Modelling and Parameterisation of Ice-Shelf Melting in the Amundsen Sea, Antarctica

Abstract. Ice loss from the Amundsen Sea sector of West Antarctica is a major contributor to global sea-level rise, and is a key source of uncertainty in projections of sea level over the coming centuries. This ice loss is ultimately driven by changes in ocean melting, which must therefore be represented in ice-sheet model forecasts. In this study we use high-resolution ocean simulations to understand the mechanisms controlling ice-shelf melting in the eastern Amundsen Sea. Melting is focussed on four ‘hot spots’ of melting of the deep ice where the main glacier trunks cross the grounding line. Secondary areas of elevated melting occur beneath the associated buoyant ‘meltwater outflows’, which are guided by ice topography and Coriolis force. The simulations are then used to test simple local parameterisations of melting. The best parameterisation expresses melt rate as a simple function of ocean temperature to the power 3/2, ice slope to the power 1/2, and tapered to zero near the grounding line. This matches the simulated melting with r²=0.65, capturing melting hot spots near the grounding line but failing to represent melting along meltwater outflow paths. This parameterisation also broadly captures the strong melting feedbacks that appear when the model is applied to possible future ice geometries. It is possible that simple local melting parameterisations may be sufficient wherever ice shelf buttressing is focussed near the grounding line (such as Thwaites Glacier), but may be inadequate in regions where melting beneath shear margins controls buttressing (such as Pine Island Glacier).