AISS-mélange. Antarctic ice sheet stability: ocean-ice-mélange model to reduce uncertainties

NAISS 2023/5-427


NAISS Medium Compute

Principal Investigator:

Lars Arneborg



Start Date:


End Date:


Primary Classification:

10503: Geosciences, Multidisciplinary

Secondary Classification:

10509: Oceanography, Hydrology and Water Resources

Tertiary Classification:

10599: Other Earth and Related Environmental Sciences



The ultimate goal of this project is to enhance our comprehension of the processes occurring at the ice-ocean interface below ice shelves, which become more fractured. This understanding will contribute to the reduction of uncertainties in predictions related to rising sea levels. In recent decades, ice shelves, which have a butressing role in the stability of Antarctica, have experienced tremendous changes in their thickness, melt pattern or geometry but have also, in general, gradually build damage (more surface and basal crevasses). This leads to dynamic vulnerability and in the worse case scenario to the complete disintegration of the ice shelf into ice melange (a mixture of icebergs and sea ice) as it is the case for the western part of Thwaites glacier, in the Amundsen Sea. This region has been the focus of this project as the ice shelves (including Pine Island Ice Shelf and Thwaites Ice Shelf) are the most sensitive to climate change so far. Melt rate is governed by the so called “three equation parameterisation”, which depends on the shear between the ice and the ocean depending on the velocity a a certain distance from the ice and a coefficient of drag usually kept constant. In NEMO, it is calculated in a “Losch layer”, which is typically equal to 20-30 and can span a few cells depending on the vertical resolution. A more physical way of tackling this boundary layer problem would be to use the law-of-the-wall, which relates the shear to the roughness length (small-scale roughness). Unfortunately very few observations are available for this variable. The question(s) I want to answer is (are): Does changing roughness in time and space below ice shelves matter for ice melt? This small-scale roughness has a direct impact on the ocean drag but what about the larger-scale roughness? Does the assumption of a constant coefficient of drag spatially and temporally is sound enough? In the first study, we discuss the usefulness of a spatially variable coefficient of drag depending on the topography and calculated in the 1st wet cell rather than in the Losch layer. In a second study, I would like to compare the new parameterisation to an implementation of a form drag parameterisation from sea ice into ice shelves. Finally, I want to start testing the new parameterisation in future scenarios offline-coupled with climate and ice sheet models.