $\Lambda$ of the flow, the $s$-th derivative of the transported perturbation gains a factor of $\Lambda^s$. This results in norm inflation, contradicting the assumption of boundedness.

Step 5: Baire Category Argument

Finally, they combine the previous steps to prove genericity.

• They define sets $U_N = \{ \omega_0 : \sup_{t \ge 0} \|\omega(t)\|_{H^s} > N \}$.
• Using the dichotomy and the instability mechanism, they prove that $U_N$ is dense.
• Using the continuity of the solution map, they prove $U_N$ is open.
• By the Baire Category Theorem, the intersection $\bigcap U_N$ is a dense $G_\delta$ set, proving that generic initial data lead to unbounded growth.

4. Key Technical Innovations

• Paracomposition Operators: The authors utilize a streamlined version of Alinhac's paracomposition operators to handle the composition of functions with diffeomorphisms ($f \circ \Phi$), allowing for a precise paralinearization of the Lagrangian dynamics.
• Microlocal Cusp Analysis: Instead of relying on boundary effects, they use the specific spectral properties of Hölderian cusps to extract growth information from the internal nonlinear dynamics.
• Dichotomy Principle: The separation of the problem into "Vorticity Blow-up" vs. "Flow Blow-up" allows them to bypass the difficulty of directly proving vorticity blow-up for all cases, instead proving that if the flow stretches enough, the vorticity must blow up.
• Whole Space Scaling: For $\mathbb{R}^2$, they leverage scaling symmetries (Shankar's identity) where spatial infinity acts as a stagnation point, simplifying the analysis compared to the periodic case.

5. Significance

• Confirmation of Yudovich's Conjecture: This work provides the first rigorous confirmation of generic turbulent behavior for the 2D Euler equation driven solely by internal nonlinear dynamics, independent of boundaries or symmetries.
• Regularity Loss: It demonstrates that even for arbitrarily smooth initial data, the solution can lose regularity in infinite time, challenging the intuition that 2D Euler is "tamer" than 3D Euler.
• Methodological Advancement: The combination of paradifferential calculus with topological fixed-point arguments offers a new toolkit for studying instability in quasilinear PDEs, potentially applicable to other active scalar equations (like SQG) and higher dimensions.

In summary, the paper proves that the "tame" behavior of 2D Euler is the exception rather than the rule: for a generic set of smooth initial conditions, the fluid develops increasingly fine scales, causing the vorticity gradients to grow indefinitely.