Modelling Antarctic ice shelf-ocean interaction at high-resolution

Understanding the processes involved in basal melting of Antarctic ice shelves is important to quantify the rate at which Antarctica will lose mass in the future. We implement a state-of-the-art ocean model to derive a new estimate of continent-wide ice shelf basal melting at 2 km horizontal resolution, then explore the underlying oceanic mechanisms. A new circum-Antarctic ice shelf-ocean model is developed, which is based on the Regional Ocean Modeling System (ROMS). Improvements over previous models of this scale are the inclusion of tides at a horizontal resolution of 2 km that is su_cient to resolve critical onshelf heat transport by bathymetric troughs and eddies. For this study we choose to run the model with conditions from the year 2007 to represent nominal present day climate. Results agree with available observations of tides, ocean state and ice shelf-ocean interaction, giving us con_dence in modelled mass loss estimates. We estimate a total basal mass loss of 1207 Gt/yr of which 79 % comes from ice shallower than 400 m deep. Ice shallower than 200 m deep contributes 33 % to the total mass loss and this ice is often con_ned to the ice front, areas which are not resolved by methods using satellite data. Modelled melting at depths shallower than 200 m triples in summer, when solar heated surface waters advect under the ice. These results suggest that changes in the surface ocean will need to be considered when constraining the future Antarctic mass loss and not only deep warm water intrusions. Further, we estimate the impact of tides on Antarctic ice shelf melting and the continental shelf ocean using a downscaled version of the model with 4 km horizontal resolution. Tides contribute 57 Gt/yr (4 %) to the continent-wide ice shelf basal mass loss, while continental shelf temperatures are reduced by tidal exchange, on average by about 0.04 °C. This indicates that tides act to increase the efficiency at which ocean heat is converted to melting. Regional variations are larger, with melt rate modulations exceeding 500 % and temperatures changing by more than 0.5 °C. In particular, tides tend to heat the Weddell Sea continental shelf with warm waters reaching under the Ronne Ice Shelf and increasing its area-averaged basal melt rate by 150 %. Finally, we explore the tidal processes that cause variations in melting using singular spectrum analysis over periods of up to one month. At most places along the ice base, friction velocity varies at tidal timescales (one day or faster), while thermal driving changes at rates slower than one day. In some key regions under the large cold-water ice shelves, however, thermal driving varies faster than friction velocity and this can not be explained by tidal modulations in boundary layer exchange rates alone. These results show that large scale ocean models aiming to predict accurate ice shelf melt rates will need to explicitly resolve tides. This thesis shows that kilometre-scale processes on the Antarctic continental shelf that are related to the surface ocean and tides exert large controls on the mass balance of the Antarctic Ice Sheet. Global circulation models that aim to provide projections of future climate, will need to capture these processes in order to accurately predict sea level rise resulting from changes in Antarctic basal mass loss.