Random divergence-free drifts and the Onsager-Richardson threshold

This paper proves the absence of anomalous dissipation for passive scalars driven by random divergence-free autonomous vector fields in Hölder classes with regularity $\alpha > 1/3$ by utilizing dimension-theoretic arguments rather than commutator estimates, thereby establishing that anomalous regularization does not occur for this class of fields.

The Gist

Here is an explanation of the paper "Random Divergence-Free Drifts and the Onsager-Richardson Threshold" using simple language, analogies, and metaphors.

The Big Picture: The "Perfect Stir" vs. The "Messy Stir"

Imagine you are making a cup of coffee. You drop a cube of sugar into hot water.

• Scenario A (No Stirring): The sugar dissolves very slowly, just by the heat of the water (diffusion).
• Scenario B (Stirring): You take a spoon and swirl the water. The sugar dissolves almost instantly.

In physics, this "stirring" is called advection. The water (the fluid) moves the sugar (the passive scalar) around.

Now, imagine the water is turbulent. It's not just a smooth swirl; it's a chaotic, chaotic mess of tiny whirlpools and eddies. Physicists have long believed that in this chaotic state, the sugar dissolves so efficiently that even if you remove the "heat" (the diffusion) entirely, the sugar would still disappear instantly. This is called Anomalous Dissipation.

It's as if the chaos of the water itself acts like a super-powerful blender, destroying the sugar without needing any heat at all.

The Mystery: How Rough Can the Stirring Be?

For decades, mathematicians have been asking: How chaotic does the water need to be for this "magic blender" effect to happen?

There is a famous rule in fluid dynamics called Onsager's Conjecture (named after Lars Onsager). It says that for a fluid to behave in this wild, energy-losing way, the flow has to be "rough" enough. Specifically, the smoothness of the flow is measured by a number called $\alpha$ (alpha).

• If the flow is very smooth ($\alpha > 1/3$), it behaves nicely. Energy is conserved.
• If the flow is very rough ($\alpha < 1/3$), it behaves wildly. Energy can vanish mysteriously.

The number $1/3$ is the magic threshold. It's the border between "orderly" and "chaotic."

What This Paper Does

The authors of this paper (Boutros, De Lellis, and Mayboroda) wanted to test this rule for the "sugar in coffee" problem (passive scalar transport).

They asked: "If we have a random, chaotic flow that is slightly rough (but still smoother than the $1/3$ limit), will the sugar still dissolve instantly even without heat?"

The Answer: No.

They proved that if the flow is "random" but still has a smoothness level greater than $1/3$, the "magic blender" effect does not happen. The sugar will not dissolve instantly if you turn off the heat. The system behaves normally.

The "Magic" of the Proof: A Geometric Puzzle

How did they prove this? They didn't use the usual heavy math tools (like complex energy calculations). Instead, they used a clever geometric trick involving shapes and dimensions.

Here is the analogy:

1. The Stream Function (The Map): Imagine the chaotic flow is drawn on a map. This map has "hills" and "valleys." The water flows along the contours of these hills.
2. The Critical Points (The Peaks): In any map, there are peaks and valleys where the slope is flat. These are "critical points."
3. The Morse-Sard Property: A famous math rule says that for a smooth enough map, the "heights" of these peaks (the critical values) are so rare that if you picked a random height, you would almost never land on a peak. The set of these special heights has zero size (like a single dot on a page has zero area).

The Authors' Insight:
They showed that if the flow is random and smooth enough ($\alpha > 1/3$), the "map" of the flow is so well-behaved that the "peaks" are mathematically negligible.

• Because the peaks are negligible, the water flow follows a very predictable path (called a Lagrangian flow).
• If the water follows a predictable path, it cannot "mix" the sugar in that magical, instant way.
• Therefore, the "Anomalous Dissipation" (the magic blender) cannot exist.

Why is the Number $1/3$ Special Here?

You might wonder: "Why is $1/3$the cutoff? Why not$1/2$or$1/4$?"

In the paper, they explain that this number comes from a battle between two things:

1. How rough the flow is (The "jaggedness" of the map).
2. How many dimensions the "peaks" take up.

If the flow is too rough (below $1/3$), the "peaks" become so numerous and spread out that they cover a significant area, breaking the rules of the game. But if the flow is just a tiny bit smoother (above$1/3$), the "peaks" shrink down to nothingness, and the system becomes stable again.

It is a surprising coincidence that the same number ($1/3$) that limits energy conservation in swirling fluids (Onsager's original theory) also limits how well random flows can mix things up.

The Takeaway for Everyday Life

Think of this paper as a safety check for a chaotic system.

• Old Belief: "If the world is chaotic enough, it will mix everything perfectly, even without help."
• New Finding: "Actually, if the chaos is too smooth (even just a little bit), it loses its super-mixing power. It needs to be truly, violently rough to break the rules."

The authors proved that for random, autonomous (time-independent) flows, you cannot get "anomalous" mixing unless the flow is extremely rough. If the flow is even slightly smoother than the $1/3$ threshold, the laws of physics hold steady, and the sugar won't vanish unless you add heat.

In short: Chaos has a limit. If it's not rough enough, it can't break the rules of thermodynamics.

Technical Summary

Here is a detailed technical summary of the paper "Random Divergence-Free Drifts and the Onsager-Richardson Threshold" by Boutros, De Lellis, and Mayboroda.

1. Problem Statement

The paper addresses the phenomenon of anomalous dissipation in the context of passive scalar transport driven by random, autonomous, divergence-free vector fields.

• Context: In the vanishing viscosity limit ($\varepsilon \to 0$), the advection-diffusion equation $\partial_t \theta_\varepsilon + (v \cdot \nabla)\theta_\varepsilon = \varepsilon \Delta \theta_\varepsilon$ is expected to exhibit non-vanishing energy dissipation ($\limsup_{\varepsilon \to 0} \varepsilon \int |\nabla \theta_\varepsilon|^2 > 0$) in turbulent regimes. This is linked to the "zeroth law of turbulence" and the Obukhov-Corrsin-Yaglom (OCY) law, which predicts a relationship between the Hölder regularity $\alpha$ of the velocity field $v$ and the effective regularity $\beta$ of the scalar: $\alpha + 2\beta = 1$.
• The Gap: While Onsager's conjecture for the Euler equations establishes $1/3$as the sharp threshold for energy conservation (rigidity vs. flexibility), the threshold for anomalous dissipation in passive scalar transport with **random** drifts was not rigorously established, particularly for regularities$\alpha > 1/3$.
• Goal: To determine if random divergence-free vector fields with Hölder regularity $\alpha > 1/3$ can induce anomalous dissipation or anomalous regularization (effective smoothing) without assuming regularity on the scalar itself.

2. Methodology

The authors employ a novel combination of probabilistic arguments and geometric measure theory, avoiding traditional commutator estimates used in deterministic fluid dynamics.

A. The DiPerna-Lions Renormalization Property

The core strategy is to prove that the random vector field $v$ satisfies the DiPerna-Lions renormalization property almost surely.

• If a vector field has this property, bounded weak solutions to the transport equation are unique, and the $L^p$ norms of the solution are conserved.
• Crucially, this property implies that the vanishing viscosity limit converges strongly to the unique inviscid solution, thereby excluding anomalous dissipation.

B. Reduction to the Morse-Sard Property

The authors rely on a deep result by Alberti, Bianchini, and Crippa [2], which states that if the stream function (or Hamiltonian) associated with the vector field satisfies the Morse-Sard property (i.e., the set of its critical values has zero Lebesgue measure), then the vector field possesses the renormalization property.

C. Dimension-Theoretic Arguments

The proof is split into two main steps to establish the Morse-Sard property for the random field:

1. Probabilistic Step (Lemma 3.1): Using the assumed non-degeneracy of the probability measure (bounded density of the pointwise distribution), the authors prove that the Hausdorff dimension of the critical set $Z = \{x : \text{rank}(D\phi(x)) \le d-2\}$ of the stream function $\phi$ is bounded almost surely by:
   $$\dim_H(Z) \le d - 2\alpha$$
2. Deterministic Step (Lemma 3.2): They prove a covering argument showing that if $\phi \in C^{1,\alpha}$ and $\dim_H(Z) \le d - 2\alpha$, then the image of the critical set $\phi(Z)$ has zero Lebesgue measure (Morse-Sard property) if and only if:
   $$1 + \alpha > 2 - 2\alpha \implies \alpha > 1/3$$
   This inequality arises from balancing the Hölder regularity of the map against the dimension of the critical set.

3. Key Contributions and Results

Main Theorems

The paper establishes that for random autonomous vector fields with Hölder regularity $\alpha > 1/3$, anomalous dissipation does not occur almost surely.

• Theorem 1.2 (2D Case): For a probability measure on divergence-free, mean-zero vector fields on $\mathbb{T}^2$ with:

  1. $v \in C^\alpha(\mathbb{T}^2)$ with $\alpha > 1/3$ a.s.
  2. A mild non-degeneracy condition (bounded density of $v(x)$).
     Result: $v$ satisfies the DiPerna-Lions renormalization property a.s. No dissipation anomalies occur.

• Theorem 1.3 (3D Case): For a probability measure on maps $\phi = (\phi_1, \phi_2): \mathbb{T}^3 \to \mathbb{R}^2$ with $\phi \in C^{1,\alpha}$ ($\alpha > 1/3$) and a similar non-degeneracy condition on the derivative $D\phi$.
  Result: The vector field $v = \nabla \phi_1 \times \nabla \phi_2$ satisfies the renormalization property a.s.

• Theorem 2.2 (General Dimension): A general formulation for $d$ dimensions using the Hodge star operator and $(d-1)$-forms, unifying the 2D and 3D cases.

Technical Nuances

• No Regularity on Scalar: Unlike previous results, the passive scalar $\theta$ is not assumed to have any regularity beyond being bounded initially. The results are "Onsager-supercritical."
• Mild Non-degeneracy: The condition on the probability measure (bounded density) is extremely mild and satisfied by standard constructions (e.g., random Fourier series with non-degenerate coefficients).
• Autonomous Fields: The results apply to time-independent (autonomous) fields, distinguishing them from stochastic PDE results (like the Kraichnan model) which rely on white noise in time.

4. Significance and Implications

• **Sharpness of the $1/3$Threshold:** The paper demonstrates that the exponent$1/3$ is a sharp threshold for passive scalar transport in random autonomous fields, mirroring Onsager's conjecture for the Euler equations.
  • Surprise: This is surprising because the mechanism is purely dimension-theoretic (geometric), whereas Onsager's original result relied on nonlinear energy flux arguments.
  • Contradiction to Heuristics: The result contradicts the "Obukhov-Corrsin-Yaglom" law ($\alpha + 2\beta = 1$) which suggests anomalous regularization for $\alpha > 1/3$. The authors show that for $\alpha > 1/3$, the scalar does not smooth out; instead, it preserves the initial data's distribution (e.g., discontinuities remain), preventing the "anomalous regularization" predicted by dimensional analysis.
• Geometric Mechanism: The work highlights that the rigidity of the transport equation in this regime is driven by the geometric structure of the stream function's critical values (Morse-Sard property) rather than statistical homogeneity or energy cascades.
• Probabilistic Rigidity: It establishes that "generic" random vector fields with sufficient regularity ($\alpha > 1/3$) are too rigid to support the chaotic mixing required for anomalous dissipation.

5. Conclusion

The authors prove that for random, autonomous, divergence-free vector fields with Hölder regularity $\alpha > 1/3$, the passive scalar transport equation possesses the DiPerna-Lions renormalization property almost surely. Consequently, the vanishing viscosity limit yields strong convergence to the unique inviscid solution, and anomalous dissipation is absent. This result identifies $1/3$ as the critical threshold for the onset of anomalous dissipation in this probabilistic setting, driven by a geometric mechanism involving the Hausdorff dimension of critical sets.