A mathematical study of an elastic-viscous-plastic sea-ice model with the Kelvin-Voigt rheology

Daniel W. Boutros , Xin Liu , Marita Thomas , and Edriss S. Titi · 2026

Read A mathematical study of an elastic-viscous-plastic sea-ice model with the Kelvin-Voigt rheology by Daniel W. Boutros , Xin Liu , Marita Thomas , and Edriss S. Titi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-26295.txt.

Introduction

Boutros and three colleagues built a sea-ice model that treats ice as an elastic-viscous-plastic material with Kelvin-Voigt regularization, meaning stress depends on strain rate as well as strain. Earlier work placed that regularization in the stress law; here it enters the momentum balance instead. The team proved local well-posedness with advection and global well-posedness without it, via a new bound on the stress tensor. The proofs also allow viscosity coefficients with no upper cutoff and rougher starting data.

Catalog record

CREATORDaniel W. Boutros , Xin Liu , Marita Thomas , and Edriss S. Titi
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE67738
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.26295

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A mathematical study of an elastic-viscous-plastic sea-ice model with the Kelvin-Voigt rheology Daniel W. Boutros , Xin Liu , Marita Thomas , and Edriss S. Titi Abstract Motivated by the elastic-viscous-plastic (EVP) sea-ice model [E. C. Hunke and J. K. Dukowicz, J. Phys. Oceanogr., 27, 9 (1997), 1849–1867], which is used in large-scale numerical climate simulations, we proposed in [D. W. Boutros, X. Liu, M. Thomas and E. S. Titi, arXiv:2505.03080 (2025)] the use of the inviscid Voigt regularisation for the constitutive (stress-tensor) relation and proved the global well-posedness of the resulting model. The EVP model treats sea ice as a non-Newtonian fluid. In turn, elastic-viscous-plastic solids often involve a Kelvin-Voigt viscosity in terms of the strain rate. Therefore, in the present work we formulate an elastic-viscous-plastic sea-ice model with a Kelvin-Voigt regularisation in terms of the strain rate. In other words, we introduce the Voigt regularisation in the momentum balance rather than in the constitutive relation (for the stress tensor). We then prove the local well-posedness for the Kelvin-Voigt EVP model with the advection term, in the momentum balance, and the global well-posedness in the absence of the advection term (following a very standard approximation in the latter case). A crucial component of the proof of these results, is a new L ∞ L^{\infty} -estimate for the stress tensor which relies on the damping structure. Note that, both with and without the advection term, we are able to handle the case of viscosity coefficients without a cutoff from above, which remains a major open problem for the closely related Hibler sea-ice model. We are also able to prove the existence of solutions for much less regular initial data compared to our previous paper on the Voigt-EVP model. Keywords: sea-ice dynamics; elastic-viscous-plastic rheology; well-posedness; Hibler’s sea-ice model; Kelvin-Voigt regularisation; viscoplasticity; non-Newtonian flows Mathematics Subject Classification: 35Q86 (primary), 35A01, 35A02, 35A09, 35B65, 74D10, 74H20, 74H25, 74H30, 86A08, 86A40 (secondary) 1 Introduction 1.1 Formulation of the EVP model The evolution of the sea-ice cover plays a fundamental role in the climate system, and adequate modelling of the dynamics of sea ice is therefore of great importance. In this paper, we consider the elastic-viscous-plastic (EVP) sea-ice model, which was originally introduced in [ 33 ] . The EVP model, on the time interval [0,T] and the two-dimensional flat torus \mathbb{T}^{2}=(\mathbb{R}/\mathbb{Z})^{2} , is given by \displaystyle\partial_{t}u=\nabla\cdot\sigma+\mathcal{T}_{a}+\mathcal{T}_{w}+\Omega u^{\perp}-g\nabla H_{0}, (1.1a) \displaystyle\dfrac{1}{\mathcal{E}}\partial_{t}\sigma+\dfrac{4\mathcal{D}}{P}(\sigma-\frac{1}{2}\Tr\sigma\mathbb{I}_{2})+\dfrac{\mathcal{D}}{2P}\Tr\sigma\mathbb{I}_{2}+\dfrac{\mathcal{D}}{2}\mathbb{I}_{2}=D(u), (1.1b) \displaystyle u\lvert_{t=0}=u_{0},\hskip 10.00002pt\sigma\lvert_{t=0}=\sigma_{0}, (1.1c) where P is the internal ice strength parameter, which is a given positive constant in our setting, while u:\mathbb{T}^{2}\times[0,T]\rightarrow\mathbb{R}^{2} is the velocity field, \sigma:\mathbb{T}^{2}\times[0,T]\rightarrow\mathbb{R}^{2\times 2}_{\mathrm{sym}} is the stress tensor, which are the unknowns. The tensor field D(u) is the deformation-rate tensor, i.e., the symmetric part of the velocity gradient D(u)\coloneqq\frac{1}{2}[\nabla u+(\nabla u)^{\top}]. (1.2) In system ( 1.1 ) above we have used the strain rate, which is defined (and simplified) as follows \mathcal{D}=\lvert D(u)\rvert, (1.3) where in this paper we will use the notation \lvert\cdot\rvert for the Euclidean norm for tensors. In other words, for a tensor (A_{ij})_{i,j=1}^{2} we define the norm \lvert\cdot\rvert as follows \lvert A\rvert\coloneqq\sqrt{\sum_{i,j=1}^{2}\lvert A_{ij}\rvert^{2}}. In addition, \mathcal{T}_{a} and \mathcal{T}_{w} are the atmospheric and oceanic drag forces, which are given by \displaystyle\mathcal{T}_{a} \displaystyle\coloneqq c_{a}\rho_{a}\lvert U_{a}\rvert\bigg(U_{a}\cos\phi+U_{a}^{\perp}\sin\phi\bigg), (1.4) \displaystyle\mathcal{T}_{w} \displaystyle\coloneqq c_{w}\rho_{w}\lvert U_{w}-u\rvert\bigg[(U_{w}-u)\cos\theta+(U_{w}-u)^{\perp}\sin\theta\bigg]. (1.5) In the Coriolis term, in equation ( 1.1a ), we have used the notation v^{\perp}=(-v_{2},v_{1})^{\top} for v\in\mathbb{R}^{2} . Like in the original paper [ 33 ] , we have taken the mean ice thickness h and the ice compactness A to be constant, which in turn implies that the mass m and the internal ice strength P are also constants (as we have already assumed above). The given function H_{0} in equation ( 1.1a ) describes the ocean surface topography. All the remaining parameters in system ( 1.1 ) and equations ( 1.4 )-( 1.5 ) (in particular the angles \phi and \theta ) are introduced in Table 1 , below. Symbol Meaning Typical value Equation of first appearance A ice compactness (area covered by ‘thick’ ice) 0\leq A\leq 1 c_{a} air drag coefficient 1.2\cdot 10^{-3} ( 1.4 ) c_{w} ocean drag coefficient 5.5\cdot 10^{-3} ( 1.5 ) \mathcal{E} elastic modulus 0.25 ( 1.1b ) g gravitational constant 9.81\;\text{m\,s}^{-1} ( 1.1a ) H_{0} sea surface height ( 1.1a ) \Omega rotation parameter 1.46\cdot 10^{-4}\;\text{s}^{-1} ( 1.1a ) P internal ice strength ( 1.1b ) \phi air turning angle 25^{\circ} ( 1.4 ) \rho_{a} air density 1.3\;\text{kg/m}^{3} ( 1.4 ) \rho_{w} ocean water density 1026\;\text{kg/m}^{3} ( 1.5 ) \theta water turning angle 25^{\circ} ( 1.5 ) U_{a} geostrophic wind ( 1.4 ) U_{w} geostrophic ocean current ( 1.5 ) Table 1: An overview of the notation used in the EVP model, the typical values are taken from: [ 33 , 32 , 52 , 43 ] . The EVP model was introduced in [ 33 ] as a numerical regularisation of the viscous-plastic Hibler sea-ice model [ 31 ] . The Hibler model corresponds to formally setting \mathcal{E}=\infty in equations ( 1.1a ) and ( 1.1b ). In a previous paper by the authors [ 10 ] , it has been observed that the 1D EVP model is (formally) linearly ill-posed in Sobolev spaces (in the absence of the advection term and utilising a regularisation of the strain rate \mathcal{D} of the type ( 1.10 ), below). This led to the introduction of the Voigt-regularisation -\alpha^{2}\partial_{t}\Delta\sigma , with \alpha>0 , on the left-hand side of equation ( 1.1b ) in [ 10 ] in order to regularise the system. The motivation for this choice of regularisation is like that in the original EVP model [ 33 ] , namely, it preserves the steady states as well as the formal asymptotic in time statistical (infinite-time average) solutions of the Hibler model. In addition, it is a modification of the elastic term (which was introduced for numerical purposes) rather than the viscous-plastic terms in the constitutive relation (for the stress tensor). In this work we consider the other case, namely introducing a Kelvin-Voigt regularisation -\alpha^{2}\partial_{t}\Delta u , with \alpha>0 , on the left-hand side of equation ( 1.1a ). The motivation for doing so is twofold. On the one hand, this choice of regularisation is purely an introduction of a Kelvin-Voigt component into the rheology and therefore has a clear physical interpretation. On the other hand, from the mathematical analysis perspective the nonlinearities of the EVP model contain derivatives of u rather than \sigma , which is a reason to introduce a regularisation for u instead of \sigma . Due to the Kelvin-Voigt regularisation, we are able to prove the existence of solutions for initial data of lower regularity compared to the case of the Voigt-regularisation of the constitutive relation (for the stress tensor) in [ 10 ] . In the present work we will establish two results, specifically, the local well-posedness of the EVP model with the advection term and Voigt-regularisation in the momentum balance ( 1.1a ) in Theorem 1.1 , and in Theorem 1.2 the global well-posedness of the EVP model without the advection term but including the Voigt regularisation in the momentum balance ( 1.1a ). 1.2 Main results of this paper In this paper, we consider two different formulations of the EVP model

Read the complete text