Sea Ice Modeling with an Elastic-Decohesive Constitutive Model
Abstract: Sea ice has an important effect on global climate by reflecting incoming solar radiation and by reducing the heat transfer between the atmosphere and the ocean. Understanding these coupling effects is critical for producing reasonable climate change predictions. On a more local level, predicting ice formation and motion is important for navigation. For both of these applications an efficient and accurate sea ice model is needed. Elements of a sea ice model include a two-dimensional momentum equation derived by integrating the the full three-dimensional equation through the ice thickness, a constitutive model used to calculate the internal stresses from the deformation, a one-dimensional thermodynamic model to calculate the evolution of ice thickness based on uxes at the atmosphere and ocean interface, and an evo- lution equation for the ice thickness distribution to tie the momentum equation to the thermodynamics. Using this framework, I am working with Professor Deborah Sulsky of the UNM Department of Mathematics and Statistics and Professor Howard Schreyer of the UNM Department of Mechanical Engineering to develop a new sea ice model. Our approach differs from standard models in two fundamental ways. First, we use an elastic-decohesive constitutive model, which explicitly predicts the initiation of leads or cracks in the ice and then evolves the leads in time. This model is inherently anisotropic in constrast with the standard isotropic viscous-plastic sea ice constitutive model. Second, we use the Material-Point Method (MPM) to solve the equations of motion. MPM is a numerical method that combines unconnected Lagrangian material points that carry material properties with the ow and a back- ground grid where the momentum equation is solved. The material points model advection naturally and allow the determination of a sharp ice boundary, which are advantages over the standard Eulerian schemes traditionally used in sea ice mod- eling. In this talk I will give an overview of our model, focusing primarily on the elastic-decohesive constitutive model, and show some preliminary results.