The L p spectrum of the Stokes operator on hyperbolic 3-space and exponential decay for the Navier-Stokes equations

arXiv:2605.22212 · math-ph, math.AP, math.DG, math.MP, physics.flu-dyn · Submitted 2026-05-21 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "The L p spectrum of the Stokes operator on hyperbolic 3-space and exponential decay for the Navier-Stokes equations".

Mira: Here's the gist of the paper. We prove that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3(H3), and that this solution decays exponentially to zero.

Kai: Global Well-Posedness and Decay The paper establishes that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3(H3), and that this solution decays exponentially to zero.

Mira: The exponential decay rate is µλ(3)Def, where µ is the kinematic viscosity and λ(3)Def = 26/9 is the effective spectral gap of the deformation Laplacian in L 3(H3).

Kai: On flat R cubed, the corresponding Kato-type result gives only algebraic decay.

Title and authors: Mira: The authors are Zhi-Wei Wang and Samuel L Braunstein, and their title really tells you what they're tackling: connecting the spectrum of a specific operator on hyperbolic three-space to the decay rate of Navier-Stokes solutions <ref:2605.22212#pg1>.

Kai: They’re focusing on how negative curvature changes the behavior compared to flat space, specifically looking at that L p spectrum. It suggests that this geometric structure provides something fundamentally different for fluid dynamics when you move away from Euclidean space.

Lev: I wonder what the implication is for numerical simulations? Because if we can predict exponential decay, that’s a huge win over just algebraic decay which is what we usually see on flat R3.

Mira: That's the main thing. The paper shows that this curvature provides an exponential factor in the decay rate of L q norms for p greater than or equal to three <ref:2605.22212#pg1>. This is a macroscopic consequence of having that spectral gap from the negative curvature itself, which they call lambda(three) Def, which is twenty-six over nine <ref:2605.22212#pg1,is a macroscopic consequence of>.

Kai: So we’re talking about turning power-law decay into exponential decay just by changing the geometry of the space we are simulating on. That's a big conceptual shift for long-term prediction.

Lev: And from an engineering standpoint, that means we might need less simulation time to get a good result for large time scales, provided our initial data is small enough to start with this exponential stability guarantee.

The paper's summary: Kai: The core summary of the paper centers on proving that for sufficiently small initial data in L three of H3, there is a unique global mild solution to the Navier-Stokes equations with this deformation Laplacian.

Mira: They are showing uniqueness and global existence, which is always a big deal in these kinds of problems. But more importantly, they establish that this solution decays exponentially to zero at a rate determined by the kinematic viscosity multiplied by that spectral gap, µ times lambda(three) Def <ref:2605.22212#pg1,that this solution decays exponentially to zero>.

Lev: That explicit exponential decay rate is what makes this paper interesting for error correction researchers. It gives us a clear target for stability analysis on hyperbolic manifolds if we were to try and model something complex there.

Kai: And they explicitly contrast this with flat space R3, where the corresponding result only yields algebraic decay, like t to the minus one over four term alone. So the curvature is what upgrades the long-time behavior from a slow power-law decay to something much faster.

Mira: They also discuss some limitations upfront—the paper points out that while they have this nice exponential decay for large times, it doesn't help with getting better regularity for data that is rougher than L three.

Lev: I see what they mean though; the obstruction is still there, and it comes from the local scaling behavior of the heat kernel. So you still can't get infinite regularity just by having negative curvature when your starting data is too rough.

The paper's improvements: Kai: The authors suggest a few things that are really important for understanding this result. One major improvement they highlight is that the choice of the deformation Laplacian is not arbitrary; it’s actually the analytically optimal one for negative curvature because it gives them that full spectral gap of four and that specific decay rate of twenty-six over nine <ref:2605.22212#pg1>.

Mira: They make a strong point about how this operator choice matters beyond just fitting the kinematics. If you used something else, like the Hodge Laplacian, they say it would have a zero spectral gap on divergence-free fields, which wouldn't give you any of that exponential benefit at all.

Lev: That makes sense for real hardware implementation. If we were designing an error correction scheme on a manifold, picking the operator that gives you the maximal spectral gap is crucial because it dictates the stability bounds you can achieve.

Kai: And they also point out how this analysis sets up a sharp boundary between what global geometry can improve and what local Euclidean structure still controls. It shows that curvature helps with long-term decay but doesn't fix short-time singularities for rougher data, which is a key distinction.

Mira: They establish this boundary by showing that the scaling integral for short times, the Fujita-Kato contraction argument, depends only on local Euclidean structure and is independent of the spectral gap when p is less than three.

Lev: So even with all this geometric machinery, if your initial data isn't smooth enough in L three, you’re still hitting that short-time singularity dictated by how the heat kernel scales locally. That’s a hard limit on what curvature can do for regularity.

Conclusion: Kai: To wrap things up, this paper on "The L p spectrum of the Stokes operator on hyperbolic three-space and exponential decay for the Navier-Stokes equations" shows that negative curvature gives us exponential time decay for fluid solutions when you start with small enough initial data in L three <ref:2605.22212#pg1>.

Mira: The main implication is that the geometry of a space like H3 fundamentally alters how fluid problems stabilize over long periods, moving us from algebraic decay to exponential decay.

Lev: For me, the most concrete thing is that the paper identifies the deformation Laplacian as analytically optimal for this setting because it provides the full spectral gap of four and results in that twenty-six over nine decay rate <ref:2605.22212#pg1>.

Kai: And they also clearly delineate what curvature can do—improve long-term decay—and what it cannot do—cure short-time singularities for rougher data.

Mira: So, even with a geometrically favorable space, you still have to worry about the local scaling integral determining the short-time behavior, which is independent of the spectral gap.

Lev: Exactly. We need small initial data to get that exponential decay from this paper, and we have to be careful about how rough that initial data is for any regularity gains at all.

College of Physics, Jilin University · Computer Science, University of York

math-ph, math.AP, math.DG, math.MP, physics.flu-dyn

Submitted: 2026-05-21

Updated: 2026-11-07

Comments: 15 pages

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 85/100

The gist: We prove that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial…

Key concepts

Deformation Laplacian
This specific differential operator is chosen because it correctly follows Lagrangian kinematics on hyperbolic space. It is crucial because it possesses a full spectral gap of 4, which dictates the optimal exponential decay rate for the fluid solutions.
Spectral Gap ($\lambda(3)Def$)
The spectral gap refers to the lowest non-zero eigenvalue associated with the Stokes operator on H3. In this context, it is $\lambda(3)Def = 26/9$. This value quantifies how quickly the solution decays over time, converting power-law decay into exponential decay.
UV/IR Boundary
This concept defines a sharp limit on what geometry can achieve for fluid equations. Curvature provides Infrared (IR) improvement—exponential decay at large times—but it cannot fix Ultraviolet (UV) obstructions, which are short-time singularities determined by the local Euclidean structure of the manifold.

Terminology

Summary

We prove that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3(H3), and that this solution decays exponentially to zero. p. 0

Where the paper says it

braunstein@york. uk June 16, 2026 Abstract We prove that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3 (H3 ), and that this solution decays exponentially to zero.

Global Well-Posedness and Decay

The paper establishes that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3(H3), and that this solution decays exponentially to zero. The exponential decay rate is µλ(3)Def, where µ is the kinematic viscosity and λ(3)Def = 26/9 is the effective spectral gap of the deformation Laplacian in L 3(H3). p. 0

Where the paper says it

uk June 16, 2026 Abstract We prove that the three-dimensional incompressible Navier-Stokes equations with the deformation Laplacian on hyperbolic 3-space H3 admit a unique global mild solution for sufficiently small initial data in L 3 (H3 ), and that this solution decays exponentially to zero. The exponential decay rate is µλ(3) Def, where µ is the kinematic viscosity and λ (3) Def = 26/9 is the effective spectral gap of the deformation Laplacian in L 3 .

On flat R cubed, the corresponding Kato-type result gives only algebraic decay. p. 0
Where the paper says it

The exponential decay rate is µλ(3) Def, where µ is the kinematic viscosity and λ (3) Def = 26/9 is the effective spectral gap of the deformation Laplacian in L 3 . On flat R 3 , the corresponding Kato-type result gives only algebraic decay.

The exponential stability is a macroscopic consequence of the spectral gap provided by negative curvature. p. 0
Where the paper says it

On flat R 3 , the corresponding Kato-type result gives only algebraic decay. The exponential stability is a macroscopic consequence of the spectral gap provided by negative curvature.

Spectral Theory and Operator Properties

The analysis relies on the deformation Laplacian, which is established as the unique choice consistent with Lagrangian kinematics. The Stokes operator A = -P∆Def on L squared divergence-free vector fields on H3 has spectrum σ(A) ⊆ [4,∞). p. 2

Where the paper says it

2 The L 2 spectral gap Proposition 2. The Stokes operator A = −P∆Def on L 2 divergence-free vector fields on H3 has spectrum σ(A) ⊆ [4,∞).

The effective bilinear spectral gap γ = 26/9 collects the exponential factors. p. 4
Where the paper says it

Substituting into (10): ∥e −τµAP∇ · (u ⊗ v)∥Lq ≤ C (τ −δ + τ −δ0 ) e −µγτ ∥u∥Lq ∥v∥Lq , (13) where γ = 26/9 collects the exponential factors. The exponent δ = 1/2 + 3/(2q) and the large-time correction δ0 = 3/(2q) capture the temporal singularities of the bilinear term.

With p = 3 and q = 6, β = 1/4, δ = 3/4, and δ0 = 1/4. p. 4
Where the paper says it

Proposition 4. For 1 < r ≤ q < ∞ and τ > 0: ∥e −τµAP∇ · F∥Lq ≤ C (τ −δ + τ −δ0 ) e −µγτ ∥F∥Lr , (10) where F is a symmetric tensor field, δ = 1 2 + 3 2 ( 1 r − 1 q ), δ0 = 3 2 ( 1 r − 1 q ), and γ = 2+λ (r ′ ) 0 /2+λ (r) 0 /2 is the effective bilinear spectral gap.

The map u -> e−tµAu0+B(u, u) is a contraction on the ball where∥u0∥L3 <ϵ0 ≡ (4C1C2)−1. p. 6
Where the paper says it

The map u 7→ e −tµAu0+B(u, u) is a contraction on the ball {u ∈ X : ∥u∥X ≤ 2C1∥u0∥L3 } provided 4C1C2∥u0∥L3 < 1, i. , ∥u0∥L3 < ϵ0 ≡ (4C1C2) −1 .

The unique fixed point satisfies∥u(t)∥L6 ≤ 2C1∥u0∥L3 t−1/4 e−µγt, giving (3) with the full exponential decay rate γ = λ(3)Def = 26/9. p. 6
Where the paper says it

, ∥u0∥L3 < ϵ0 ≡ (4C1C2) −1 . The unique fixed point satisfies ∥u∥X ≤ 2C1∥u0∥L3 , which unpacks to ∥u(t)∥L6 ≤ 2C1∥u0∥L3 t −1/4 e −µγt , (22) giving (3) with the full exponential decay rate γ = λ (3) Def = 26/9.

UV Obstruction and Geometric Invariance

The Fujita-Kato contraction argument shows that for p ≤ 3/2, the integral strictly diverges for all t > 0 regardless of the curvature. For p = 3/2

max(3, p), but the scaling exponent is independent of q and the spectral gap contributes only a bounded exponential factor, which is O(1) as t → 0. This shows that no amount of global curvature can change the ultraviolet scaling because it is determined by the local Euclidean structure. p. 7

Where the paper says it

The obstruction is a short-time, high-frequency phenomenon determined by the local Euclidean structure of the manifold. Because every Riemannian manifold is locally flat, no amount of global curvature can change the ultraviolet scaling.

Physical Interpretation and Operator Choice

The choice of the deformation Laplacian is justified by its unique consistency with Lagrangian kinematics. The results locate a sharp boundary between what global geometry (curvature, spectral gap) can and cannot do for the 3D Navier-Stokes equations. p. 7

Where the paper says it

On such manifolds, the deformation Laplacian’s spectral gap continues to dictate the macroscopic stabilization of the fluid, further cementing its role as the analytically optimal viscous operator. 2 The UV/IR boundary The results locate a sharp boundary between what global geometry (curvature, spectral gap) can and cannot do for the 3D Navier-Stokes equations: What curvature provides (IR improvement): exponential time decay for L p norms with p ≥ 3 (by interpolation with the semigroup bounds of Proposition 2.

The deformation Laplacian is not only kinematically correct but is also the analytically optimal one on negatively curved manifolds because it provides the full spectral gap of 4 and the corresponding decay rate λ(3)Def = 26/9. p. 7
Where the paper says it

This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations. The deformation Laplacian is not only the kinematically correct choice; it is the analytically optimal one on negatively curved manifolds.

This illustrates that the choice of viscous operator has analytical consequences beyond the formal structure of the equations. p. 7
Where the paper says it

Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ (3) Def = 26/9. This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations.

Comparison with Flat Space

On R cubed, the Kato theorem gives global existence for small L cubed data with algebraic decay∥u(t)∥Lq ≤ C t− 3/2(1 − 1/q). On H3, the decay acquires an exponential factor:∥u(t)∥Lq ≤ C t− 3/2(1 − 1/q)e−µλ(3)Def. p. 6

Where the paper says it

1 Comparison with flat space On R 3 , the Kato theorem [4] gives global existence for small L 3 data with algebraic decay ∥u(t)∥Lq ≤ C t− 3 2 ( 1 3 − 1 q ) . On H3 , the decay acquires an exponential factor: ∥u(t)∥Lq ≤ C t− 3 2 ( 1 3 − 1 q ) e −µλ(3) Deft .

The short-time behaviour (t≪1) is identical, and the improvement is entirely at large times (t≫1). p. 6
Where the paper says it

On H3 , the decay acquires an exponential factor: ∥u(t)∥Lq ≤ C t− 3 2 ( 1 3 − 1 q ) e −µλ(3) Deft . The short-time behaviour (t ≪ 1) is identical; the improvement is entirely at large times (t ≫ 1).

The smallness threshold epsilon0 = (4C1C2)−1 is the same as on R cubed because the Beta-function constant B(1/2, 1/4) is geometry-independent. p. 6
Where the paper says it

The short-time behaviour (t ≪ 1) is identical; the improvement is entirely at large times (t ≫ 1). The smallness threshold ϵ0 = (4C1C2) −1 is the same as on R 3 (since the Beta-function constant B(1/2, 1/4) is geometry-independent).

The Role of Curvature

What curvature provides is IR improvement, namely exponential time decay for L p norms with p ≥ 3 by interpolation with the semigroup bounds of Proposition 2.3. What curvature cannot provide is improved regularity for data rougher than L cubed because the obstruction is a short-time, high-frequency phenomenon determined by the local Euclidean structure of the manifold. p. 7

Where the paper says it

What curvature cannot provide (UV obstruction): improved regularity for data rougher than L 3 . The obstruction is a short-time, high-frequency phenomenon determined by the local Euclidean structure of the manifold.

The results extend to broader classes of complete, simply connected 3-manifolds with pinched negative sectional curvature where the deformation Laplacian’s spectral gap continues to dictate macroscopic stabilization. p. 7
Where the paper says it

the exact ultraviolet scaling obstruction, extends universally to the broader class of complete, simply connected 3-manifolds with pinched negative sectional curvature. On such manifolds, the deformation Laplacian’s spectral gap continues to dictate the macroscopic stabilization of the fluid, further cementing its role as the analytically optimal viscous operator.

The Role of the Deformation Laplacian

With the Hodge Laplacian, the Stokes operator on H3 would be AH = ∆H, with spectral gap 0 on divergence-free fields by Proposition 2. p. 7

Where the paper says it

3 The role of the deformation Laplacian The results are specific to the deformation Laplacian. With the Hodge Laplacian, the Stokes operator on H3 would be AH = ∆H, with spectral gap 0 on divergence-free fields (by Proposition 2.

1. Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ(3)Def = 26/9. p. 7
Where the paper says it

The Bochner Laplacian would give AB = −∆B = ∆H + 2, with gap 2. Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ (3) Def = 26/9.

This highlights that the choice of viscous operator has analytical consequences beyond formal structure. p. 7
Where the paper says it

Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ (3) Def = 26/9. This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations.

The deformation Laplacian is not only kinematically correct but is also the analytically optimal one on negatively curved manifolds because it provides the full gap of 4 and the corresponding decay rate λ(3)Def = 26/9. p. 7
Where the paper says it

This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations. The deformation Laplacian is not only the kinematically correct choice; it is the analytically optimal one on negatively curved manifolds.

Where the Gains Come From

The gains come from leveraging the exact L squared spectral gap of the deformation Laplacian on H3 to derive explicit optimal exponential decay rate µλ(3)Def = µ · 26/9. The explicit computation captures the full effective spectral gap rather than a half-gap, avoiding suboptimal decay rates. p. 1

Where the paper says it

Furthermore, our explicit computation captures the full effective spectral gap rather than a half-gap. This avoids the suboptimal decay rates that occur when mathematically trading spectral bounds to absorb polynomial time singularities.

The curvature of H3 converts the large-time behaviour from power-law to exponential. p. 2
Where the paper says it

The exponential factor e −µλ(3) Deft is the qualitative novelty: on R 3 , the corresponding decay is algebraic (t −1/4 alone). The curvature of H3 converts the large-time behaviour from power-law to exponential.

This exponential factor is the qualitative novelty on H3 compared to R cubed where it is algebraic (t−1/4 alone). p. 2
Where the paper says it

The exponential factor e −µλ(3) Deft is the qualitative novelty: on R 3 , the corresponding decay is algebraic (t −1/4 alone). The curvature of H3 converts the large-time behaviour from power-law to exponential.

The UV/IR Boundary

The results locate a sharp boundary between what global geometry can and cannot do for the 3D Navier-Stokes equations. What curvature provides (IR improvement) is exponential time decay for L p norms with p ≥ 3 by interpolation with the semigroup bounds of Proposition 2. p. 7

Where the paper says it

On such manifolds, the deformation Laplacian’s spectral gap continues to dictate the macroscopic stabilization of the fluid, further cementing its role as the analytically optimal viscous operator. 2 The UV/IR boundary The results locate a sharp boundary between what global geometry (curvature, spectral gap) can and cannot do for the 3D Navier-Stokes equations: What curvature provides (IR improvement): exponential time decay for L p norms with p ≥ 3 (by interpolation with the semigroup bounds of Proposition 2.

3. What curvature cannot provide (UV obstruction) is improved regularity for data rougher than L cubed because the obstruction is a short-time, high-frequency phenomenon determined by the local Euclidean structure of the manifold. p. 7
Where the paper says it

What curvature cannot provide (UV obstruction): improved regularity for data rougher than L 3 . The obstruction is a short-time, high-frequency phenomenon determined by the local Euclidean structure of the manifold.

The boundary between what curvature helps and what it cannot is located precisely at Kato’s critical space L cubed. p. 2
Where the paper says it

The spectral gap contributes only the exponential factor, which controls the large-time behaviour but cannot cure the short-time singularity. The boundary between what curvature helps and what it cannot is located precisely at Kato’s critical space L 3 .

The exponent 1/2−3/(2p) is determined by the local (short-time) scaling of the heat kernel, which is identical on H3 and R cubed. p. 2
Where the paper says it

In particular, L 2 data remains supercritical on H3 . The exponent 1/2−3/(2p) is determined by the local (short-time) scaling of the heat kernel, which is identical on H3 and R 3 .

The spectral gap contributes only the exponential factor, which controls large-time behaviour but cannot cure short-time singularity. p. 2
Where the paper says it

The exponent 1/2−3/(2p) is determined by the local (short-time) scaling of the heat kernel, which is identical on H3 and R 3 . The spectral gap contributes only the exponential factor, which controls the large-time behaviour but cannot cure the short-time singularity.

The results are specific to the deformation Laplacian. With the Hodge Laplacian, the Stokes operator on H3 would be AH = ∆H, with spectral gap 0 on divergence-free fields by Proposition 2.1. Only the deformation Laplacian gives the full gap of 4 and corresponding decay rate λ(3)Def = 26/9. p. 7

Where the paper says it

The Bochner Laplacian would give AB = −∆B = ∆H + 2, with gap 2. Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ (3) Def = 26/9.

This illustrates a point made in [2] that the choice of viscous operator has analytical consequences beyond formal structure. p. 7
Where the paper says it

Only the deformation Laplacian gives the full gap of 4 and the corresponding decay rate λ (3) Def = 26/9. This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations.

The deformation Laplacian is not only kinematically correct but is also the analytically optimal one on negatively curved manifolds because it provides the full gap of 4 and corresponding decay rate λ(3)Def = 26/9. p. 7
Where the paper says it

This illustrates a point made in [2]: the choice of viscous operator has analytical consequences beyond the formal structure of the equations. The deformation Laplacian is not only the kinematically correct choice; it is the analytically optimal one on negatively curved manifolds.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI can predict long-term fluid behavior on hyperbolic manifolds by utilizing exponential decay rates rather than just algebraic ones. The improved system can leverage Theorem 1.1 to forecast that the curvature of H3 converts the large-time behaviour from power-law to exponential for Navier-Stokes solutions under small initial data.

  2. Improved AI can perform rigorous short-time analysis of fluid regularity by identifying the critical scaling exponent based on initial data integrability, which is geometry independent. This allows the system to determine that the scaling exponent is independent of q and the spectral gap, meaning it can accurately predict when solutions remain supercritical regardless of curvature.

  3. Improved AI can distinguish between geometric stabilization effects and local Euclidean obstructions in fluid dynamics problems. By using Theorem 1.2, the improved system can identify that no amount of global curvature can change the ultraviolet scaling for rougher data, allowing it to correctly assess the limits of geometric influence on short-time regularity.

  4. Improved AI can select the analytically optimal viscous operator for fluid simulations on curved spaces by prioritizing spectral advantages over kinematic convenience. The improved system recognizes that The deformation Laplacian is not only the kinematically correct choice; it is the analytically optimal one on negatively curved manifolds because it provides a full spectral gap where other operators do not.

  5. Improved AI can implement a dual-path analysis to bound nonlinear terms in fluid equations with high precision. The improved system uses Proposition 4.1 to show that for Navier-Stokes, The effective bilinear spectral gap gamma = 26/9 collects the exponential factors, leading to tighter stability estimates than those derived from simpler bounds.

Abstract

We determine the L p spectrum, 1<p< infinity, of the Stokes operator A=- on divergence-free vector fields of hyperbolic 3-space of curvature-a squared, where =Δ B+ is the Ebin-Marsden deformation Laplacian and is the Leray projector. The spectrum is the parabolic region a 24+z 2: z in, z at most1-2/p. For p>2 every interior point is an eigenvalue, realized by an explicit family of divergence-free swirl fields which leave L p exactly at the boundary of the region. It follows that the Stokes semigroup on has exponential type exactly-μ p with μ p=a 2 (4-(1-2/p) 2). At p=3 this is 35a 2/9, which lies strictly above the rates 26a 2/9 and 10a 2/3 that follow from pointwise domination by the scalar heat semigroup and from interpolation, and strictly below the L squared gap 4a squared. For the Navier-Stokes equations with small data in L cubed, whose global solvability on cubed is due to Pierfelice, the solutions decay in L cubed and in L 6 at every rate below μ times35a 2/9, where μ is the viscosity, and no faster rate holds for the linearized flow. The exponent 3 is the usual scaling threshold. Curvature fixes the time scale of the decay and introduces no new threshold.

Sources

Related papers