Global well-posedness of the elastic-viscous-plastic sea-ice model with the inviscid Voigt-regularisation

This paper establishes the global well-posedness of the elastic-viscous-plastic (EVP) sea-ice model under inviscid Voigt-regularization, thereby resolving the long-standing issue of handling viscosity coefficients without cutoff that has hindered the analysis of related sea-ice models.

The Gist

Imagine the Arctic Ocean as a giant, frozen puzzle. The pieces of this puzzle are chunks of sea ice, and they are constantly being pushed, pulled, and twisted by the wind and the ocean currents. Sometimes the ice flows smoothly like honey; other times, it cracks, piles up, and acts like a rigid, brittle solid.

For decades, scientists have tried to write a computer program (a mathematical model) to predict how this ice moves. This is crucial because the ice acts like a giant mirror for the sun and a blanket for the ocean, playing a huge role in our global climate.

The most famous model for this is called the Hibler model. Think of it as a very accurate but incredibly stubborn recipe. It works well in theory, but when you try to run it on a computer, it often crashes or takes forever to compute because the math gets "stuck" when the ice stops moving or moves very slowly. It's like trying to drive a car where the steering wheel locks up the moment you stop turning it.

To fix this, scientists invented a new recipe called the EVP (Elastic-Viscous-Plastic) model. Instead of treating the ice as a rigid block that instantly snaps into place, the EVP model pretends the ice has a tiny bit of "springiness" (elasticity). It's like saying the ice is made of a very stiff rubber band. This makes the computer math much easier to solve, allowing for faster simulations and parallel computing (using many computers at once).

However, there was a catch.

While the EVP model was great for computers, mathematicians were worried: Is this model actually mathematically sound? Could it produce impossible results, like ice moving at infinite speed or behaving in a way that defies physics? In fact, the original EVP model had a hidden flaw: under certain conditions, the math could become "ill-posed," meaning the solution could blow up or become unpredictable, much like a house of cards collapsing in a breeze.

The Breakthrough: The "Voigt" Safety Net

In this paper, the authors (Boutros, Liu, Thomas, and Titi) decided to tackle this problem head-on. They didn't just run simulations; they built a rigorous mathematical proof to show that a regularized version of the EVP model works perfectly.

Here is the simple analogy for what they did:

1. The Problem: The original EVP model is like a trapeze artist swinging without a safety net. If they slip (mathematically speaking), they fall.
2. The Solution: The authors added a "Voigt regularisation." Imagine this as a safety net made of invisible springs attached to the trapeze artist.
   • This "net" doesn't change the final destination (the ice still ends up where it should).
   • It doesn't change the speed of the swing significantly.
   • But, if the artist starts to wobble dangerously, the net catches them and stabilizes the motion.

In mathematical terms, they added a term involving the "Laplacian" (a measure of curvature) to the stress equation. This acts like a smoothing agent. It prevents the math from getting jagged, sharp, or infinite.

What Did They Prove?

The authors proved Global Well-Posedness. In plain English, this means:

• Existence: A solution always exists. The ice will always have a defined path, no matter how long you wait.
• Uniqueness: There is only one correct path for the ice. If you start with the same wind and ocean conditions, you will always get the same result.
• Stability: Small changes in the starting conditions (like a tiny shift in wind) lead to small changes in the result, not a total disaster.

They showed that even if you remove the artificial "cutoffs" (the safety valves scientists usually add to stop the math from breaking), the model remains stable thanks to their new "safety net."

Why Does This Matter?

Think of climate models as a giant, complex orchestra. The sea ice is one of the most important instruments. If the instrument is out of tune (mathematically unstable), the whole symphony (the climate prediction) sounds wrong.

• Before this paper: We had a great-sounding instrument (EVP) that was fast and efficient, but we weren't 100% sure it wouldn't break the music if played too long or too loudly.
• After this paper: We now have a mathematical guarantee that this instrument is solid. It will play the right notes, stay in tune, and never collapse, even under the most extreme conditions.

The "Super-Exponential" Bound

The paper mentions a "triple-exponential" bound. Imagine you are stacking blocks.

• A normal growth is like stacking one block on top of another.
• An exponential growth is like the stack doubling in height every second.
• A triple-exponential growth is like the stack doubling, then the rate of doubling doubles, and then that rate doubles again.

The authors found that while the numbers in their equations can get huge very quickly, they are still finite. They don't go to infinity. This is a massive victory because it proves the model is safe to use for long-term climate predictions.

Summary

This paper is the "quality control" stamp for one of the most important tools in climate science. The authors took a popular, fast, but mathematically risky model for sea ice, added a clever mathematical "safety net" (Voigt regularisation), and proved that it is robust, stable, and reliable forever. This gives scientists the confidence to use these models to predict our changing climate with greater certainty.

Technical Summary

Here is a detailed technical summary of the paper "Global well-posedness of the elastic-viscous-plastic sea-ice model with the inviscid Voigt-regularisation."

1. Problem Statement and Motivation

The paper addresses the rigorous mathematical analysis of the Elastic-Viscous-Plastic (EVP) sea-ice model, a standard component in global climate models (e.g., CICE, MITgcm). The EVP model was originally introduced by Hunke and Dukowicz (1997) as a numerical regularization of the classical Hibler model (1979) to facilitate explicit time-stepping and parallel computing.

Key Challenges:

• Singularities and Degeneracy: The constitutive law in the original Hibler model (and the unregularized EVP) involves a viscosity that becomes singular as the strain rate approaches zero and degenerate as it approaches infinity. This creates significant difficulties for both numerical simulation and mathematical analysis.
• Linear Ill-posedness: The authors demonstrate that the unregularized EVP model suffers from linear ill-posedness in Sobolev spaces. Specifically, they show that the linearized system can become locally elliptic under certain conditions, leading to Hadamard instability. This instability persists regardless of the viscosity cutoff parameter ($\epsilon$), distinguishing it from previous instability results found in the Hibler model literature.
• Lack of Rigorous Theory: While the Hibler model has seen some recent mathematical progress (local well-posedness), the EVP model, despite its widespread use, lacked a rigorous global well-posedness theory.

2. Methodology

To overcome the ill-posedness and singularities, the authors propose and analyze a regularized version of the EVP model, termed the Voigt-EVP system.

The Regularization Strategy:

1. Inviscid Voigt Regularization: The authors introduce a Voigt-type regularization to the evolution equation of the stress tensor ($\sigma$). Instead of the standard time derivative $\partial_t \sigma$, they use the operator $\partial_t (\sigma - \alpha^2 \Delta \sigma)$, where $\alpha > 0$ is a fixed parameter. This transforms the stress evolution into a pseudo-parabolic equation, adding a dispersive/damping effect without introducing artificial viscosity to the momentum equation.
2. Strain Rate Regularization (Intermediate Step): To handle the non-smoothness of the strain rate $D = |D(u)|$ at zero, they first analyze an intermediate system where the strain rate is smoothed: $D_\epsilon = \sqrt{|D(u)|^2 + \epsilon^2}$.
3. Galerkin Approximation: The existence of solutions is established using a Galerkin approximation scheme (Fourier modes).
4. A Priori Estimates: The core of the proof involves deriving uniform-in-$\epsilon$ and uniform-in-$N$ (approximation level) energy estimates in Sobolev spaces ($H^2$ for velocity, $H^3$ for stress).
   • $L^2$ and $H^1$ Estimates: Standard energy methods are used, leveraging cancellations similar to those in Oldroyd-B fluid models.
   • $H^2$ and $H^3$ Estimates: Higher-order estimates require handling the nonlinearity of the strain rate and the drag terms. The authors utilize the Brezis-Gallouët-Wainger inequality (a logarithmic Sobolev inequality) to control $L^\infty$ norms by $H^2$ norms.
   • Logarithmic Gronwall Inequality: Due to the logarithmic terms arising from the Brezis-Gallouët-Wainger inequality, the authors apply a logarithmic Gronwall inequality. This leads to triple-exponential bounds on the solution norms ($C \exp(\exp(\exp(CT)))$), which, while large, are finite for any finite time $T$.
5. Limit Passage:
   • First, the limit $N \to \infty$ is taken to establish the existence of strong solutions for the intermediate system (with $\epsilon > 0$).
   • Second, the limit $\epsilon \to 0$ is taken to recover the Voigt-EVP system with the original strain rate definition. The uniform estimates ensure the stability of this limit.

3. Key Contributions

• First Rigorous Analysis of EVP: This paper provides the first rigorous mathematical proof of global well-posedness for the EVP sea-ice model.
• Identification of New Instability: The authors rigorously prove that the unregularized EVP model is linearly ill-posed in Sobolev spaces due to a specific structural instability, distinct from the viscosity-cutoff issues in the Hibler model. This justifies the necessity of the Voigt regularization for mathematical well-posedness.
• Handling Unbounded Viscosity: Unlike previous analyses of the Hibler model which required a viscosity cutoff to avoid singularities, the Voigt regularization allows the authors to handle the case where viscosity coefficients are not cut off (i.e., taking the limit $\epsilon \to 0$).
• Structural Connection to Viscoelastic Fluids: The paper highlights the structural similarities between the EVP model and the Oldroyd-B model for viscoelastic non-Newtonian fluids, adapting techniques from that field to the geophysical context.

4. Main Results

Theorem 1.2 (Global Well-Posedness):
Given initial data $u_0 \in H^2(\mathbb{T}^2)$ and symmetric $\sigma_0 \in H^3(\mathbb{T}^2)$, and assuming appropriate regularity for external forces (wind, ocean currents, topography), the Voigt-EVP system admits a unique global strong solution $(u, \sigma)$ on any time interval $[0, T]$.

• Regularity:
  • Velocity: $u \in C([0, T]; H^2(\mathbb{T}^2))$
  • Stress: $\sigma \in C([0, T]; H^3(\mathbb{T}^2))$
  • Time derivatives: $\partial_t u \in L^\infty(0, T; H^2)$ and $\partial_t \sigma \in L^\infty(0, T; H^3)$.
• Symmetry: The stress tensor $\sigma$ remains symmetric for all time if it is initially symmetric.
• Stability: The solution depends continuously on the initial data, ensuring uniqueness.
• Bound: The solution norms satisfy a triple-exponential bound in time, derived from the repeated application of logarithmic Gronwall inequalities.

Corollary 1.4: The intermediate system (with the smoothed strain rate $D_\epsilon$) is also globally well-posed with the same regularity.

5. Significance and Implications

• Mathematical Foundation: This work lays the theoretical foundation for the EVP model, confirming that with the Voigt regularization, the model is mathematically sound and stable.
• Computational Relevance: The Voigt regularization is "inviscid" (it does not add artificial viscosity to the momentum equation) and shares the same steady states as the original Hibler model. This suggests that the regularization is a minimal modification that preserves the physical asymptotic behavior while ensuring numerical stability and mathematical well-posedness.
• Future Directions: The authors note that while the Voigt-EVP model is well-posed, the question of whether it converges to the original Hibler model as the elastic modulus $E \to \infty$ remains an open and complex problem, as numerical studies have shown discrepancies between the two. However, this paper establishes that the regularized EVP model itself is a robust dynamical system.
• Broader Impact: The techniques developed (handling non-Newtonian rheologies with strain-rate-dependent viscosities and Voigt regularization) are applicable to other complex fluid models and geophysical flow problems.

In summary, the paper successfully resolves the mathematical ill-posedness of the widely used EVP sea-ice model by introducing a Voigt regularization, proving the existence and uniqueness of global strong solutions, and providing a rigorous framework for future analysis of sea-ice dynamics.