On Local Regularity of Distributional Solutions to the Navier--Stokes Equations

This paper establishes a sharp regularity result in spatial variables for distributional solutions to the Navier-Stokes equations that satisfy the Prodi-Serrin condition, without requiring the solutions to belong to the local Leray-Hopf class.

The Gist

Imagine you are trying to predict the weather. You have a massive, chaotic system of wind, rain, and pressure (the Navier-Stokes equations). Mathematicians have been trying to solve these equations for nearly a century to understand how fluids move.

For a long time, the "gold standard" for a solution was the Leray-Hopf solution. Think of this as a "Gold Card" holder. To get this card, a solution had to prove it had a finite amount of "energy" (like a car with a full tank of gas). If you had the Gold Card, you were guaranteed to be smooth, predictable, and unique.

However, there was a nagging question: Is the Gold Card actually required? Or, could a solution be smooth and predictable even if it didn't start with a full tank of gas (finite energy)?

This paper by Giovanni P. Galdi answers that question with a resounding "Yes, but with a catch."

Here is the breakdown using simple analogies:

1. The Problem: The "Ghost" Solutions

In the world of fluid dynamics, there are "ghost" solutions. These are mathematical tricks where the fluid looks like it's moving, but it's actually just a static, invisible force field (like a potential) that doesn't really flow.

• The Old Rule: To prove a fluid is smooth and real, you had to assume it was a "Gold Card" holder (finite energy) everywhere.
• The Catch: If you didn't assume the Gold Card, those "ghost" solutions could sneak in and ruin the proof. They look smooth in space but behave wildly in time.

2. The New Discovery: The "Split Personality"

Galdi's breakthrough is realizing that any fluid solution can be split into two distinct personalities (a mathematical trick called the Helmholtz-Weyl decomposition):

1. The Swirly Part ($u_\sigma$): This is the actual fluid moving, swirling, and eddying. This is the part that does the "work."
2. The Pressure Part ($\nabla \pi$): This is the invisible force field pushing the fluid around. It's like the background music of the system.

The Old Assumption: We thought we needed to check the entire system (both parts) to have finite energy.
Galdi's Insight: We only need to check the Swirly Part.

3. The "Prodi-Serrin" Condition: The "Smoothness Test"

There is a famous test called the Prodi-Serrin condition. Think of it as a "Smoothness Test" for the Swirly Part.

• If the Swirly Part passes this test (meaning it isn't too wild or jagged in space and time), it is guaranteed to be smooth and well-behaved.
• The Big Question: Does the Swirly Part need to be a "Gold Card" holder (finite energy) to pass this test?

Galdi's Answer: No!
You don't need the whole system to have finite energy. You just need the Swirly Part to pass the Smoothness Test.

4. The Catch: The "Background Music"

So, if the Swirly Part is smooth, what about the Pressure Part?

• Galdi shows that the Pressure Part (the background music) needs to be calm in time. It doesn't need to be finite in energy, but it can't be screaming or changing instantly. It just needs to be "bounded" (not going crazy).
• Analogy: Imagine a dance floor. The dancers (the Swirly Part) are doing complex, smooth moves. The DJ (the Pressure Part) doesn't need to be a professional dancer, but they can't suddenly blast a siren or change the tempo every millisecond. As long as the DJ keeps a steady rhythm, the dancers will remain smooth.

5. Why This Matters

Before this paper, mathematicians were stuck thinking, "We can't prove the fluid is smooth unless we assume it has finite energy everywhere."

• The Old View: "If you don't have a full tank of gas, you can't drive smoothly."
• Galdi's View: "Actually, as long as the engine (the Swirly Part) is tuned right and the fuel gauge (the Pressure) isn't broken, you can drive smoothly even if the tank isn't full."

Summary in One Sentence

Galdi proves that to guarantee a fluid flow is smooth and predictable, you don't need the whole system to be "perfectly contained" (finite energy); you just need the moving parts to be well-behaved and the pressure to stay calm.

This removes a huge, unnecessary assumption from the theory, making our understanding of fluid dynamics more flexible and closer to reality.

Technical Summary

Here is a detailed technical summary of the paper "On Local Regularity of Distributional Solutions to the Navier–Stokes Equations" by Giovanni P. Galdi.

1. Problem Statement and Context

The Navier-Stokes equations (NSE) are central to fluid dynamics, yet fundamental questions regarding the uniqueness, regularity, and energy balance of their solutions remain open. Historically, the Leray-Hopf class of solutions ($L^\infty(0, T; L^2) \cap L^2(0, T; W^{1,2})$) has been the standard framework. These solutions are known to exist globally but their regularity is conditional.

The Prodi-Serrin Condition:
It is a classical result that if a Leray-Hopf solution satisfies the Prodi-Serrin condition:
$$u \in L^{q'}(0, T; L^q(\mathbb{R}^3)), \quad \frac{2}{q'} + \frac{3}{q} = 1, \quad q > 3$$
then the solution is smooth ($C^\infty$) and unique.

The Gap:
A critical open question is whether the assumption that the solution belongs to the Leray-Hopf class (finite energy) is a necessary prerequisite for these regularity results. Specifically, can one prove local regularity for a distributional solution (which may not have finite energy globally or locally in the $L^2$ sense) that satisfies the Prodi-Serrin condition, without assuming it belongs to the local Leray-Hopf class?

Previous attempts to answer this required excluding "potential-like" solutions (solutions of the form $u = a(t)\nabla \psi$ where $\psi$ is harmonic) by imposing strong time-boundedness ($L^\infty$ in time) or assuming the solution solves an initial value problem.

2. Methodology

Galdi addresses this by analyzing the Helmholtz-Weyl decomposition of the velocity field $u$ in a space-time cylinder $O = B \times (t_1, t_2)$:
$$u = u_\sigma + \nabla \pi$$
where:

• $u_\sigma$ is the solenoidal (divergence-free) component.
• $\nabla \pi$ is the gradient (potential) component.

The paper relies on three main technical pillars:

1. Decomposition Strategy: Instead of treating $u$ as a whole, the author separates the regularity analysis into the solenoidal part ($u_\sigma$) and the potential part ($\pi$).
2. Stokes System Regularity: The author establishes and utilizes specific regularity lemmas for distributional solutions to homogeneous and non-homogeneous Stokes systems.
   • Lemma 2.2: Proves that if a distributional solution to the homogeneous Stokes system satisfies certain integrability, its solenoidal part is smooth ($C^\infty$) in the interior.
   • Lemma 2.3: Provides estimates for solutions to the non-homogeneous Stokes system driven by a tensor field $F$, establishing mapping properties between Lebesgue spaces (specifically relating $L^{q'/2, q/2}$ to $L^{r', r}$).
3. Iterative Bootstrapping:
   • Case $q \in [4, 6]$: The author uses energy estimates and the properties of the Stokes system to show that if $u_\sigma$ satisfies the Prodi-Serrin condition, it automatically gains the necessary $L^\infty(L^2)$ and $L^2(W^{1,2})$ regularity locally.
   • Case $q \in (3, 4) \cup (6, \infty)$: The author employs an iterative argument (inspired by Seregin and Šverák [3]) to "bootstrap" the integrability. By using the non-homogeneous Stokes estimates, the author shows that if $u$ satisfies the condition for a specific $q$, it implies a condition for a "better" $q$ (closer to the range $[4, 6]$). Repeating this process finitely many times allows the result to be extended to all $q > 3$.

3. Key Contributions and Results

Main Theorem (Theorem 3.1):
Let $u \in L^2(O)$ be a distributional solution to the NSE (satisfying the weak formulation). Let $u = u_\sigma + \nabla \pi$ be its Helmholtz-Weyl decomposition. If:

1. The solenoidal part satisfies the Prodi-Serrin condition: $u_\sigma \in L^{q', q}(O)$ with $\frac{2}{q'} + \frac{3}{q} = 1$ and $q > 3$.
2. The potential part satisfies a mild time-integrability condition: $\pi \in L^{\infty, 1}(O)$ (bounded in time, integrable in space).

Then:

• $u$ satisfies the local Leray-Hopf regularity conditions on any compact subdomain $O' \subset O$:
  $$u \in L^{\infty, 2}(O'), \quad \nabla u \in L^{2, 2}(O')$$
• Consequently, $u$ is smooth ($C^\infty$) in the spatial variables for almost every time $t$, and all spatial derivatives are bounded.

Crucial Distinctions from Previous Work:

• No Global Energy Assumption: The solution $u$ does not need to belong to the Leray-Hopf class ($L^\infty(L^2) \cap L^2(W^{1,2})$) a priori. The regularity is derived from the Prodi-Serrin condition on the solenoidal part.
• Handling Potential Solutions: The paper explicitly addresses the "potential-like" counterexamples ($u = a(t)\nabla \psi$). It demonstrates that the regularity failure in such cases stems entirely from the gradient part $\nabla \pi$. By imposing the condition $\pi \in L^{\infty, 1}$, these singularities are ruled out.
• Sharpness: The assumption that $\pi$ must be bounded in time ($L^\infty$ in time) is shown to be necessary. Without it, the potential-like solutions would violate regularity.

4. Significance and Implications

1. Redundancy of the Leray-Hopf Assumption: The paper proves that the requirement for a solution to be in the Leray-Hopf class is redundant for establishing local regularity under the Prodi-Serrin condition. The Prodi-Serrin condition on the solenoidal component is sufficient to force the solution into the regularity class.
2. Refinement of Serrin's Criterion: It provides a local version of Serrin's regularity criterion that does not rely on the solution solving an initial value problem (Cauchy problem) or having global energy bounds. This is a significant step toward understanding the intrinsic regularity of distributional solutions.
3. Decomposition Insight: The work highlights that the regularity of the Navier-Stokes equations is fundamentally tied to the solenoidal part of the velocity field, while the gradient part behaves like a harmonic function that only requires mild time-integrability to ensure smoothness.
4. Time Regularity Conjecture: The author suggests that if the time derivatives of the potential part ($\partial_t^k \pi$) are sufficiently integrable, then the time derivatives of the full solution ($\partial_t^k u$) will also be regular, opening a path for future research into full space-time regularity.

Conclusion

Galdi's paper resolves a long-standing question by demonstrating that the Prodi-Serrin condition applied to the solenoidal component of a distributional solution is sufficient to guarantee local spatial smoothness, provided the potential component is not too singular in time. This result decouples the regularity theory from the restrictive assumption of finite energy (Leray-Hopf class), offering a more general and sharp characterization of regular solutions to the Navier-Stokes equations.