Fine-scale ocean processes driving the basal melting of Antarctic ice shelves

The Antarctic Ice Sheet, which comprises the largest volume of ice on our planet, is losing mass due to ocean-driven melting of its fringing ice shelves. Efforts to represent basal melting in sea level projections are undermined by poor understanding of the turbulent ice shelf-ocean boundary layer (ISOBL), a meters-thick band of ocean that regulates heat, salt and momentum transfer between the far field ocean and the ice. Regional ocean models cannot resolve the ISOBL and instead rely on parameterisations to predict melting. However, observations suggest that these parameterisations only perform well for a subset of relevant ocean conditions, namely in cold, energetic environments. This thesis uses both observational data and turbulence-resolving model simulations to address this shortfall by characterising melting and ISOBL dynamics across a broad range of ocean states. Chapter 1 of this thesis outlines the motivation and context for the work that follows, highlighting the urgent knowledge gaps that will be addressed. In chapter 2, a unique set of observations from beneath the Amery Ice Shelf, including in situ basal melt rate, ocean velocity, temperature and salinity data are analysed and ocean conditions are characterised. The mean basal melt rate at the site (0.5 m yr\\(^1\\)) is a factor of 2-4 lower than predicted from common ice-ocean parameterisations. This result suggests an important role for stratification at this site, either through suppression of heat transport to the ice-ocean interface or a shoaling of the mixed layer depth. These processes cannot be unraveled from the available observational data, further motivating the need for turbulence resolving simulations. In chapter 3, large-eddy simulation is used to model the ISOBL. The model domain consists of a horizontal ice-ocean interface with a melting boundary condition at the upper surface, underlain by a stratified ocean. The domain is periodic in both horizontal directions, and is forced with a steady flow in geostrophic balance. At relatively warm, low velocity conditions a small-scale mixing process (double-diffusive convection) is shown to determine ice shelf melt rate and the properties of the mixed layer that forms beneath the ice. In double-diffusive regime, melting is found to be inherently unsteady in time and insensitive to shear from the imposed current. Simulated melt rates and water column structure are consistent with observations made near the grounding line of the Ross Ice Shelf. In chapter 4, model forcing conditions are expanded to encompass colder and more energetic cavity environments in which current shear controls melting. Two distinct mixing regimes emerge: a stratified regime in which boundary layer turbulence is strongly affected by the surface buoyancy flux due to melting and a well-mixed regime in which buoyancy has little effect. The stratified regime supports strong temperature and salinity gradients near the ice, decoupling the interface and far field conditions. The relative strength of the surface buoyancy flux and shear, characterised by the Obukhov length scale, is shown to be critical to both heat flux and boundary layer depth. Results from chapters 3 and 4 are used to develop a regime diagram for ISOBL dynamics beneath horizontal, melting ice in discussion chapter 5. This novel diagram provides new insight into the varied and nonlinear responses of basal melting and ISO BL dynamics to local conditions around Antarctica. Comparison to observed sub-ice shelf conditions and melt rates from chapter 2 and other published studies is favorable and demonstrates the relevance of these regimes over a broad range of realistic conditions. Insights from this thesis significantly extend the current understanding of the ISOBL and basal melting. The inclusion of the double-diffusive and stratified regimes in future parameterisations of ice-ocean interactions will significantly improve melt rate estimates, with consequences for predictions of ice sheet stability.