Ocean temperature forcings in glacial-interglacial Antarctic Ice Sheet simulations

Ice shelf basal melt accounts for about half the present-day ice loss from the Antarctic Ice Sheet, and is important for both ice sheet mass balance and as a source of fresh water into the Southern Ocean. In Antarctic Ice Sheet simulations over Quaternary glacial cycle time scales, neither basal melt rate nor its principal oceanographic controls (temperature and salinity of waters adjacent to ice shelves) can be reconstructed directly from proxy records. Given the strong ice-ocean-atmosphere interactions, the ideal solution is a coupled ice-ocean-atmosphere model, but computational demands currently limit this approach to short time scales. Stand-alone ice sheet simulations can cover much longer time scales at reasonable resolution, but require an alternative estimate of ocean temperatures. Here we compare the strengths and weaknesses of three options: (i) proxy reconstructions of North Atlantic and circumpolar deep water temperatures, from marine sediment cores north of 43°S; (ii) an ice sheet air temperature reconstruction, damped and lagged by a linear response function; and (iii) a glacial index method which interpolates between CMIP6 lig127k (interglacial) and lgm (glacial) end-member ocean states. We find considerable differences in the rates and magnitudes of the Antarctic Ice Sheet's contribution to past sea-level changes when applying the three methods in simulations over the last two glacial cycles, particularly during the last interglacial and Holocene. Therefore, the ocean temperature forcing remains as an important but poorly-constrained modelling choice, whether investigating past warm climates or using long simulations as a spin-up for future projections.