Influence of Grid Resolution on GIA modelling in Amundsen Sea Embayment over modern and future ice loss

Abstract:

Assessments of future ice-sheet and sea-level change require accurate predictions of glacial isostatic adjustment (GIA). This is particularly true in the vicinity of marine ice sheets, where bedrock uplift and sea level fall along ice-sheet grounding lines may have a significant negative feedback on future ice sheet dynamics (e.g. Gomez et al. 2015; Larour et al., 2019). In regions of laterally varying and low Earth viscosity such as the Amundsen Sea Embayment in the West Antarctic, high resolution GIA modeling may be required to accurately capture elastic and viscous components of the deformation due to localized ice loading changes. Improvements in modelling techniques and observational datasets now allow for km- to sub-km-scale of ice cover changes that serve as input to GIA models. In turn, advances in high resolution GIA modelling allow for capturing short-wavelength bedrock deformation at unprecedented detail, but at heavy computational expense, particularly for global GIA models capable of incorporating 3-D Earth structure.

To address the question of what model resolution is adequate for capturing GIA predictions in the vicinity of ice cover changes, we first conduct sensitivity tests with a suite of numerical grids progressively refined near the load using a finite-volume 3-D GIA model (Latychev et al., 2005). We then focus on assessing the model grid resolution required to accurately capture both the elastic and viscous GIA process due to modern and future ice-sheet changes in the rapidly evolving Amundsen Sea Embayment. Preliminary results indicate that grid resolution has the largest impact along ice margins, and that for most applications, model predictions start to converge at resolutions of ~6 km or less. Our findings suggest that the improvements in accuracy from refining grid resolution are less significant compared to the differences due to factors such as incorporating viscous deformation effects. The preparation ice loading changes that serve as input to the GIA model is also important.