The effect of late holocene ice-mass changes on glacial isostatic adjustment in West Antarctica

Abstract

Glacial isostatic adjustment (GIA) describes the Earth’s response to changing ice and water loads as ice sheets grow and diminish. GIA is difficult to model in Antarctica due to limited knowledge of ice history and Earth properties. The signal confounds satellite gravity measurements of present-day ice-mass change and needs to be accurately removed, but remains the biggest uncertainty. One problem with current Antarctic GIA models is that they neglect ice-mass changes over the past few thousand years, which, in regions of low viscosity mantle, may dominate the present-day bedrock uplift. This study investigates deficiencies in millennial-scale GIA models arising from omission of Late Holocene and present-day ice-mass changes. In the Antarctic Peninsula increasing accumulation observed in ice cores since the 1850s has been shown to cause loading and present-day GIA-related subsidence, although results are dependent on the Earth model. This missing signal may help to reconcile the misfit between GIA model predictions and GPS-observed uplift. GPS records from the northern Peninsula provide an opportunity to place bounds on the regional Earth properties. Since 1995 several ice shelves have collapsed triggering ice-mass unloading that invokes a solid Earth response. However, non-linear GPS-observed uplift cannot be explained by elastic deformation alone. Using a viscoelastic model to predict uplift due to recent ice loss and testing the fit to GPS time series, an Earth model has been constrained with upper mantle viscosity much lower than previously suggested. Elsewhere, the stagnation of Kamb Ice Stream on the Siple Coast ~165 years ago has caused localised thickening of ice which may cause significant GIA-related subsidence if the regional mantle viscosity is low. Combining with an LGM deglacial history and comparing with an empirically-derived GIA model shows large misfits, indicating that the regional mantle viscosity is high and highlighting potential errors in the LGM deglacial model.