Investigating the impact of ocean warming on Antarctic ice shelves

• Main Author: Gwyther, DE
• Format: Thesis
• Language: English
• Published: 2016
• Subjects: Antarctica; oceanography; glaciology; computer modelling; basal melting; Antarctic; Buttress; East Antarctic Ice Sheet; East Antarctica; Southern Ocean; The Antarctic; Totten Glacier; Antarc*; Ice Sheet; Ice Shelf; Ice Shelves; Totten Ice Shelf
• Online Access: https://eprints.utas.edu.au/23052/; https://eprints.utas.edu.au/23052/1/Gwyther_whole_thesis.pdf

The Antarctic ice sheet contains enough ice to raise global sea levels by over 50 metres. The stability of the Antarctic ice sheet and subsequently, projections of future sea level rise, depends strongly on interactions with the Southern Ocean. The ice sheet begins to oat around the coastal margins of Antarctica, forming floating ice shelves. These ice shelves buttress the flow of ice sheets into the ocean, and control contributions to sea level rise. Changes to the thickness of ice shelves, such as through melting at the base (basal melting), is thus a critically important factor controlling Antarctic ice discharge, and future sea level. Scientists and field technicians are limited logistically in their ability to observe basal melting in situ and so numerical modelling, such as with the Regional Ocean Modelling System (ROMS), forms an important tool for investigating these critical environments. In order to produce credible modelling results, key uncertainties in modelling must first be addressed. One such uncertainty is the effect of basal roughness on modulating turbulence within the ice{ocean boundary layer, and how the parameterisation of this affects melting and freezing. Using ROMS with an idealised ice shelf geometry, the link between basal roughness and melting and freezing was explored. A rougher interface led to more turbulent exchange and stronger melting or freezing. This showed that the scheme employed to simulate turbulent mixing across the boundary layer was important for simulating refreezing. Lastly and most importantly, employing a spatially-varying drag coefficient, which was rougher for areas of refreezing and smoother for areas of melting, led to variations in the melt/freeze rate across the entire ice shelf. This is strong motivation for geophysical estimation of this parameter and improving ice shelf-ocean models to account for spatial variability in basal roughness. Idealised models were also used to investigate how melting responds to changes in the thermal environment of the ocean cavity. The simulations showed that while circulation is weak in a cold ocean cavity environment, it is the strongest driver of melting. This is opposed to a hot ocean cavity environment, where buoyancy driven circulation is strong but melting is mostly strongly driven by heat availability. These simulations further reinforce that melting is not necessarily primarily driven by heat availability and strongest at the deep, grounding line. Instead, melting can be driven by strong flow, such as tides. The understanding that was further developed through idealised modelling was applied to a realistic scenario: basal melting beneath the Totten Glacier ice shelf, East Antarctica. The Totten Glacier drains a significant portion of the East Antarctic Ice Sheet. The modelled area average basal melt rate was 9.1 m yr-¹, which varies on seasonal and interannual (≈ 6-year) time scales. Furthermore, a causal link was demonstrated between cold, dense water production in the nearby Dalton ice tongue polynya blocking access of warmer off-shelf waters to the Totten ice shelf cavity environment. Several models of the Totten Glacier ice shelf were run, where the forcing was identical but the shape and geometry of the cavity was varied. Comparing the melt rate of different cavity geometries gives us insight into the sensitivity of melting to geometry and internal variability. The melt rate is relatively insensitive to the shape of the cavity, and instead is controlled by the heat flux into the cavity. These model runs were climatologically forced, and yet displayed significant variability. As a result, any glaciological and oceanographic observations of the Totten Glacier ice shelf should be long enough in order to filter this internal variability. Lastly, a simple linear relationship between ocean water on the adjacent continental shelf to melt rate is developed, which suggests that the current contribution to sea level rise from the Totten Glacier outflow is ≈ 0:20 mm yr-¹, which under the RCP8.5 scenario, will rise by 80% by 2100. This thesis improves understanding of the Totten Glacier ice shelf, a possible future cause of significant sea level rise, through explaining the main drivers of melt. Furthermore, these results improve understanding of the uncertainties within ice shelf-ocean numerical models, and help to further constrain projections of sea level change in the future.