Nonlocal ice strength regularization of the Viscous-Plastic sea ice rheology: resolving ill-posedness and resolution-dependent fracture patterns
Abstract. We implement a nonlocal ice strength regularization in the standard Viscous-Plastic (VP) and Elastic-Viscous-Plastic (EVP) models. To this end, we introduce an inhomogeneous Helmholtz equation for the ice strength that provides a physically based regularization of the ill-posed VP and EVP formulations while preserving the rate-independence of the plastic regime and the material instabilities that give rise to narrow deformation bands. A linear stability analysis of the new system of equations shows that the nonlocal regularization imposes an upper bound on the growth rates of perturbations, a necessary condition for well-posed problems. Those results are confirmed in idealized divergent numerical experiments: the standard (E)VP models yield resolution-dependent solutions with an increasing number of fractures as the spatial resolution is increased, whereas the nonlocal models produce a resolution-independent solution with a single fracture at the center of the ice plate. Numerical experiments demonstrating the ill-posedness of the hyperbolic tangent viscosity regularization previously identified through a linear stability analysis are also presented. With the advent of kilometer-scale sea ice simulations (Δx ≤ 2–3 km), a well-posed rheology is essential to ensure that fracture statistics reflect physical deformation processes rather than numerical instabilities.
This paper introduces a new regularization on the pressure potential for the viscous-plastic sea ice rheology. The goal is to resolve the ill-posedness and resolution-dependent fracture patterns. The references are probably cited and the paper is written very clear.
The basic idea of the regularization of this paper is to remove the ellipticity in space-time by introducing the Helmholtz equation for the acoustic wave/the ice strength. Therefore, the ice rheology will no longer interact with the acoustic wave in an ill-posed way as discussed in Guba-Lorenz-Sulsky-2013 and other references.
The analytical stability is based on the study of a linearized system, around a background flow similar to those in the literature. The numerical study of the paper also demonstrates strong stability of the regularization.
However, there are a few points the authors need to address.
1. Page 2, Line 35: `Regularization techniques typically increase the power of existing derivatives or add new higher-order derivatives in the system of Partial Differential Equations (PDEs)' is not a correct description. Regularization techniques should increase the regularity of the solutions by making the leading order operators admit compact solvers. This is the same as what the authors have done in eq 13, i.e., the Helmholtz equation: the solving map from P_H to P in eq 13 is compact, by dividing the every wavenumber (k-th) mode with 1+ l^2 k^2 (therefore adding a bound to the growth rate).
2. The authors mention that their non local regularization should work for EVP and demonstrate some evidence. As mentioned above, the regularization is on the ice strength. However, the ill-posedness of Boutros etc. 2026 is NOT based on the ice strength, where the ice strength is assumed to be constant. It is therefore unclear that the regularization proposed in this work should generally work for EVP. In fact, since the well-posedness/stability of this work is based on a bench mark around a particular configuration of the background flow (as explained in Figure 1), it naturally should not expect the regularization works for all configurations. It would be better for the authors to clarify this, and leave the rigorous well-posedness study as future work.
Overall, this paper proposes an interesting regularization of the viscous-plastic sea ice rheology. It potentially opens a research direction for this important field. Upon fixing the questions above, it is recommended the publication of this work.