| Summary: | Thesis (Ph.D.)--University of Washington, 2020 Ice mass loss from Antarctica is the main source uncertainty in projections of sea level rise. In Antarctica, the rate at which grounded ice discharges into the ocean and raises sea level is largely controlled by ice shelves, the floating extensions of the ice sheet. Ice shelves provide resistance to upstream grounded ice flow in a process known as buttressing, and reductions in buttressing are responsible for the majority of Antarctic ice mass loss and sea level contribution. Current reductions in buttressing are largely associated with ice shelf thinning from basal melt. However, ice shelf fracture processes such as mechanical weakening and tabular calving also decrease buttressing, and ice shelves become more prone to fracture as they thin. These fracture processes are poorly understood and parameterized in projections of ice sheet evolution. In this dissertation, I present a new modeling framework for large-scale shallow shelf ice flow and damage that can be used to better project ice sheet evolution and study these ice shelf fracture processes. In the first of the three studies that comprise this dissertation, I introduce a generalized interpolation material point method (GIMPM) for the shallow shelf approximation (SSA) of ice flow and thickness evolution. The GIMPM is a hybrid Eulerian-Lagrangian framework for large-deformation simulations that is well suited for coupling with the damage model that is employed in the subsequent studies, and can be viewed as a particle variation of the finite element method. The primary advantages of the GIMPM-SSA framework are computationally efficient advection of history or internal state variables without artificial diffusion errors and automated tracking of the ice front and grounding line at sub-element scales. By avoiding artificial diffusion errors, the GIMPM-SSA framework is ideal for capturing sharp features such as rifts in damage studies. I demonstrate the accuracy and stability of the GIMPM-SSA framework using ...
|
|---|