Modeling the Dynamics of Thwaites Glacier, West Antarctica

摘要

Thwaites Glacier (TG), West Antarctica, has been experiencing rapid mass loss and grounding line retreat in the past few decades. The mass loss of TG is now responsible for 4% of global sea level rise. It is therefore crucial to simulate the future evolution of TG to make projections for future sea level rise. The cause of the dramatic changes is dynamic through the loss of buttressing from its ice shelf due to calving and ice shelf melting. In this thesis, we employ various numerical ice sheet models to study the calving dynamics of TG and the response of TG to enhanced ice shelf melting. We combine a two-dimensional ice flow model with the linear elastic fracture mechanics (LEFM) theory to model crevasse propagation and ice fracturing. We find that the combination of a full-Stokes (FS) model and LEFM produces surface and bottom crevasses that are consistent with the distribution of depth and width of surface and bottom crevasses observed, whereas the combinations of simplified models with LEFM do not. We find that calving is enhanced when pre-existing surface crevasses are present, when the ice shelf is shortened, or when the ice shelf front is undercut. We show that the FS/LEFM combination yields substantial improvements in capturing the stress field near the grounding line of a glacier for constraining crevasse formation and iceberg calving. We then simulate the evolution of TG under different ice shelf melt scenarios and different ice sheet model configurations. We find that the grounding line retreat and its sensitivity to ocean forcing is enhanced when a full-Stokes model is used, ice shelf melt is applied on partially floating elements, and a Budd friction is used. Initial conditions also impact the model results. Yet, all simulations suggest a rapid, sustained retreat along the same preferred pathway. The highest retreat rate occurs on the eastern side of the glacier and the lowest rate on a subglacial ridge on the western side. Combining the results, we find the differences among simulations are small in the first 30 years, with a cumulative contribution to sea level rise of 5 mm, similar to the current rate. After 30 years, the mass loss highly depends on the model configurations, with a 300% difference over the next 100 years, ranging from 14 to 42 mm.