Understanding Antarctic ice-stream flow using ice-flow models and geophysical observations
Lilien, David A
MetadataShow full item record
Ice streams are the primary pathway by which Antarctic ice is evacuated to the ocean. Because the Antarctic ice sheets lose mass primarily through oceanic melt and calving, ice-stream dynamics exert a primary control on the mass balance of the ice sheets. Thus, changes in melt rates at the ice-sheet margins, or in accumulation in the ice-sheet interiors, affect ice-sheet mass balance on timescales modulated by the response time of the ice streams. Even abrupt changes in melt at the margins can cause ice-stream speedup and resultant thinning lasting millennia, so understanding the upstream propagation of marginally forced changes across timescales is key for understanding the ice sheets’ ongoing contribution to sea-level rise. This dissertation is comprised of three studies that use observations and models to understand changes to Antarctic ice-stream dynamics on timescales from decades to millennia. The first chapter synthesizes remotely sensed observations of Smith, Pope, and Kohler glaciers in West Antarctica to investigate the causes and extent of their retreat. These glaciers have displayed some of the largest measured grounding-line retreat, most rapid thinning, and largest speedup amongst Antarctic ice streams. This retreat has drawn interest in their stability both in its own right and as a harbinger of future changes to larger neighboring ice streams. In this study, recent melt rates were determined using flux divergence estimates derived from observations of ice thickness and surface velocity. Out-of-balance melt at the beginning of the study period indicates that the imbalance of this system predates the beginning of satellite velocity observations in 1996. Throughout much of 1996-2010, there was both greater melt over the ice shelves than flux across the grounding line, implying loss of floating ice and elevated melt forcing, and greater grounding-line flux than accumulation, implying adjustment of the grounded ice in response to the ongoing imbalance. The grounding line position of Kohler glacier, and a large melt channel that is unlikely to be a steady-state feature, suggest that the perturbation to this system began on Kohler glacier sometime around the 1970s. Viscosity of the ice shelves, inferred using a numerical model, indicates that weakening of the Crosson ice shelf was necessary to allow the observed speedup, though it is unable to determine whether the weakening was a cause or effect of the ongoing retreat. The second chapter uses a suite of numerical model simulations to determine the dominant drivers of the recent retreat of Smith, Pope, and Kohler glaciers, and extends those simulations that best match observations to evaluate likely future retreat. Similar to the findings of previous studies, the distribution of sub-shelf melt is found to be the primary control on the rate of grounding-line retreat, while the shelf-averaged melt rate exerts a secondary control. The model simulations indicate that, despite ongoing imbalance, the grounding-line position in 1996 was not inherently unstable, but rather elevated melt at the grounding line was required to cause the observed retreat. A weakening of the ice-shelf margins was found to hasten the onset of grounding-line retreat and led to greater speedup. However, without increases in melt beyond 1996 levels, marginal weakening was insufficient to initiate grounding-line retreat. All simulations that capture the observed retreat continue to lose mass until at least 2100, suggesting that ice in this basin may contribute over 8 mm to global mean sea level by 2100. The magnitude of thinning deep in the catchment suggests that the retreat of Kohler and Smith glacier may hasten the destabilization of the neighboring Thwaites glacier catchment. The third chapter uses the timescale of the recently drilled South Pole Ice Core (SPICEcore) and nearby geophysical observations to infer the history of ice flow near the South Pole during the last 10,000 years. The South Pole is located 180 km from the nearest ice divide and drains from the East Antarctic plateau through Academy glacier/Foundation ice stream. As a result, ice flow near the South Pole is potentially affected by the dynamics of these ice streams, and so the history of ice flow in this region has the potential to inform understanding of how marginally forced changes affect the ice-sheet interior. Because the South Pole is far from an ice divide, the accumulation record in SPICEcore incorporates both spatial variations in accumulation upstream and temporal variations in regional accumulation. Comparison between the SPICEcore accumulation record, derived by correcting measured layer thicknesses for thinning, with an accumulation record derived from new GPS and radar measurements upstream, yields insight into past ice flow and accumulation. When ice speeds are modeled as increasing by 15% since 10 ka, the upstream accumulation explains 77% of the variance in the SPICEcore-derived accumulation (vs. 22% without speedup). This correlation is only expected if the ice-flow direction and spatial pattern of accumulation were stable throughout the Holocene. The 15% speedup in turn suggests a slight (3-4%) steepening or thickening of the ice-sheet interior and provides a new constraint on the evolution of the East Antarctic Ice Sheet following the glacial termination.