Collaborative Research: Evaluating Retreat in the Amundsen Sea Embayment: Assessing Controlling Processes, Uncertainties, and Projections

点击翻译按钮获取中文摘要

Accurate reconstructions and predictions of glacier movement on timescales of human interest require a better understanding of available observations and the ability to model the key processes that govern ice flow. The fact that many of these processes are interconnected, are loosely constrained by data, and involve not only the ice, but also the atmosphere, ocean, and solid Earth, makes this a challenging endeavor, but one that is essential for Earth-system modeling and the resulting climate and sea-level forecasts that are provided to policymakers worldwide. Based on the amount of ice present in the West Antarctic Ice Sheet and its ability to flow and/or melt into the ocean, its complete collapse would result in a global sea-level rise of 3.3 to 5 meters, making its stability and rate of change scientific questions of global societal significance. Whether or not a collapse eventually occurs, a better understanding of the potential West Antarctic contribution to sea level over the coming decades and centuries is necessary when considering the fate of coastal population centers. Recent observations of the Amundsen Sea Embayment of West Antarctica indicate that it is experiencing faster mass loss than any other region of the continent. At present, the long-term stability of this embayment is unknown, with both theory and observations suggesting that collapse is possible. This study is focused on this critical region. We will test an ice-sheet model against existing observations, improve treatment of key processes in the model, and make projections with uncertainty assessments.This is a three-year modeling study using the open-source Ice Sheet System Model in coordination with other models to improve projections of future sea-level change. Project goals are to:1. hindcast the past two-to-three decades of evolution of the Amundsen Sea Embayment sector to determine controlling processes, incorporate and test parameterizations, and assess and improve model initialization, spinup, and performance;2. improve the model by utilizing sensitivity studies with regional process-oriented models to create numerically efficient parameterizations for key sub-grid-scale processes;3. project a range of likely evolutions of the Amundsen Sea Embayment sector and their respective contributions to sea level in the next several centuries;4. attribute sources of errors in the hindcast and provide an assessment of the uncertainties in the projections, including a range of likely outcomes given various forcings and inclusion or omission of physical processes in the model.At present, the long-term stability of the Amundsen Sea Embayment is unknown, with both theory (the "marine ice sheet instability hypothesis") and observations (rapid thinning and grounding-line retreat approaching regions where the bed deepens inland) suggesting that collapse is possible. But incompletely understood physical processes (e.g., basal hydrology, rheology, and sliding; tidal effects; ice-ocean interaction along the shelf and within the grounding zone) and lack of resolution in basal topography datasets making the ultimate outcome uncertain. Thus, there is a pressing need for high-resolution simulations of this region that include numerical representations of controlling physical processes (many of which are applicable elsewhere) within a higher-order ice-sheet model capable of assimilating recent observations and providing uncertainty analyses associated with model and data limitations. By focusing on the Amundsen Sea Embayment as a connected region across the 10-10,000-meter scales using a hierarchy of one, two, and three-dimensional models along with the sensitivity analysis tools built into the Ice Sheet System Model, this project aims to produce (1) the most reliable results to date when compared with studies that consider only one ice stream or the entire ice sheet and (2) estimates of differing dynamic responses arising from errors in data, model parameterizations, and forcings. Given the uncertainties, the project will produce a range of predictions with characteristic trends that can be recognized within future observational data sets. As new data become available, some predicted rates of change could be culled from the predictive paths generated by this study.