Well-posedness of Hibler's dynamical sea-ice model (Q2141338)

The authors consider Hibler's sea-ice model written as: \(m(\partial_{t} \mathbf{u}+\mathbf{u}\cdot \nabla \mathbf{u})+\nabla p=\operatorname{div}\mathbb{S}+ \mathcal{F}\), \(\partial_{t}h+\operatorname{div}(h\mathbf{u})=\mathcal{S}_{h}\), \(\partial_{t}A+\operatorname{div}(A\mathbf{u})=\mathcal{S}_{A}+A\operatorname{div}\mathbf{u}\cdot \chi_{\{A\geq 1\}}\), where \(m\) is the ice mass \(m=\rho_{ice}h\), \(\mathbf{u}\) the sea-ice velocity, \(p\) the pressure \(p=c_{p}h\exp(c_{a}A)\), \(\mathbb{S}\) the viscoplastic stress \(\mathbb{S}=p\frac{\nabla \mathbf{u}+\nabla \mathbf{u}^{T}}{\left\vert \nabla \mathbf{u}+\nabla \mathbf{u}^{T}\right\vert }+p\frac{ \operatorname{div}uI_{2}}{\left\vert \operatorname{div}\mathbf{u}\right\vert }\), \(\mathcal{F}=-m\eta \mathbf{u}^{\perp }+\emptyset_{a}+\emptyset_{w}\), with \(\emptyset_{a}=\rho_{a}C_{a}\left\vert U_{g}\right\vert (U_{g}\cos \phi +U_{g}^{\perp }\sin \phi )\) (air flow stress), \(\emptyset_{w}=\rho_{w}C_{w}\left\vert U_{w}-\mathbf{u}\right\vert [(U_{w}-\mathbf{u})\cos \theta +(U_{w}-\mathbf{u} )^{\perp }\sin \theta ]\) (water flow stress), \(h\) the mean ice thickness, A the ice compactness, \(\mathcal{S}_{h}=[f(h/A)A+(1-A)f(0)]\cdot \chi_{\{h>0\}}\), \(\mathcal{S}_{A}=((f(0)^{+}/h_{0}(1-A)+(-A/(2h))\cdot (\mathcal{ S}_{h})^{-}\), and \(\chi_{E}\) the characteristic function of the set \(E\). They introduce the regularized approximating problem of the preceding system taking \(\varepsilon ,\omega \in (0,1)\): \(m(\partial_{t}\mathbf{u}+\mathbf{u} \cdot \nabla \mathbf{u})+\nabla p=\operatorname{div}\mathbb{S}_{\varepsilon}+\mathcal{F}\), \(\partial_{t}h+\operatorname{div}(h\mathbf{u})=\mathcal{S}_{h,\omega}\), \(\partial_{t}A+\operatorname{div}(A\mathbf{u})=\mathcal{S}_{A,\omega}+A\operatorname{div}\mathbf{u}\cdot \chi_{A}^{\omega}\), with \(\mathbb{S}_{\varepsilon}=p\frac{\nabla \mathbf{u} +\nabla \mathbf{u}^{T}}{\sqrt{\left\vert \nabla \mathbf{u}+\nabla \mathbf{u}^{T}\right\vert^{2}+\varepsilon^{2}}}+p\frac{\operatorname{div}uI_{2}}{\sqrt{\left\vert \operatorname{div}\mathbf{u}\right\vert^{2}+\varepsilon^{2}}}\), \(\mathcal{S}_{h,\omega}=[f(h/(A+\omega ))A+(1-A)f(0)]\cdot \chi_{\{h>0\}}\), \(\mathcal{S}_{A,\omega}=\frac{f(0)^{+}}{h_{0}}(1-A)-\frac{A}{2h}\cdot \frac{\sqrt{ \left\vert \mathcal{S}_{h,\omega}\right\vert^{2}+\omega^{2}}-\mathcal{S}_{h,\omega}}{2}\), \(\chi_{A}^{\omega}=1-\frac{(1-A)^{+}}{(1-A)^{+}+\omega}\). The main result of the paper proves the local in time well-posedness of strong solutions to this regularized approximating problem in \(\Omega = \mathbb{T}^{2}\subset \mathbb{R}^{2}\).\ For given initial data \((u,h,A)\mid_{t=0}=(u_{in},h_{in},A_{in})\in (H^{3}(\Omega ))^{3}\), satisfying appropriate bounds and under regularity and boundedness properties on \(f\), the authors prove the existence of a unique strong solution \((u,h,A)\) to the regularized approximating problem in \([0,T]\times \Omega\), for some \(T\in (0,\infty )\) depending on the initial data, with \(u\in L^{\infty }(0,T;H^{3}(\Omega ))\cap L^{2}(0,T;H^{4}(\Omega ))\), \(h,A\in L^{\infty }(0,T;H^{3}(\Omega ))\), and \(\partial_{t}u,\partial_{t}h,\partial_{t}A\in L^{\infty }(0,T;L^{2}(\Omega ))\). For the proof, the authors introduce another regularized problem for which they prove the existence of a solution using a contraction mapping argument. They prove uniform bounds of the solution with respect to the further parameters and they let these further parameters go to 0.