Articles | Volume 11, issue 1
Research article
27 Jan 2017
Research article |  | 27 Jan 2017

Comparison of hybrid schemes for the combination of shallow approximations in numerical simulations of the Antarctic Ice Sheet

Jorge Bernales, Irina Rogozhina, Ralf Greve, and Maik Thomas

Abstract. The shallow ice approximation (SIA) is commonly used in ice-sheet models to simplify the force balance equations within the ice. However, the SIA cannot adequately reproduce the dynamics of the fast flowing ice streams usually found at the margins of ice sheets. To overcome this limitation, recent studies have introduced heuristic hybrid combinations of the SIA and the shelfy stream approximation. Here, we implement four different hybrid schemes into a model of the Antarctic Ice Sheet in order to compare their performance under present-day conditions. For each scheme, the model is calibrated using an iterative technique to infer the spatial variability in basal sliding parameters. Model results are validated against topographic and velocity data. Our analysis shows that the iterative technique compensates for the differences between the schemes, producing similar ice-sheet configurations through quantitatively different results of the sliding coefficient calibration. Despite this we observe a robust agreement in the reconstructed patterns of basal sliding parameters. We exchange the calibrated sliding parameter distributions between the schemes to demonstrate that the results of the model calibration cannot be straightforwardly transferred to models based on different approximations of ice dynamics. However, easily adaptable calibration techniques for the potential distribution of basal sliding coefficients can be implemented into ice models to overcome such incompatibility, as shown in this study.

Short summary
This study offers a hard test to the models commonly used to simulate an ice sheet evolution over multimillenial timescales. Using an example of the Antarctic Ice Sheet, we evaluate the performance of such models against observations and highlight a strong impact of different approaches towards modeling rapidly flowing ice sectors. In particular, our results show that inferences of previous studies may need significant adjustments to be adopted by a different type of ice sheet models.