Modelling the Antarctic ice sheet under warm Pliocene climates

Reconstructions of warm Pliocene sea level imply a large but uncertain contribution from the Antarctic ice sheet. Their high end implies collapse of the West Antarctic ice sheet, and mass loss from East Antarctic subglacial basins. Previous work has suggested that extensive East Antarctic retreat is facilitated by the marine ice cliff instability – where loss of ice shelves and mechanical failure of unbuttressed cliffs drives retreat. We explored the Antarctic ice sheet under warm Pliocene climate using BISICLES, a relatively higher-order physics ice sheet model, capable of simulating grounding line and ice stream dynamics at relatively high resolution. We used perturbed parameter ensemble experiments to explore uncertainty in Pliocene Antarctic retreat, without the marine ice cliff instability. We simulated the most dynamic regions of the ice sheet at relatively high resolution, down to 4 km in ice streams and at the grounding line . To explore uncertainties in basal sliding, surface mass balance processes, bedrock-ice sheet interactions and ice shelf basal melt sensitivity to ocean forcing, we ran a 120-member Latin Hypercube perturbed parameter ensemble. Climate forcing was based on results from four different PlioMIP2 climate models, which the ensemble was evenly divided across. This allowed us to explore the role of climate model choice in simulated Pliocene Antarctic configuration. Additional experiments explored parametric uncertainty under a modern climate, and initial condition uncertainty. We compared ensemble results to a reconstructed Pliocene sea level range, as well as geochemical provenance based reconstructions of grounding line retreat into the Wilkes subglacial basin (East Antarctica) – which some ensemble members successfully reproduced. We simulated a large Antarctic sea level contribution range, which was wider than the reconstructed Pliocene sea level contribution range. Moreover, simulated sea level contribution was highly sensitive to a perturbed basal sliding parameter. We show that our modelling framework can simulate the magnitude and location of Pliocene Antarctic mass loss consistent with proxy-based reconstructions under some modelling choices, however uncertainties remain in the representation of basal sliding.