Assessing sub-shelf melting parameterisations with the NEMO-Elmer/Ice coupled model
The purpose of this new paper published in GMD is to assess the sub-shelf melting parameterisations that are more or less commonly used in Antartic ice-sheet modelling. In West Antarctica, the floating ice shelves fringing the continent are currently thinning, mainly because of increasing sub-ice-shelf melting coming from an oceanic origin. The consequence is less ice-shelf buttressing undergone by the upper part of the ice sheet that is thus accelerating and discharging more ice to the ocean, therefore increasing sea level rise. Representing sub-ice-shelf melting in ice-sheet models is thus crucial to future sea-level projections. This can be done by coupling your favourite ice sheet model, say Elmer/Ice, to your favourite ocean model, say NEMO, but the computational cost is likely to be too large to model the whole Antarctic ice sheet over the next century. To deal with computational cost, a solution is to parameterised the relation between oceanic properties (temperature and salinity) and sub-ice-shelf melting. So far, various parameterisations have been proposed in the litterature but they have never been compared to (i) each other and (ii) to the results of a coupled model, which is what we have done in this paper.
The melting parameterisations range from simple scalings with far-field thermal driving to emulators of box and plume models, which are more complex parameterisations. In total we have assessed 19 types of melting parameterisations, forced by 6 ocean temperature and salinity scenarios over the next century, using an ideal ice-sheet/ice-shelf system resembling the Pine Island Glacier in West Antarctica. All these simulations have been compared to a small ensemble (4 members) of the NEMO-Elmer/Ice coupled model, for which different aspects of the ocean model have been changed. Figure shows a clear difference between the melting patterns obtained from the coupled simulations in the one hand and the parameterisations in the other hand. For instance, the coupled model yields relatively more melting at the grounding line and less melting when getting closer to the calving front where the ice is thinner. It is however difficult to anticipate the different responses of the ice sheet to ocean scenarios, because we don't really know how it is affected by ice-shelf buttressing distribution. Thus, one of the best ways to evaluate the parameterisations is to use the time evolution of ice loss.
To quickly sum up the results of the study, the plume parameterisation underestimates the contribution to sea level when forced by the warming scenarios. The box parameterisation compares fairly well to the coupled results in general and gives the best results using five boxes. For simple scalings, the comparison to the coupled framework shows that a quadratic dependency to thermal forcing is required, as opposed to linear. In addition, the quadratic dependency is improved when melting depends on both local and nonlocal, i.e. averaged over the ice shelf, thermal forcing. The results of both the box and the two quadratic parameterisations fall within or close to the coupled model uncertainty.
Read More: Favier L., N.C. Jourdain, A. Jenkins, N. Merino, G. Durand, O. Gagliardini, F. Gillet-Chaulet and P. Mathiot, 2019. Assessment of sub-shelf melting parameterisations using the ocean–ice-sheet coupled model NEMO(v3.6)–Elmer/Ice(v8.3). Geosci. Model Dev., 12, 2255-2283, doi:10.5194/gmd-12-2255-2019