Improved modelling of ice-ocean processes

Pine Island Glacier in West Antarctica is currently one of the single biggest contributors to sea-level rise with an estimated volume loss of 1.2mm sea-level equivalent per decade. The loss is caused, at least partly, by ocean currents in the neighbouring Amundsen Sea, which bring increased volumes of warm water to its base. This leads to increased melting of the floating ice shelf, and subsequent speed-up of the entire glacier.

This interaction between ice dynamics and the ocean plays an important role in the future of the Antarctic Ice Sheet, as it affects the health of ice shelves and glaciers all around its margins. As a consequence, computer simulations that predict future mass loss of the Ice Sheet, and associated levels of sea-level rise, should include this important interaction. However, so far, ice dynamics models often use a very crude representation of the melt rates imposed by the ocean, and ocean models do not allow the glaciers to thin or retreat.

In this work we have addressed this issue by coupling a well-tested ocean model that computes the melt rates, to a state-of-the-art ice model that allows glaciers to change. The combination of both models allows us to study feedback processes that have not been observed before, and allows us to assess the validity of previously used approximations for the melt rates underneath Antarctica's ice shelves. We have tested our setup for an idealised representation of Pine Island Glacier, and find that the retreat of the glacier from a bedrock sill sometime before the 1970s is likely irreversible under present-day ocean conditions. Moreover, we find that existing simulations of ice-shelf thinning and glacier retreat overestimate mass loss by more than 40% compared to our coupled simulations.