The influence of marine ice on ice shelf dynamics and stability

To predict the future behaviour of the Antarctic Ice Sheet, it is vital to have a sound understanding of ice rheology, which governs how ice deforms in response to stress. The rheology of laboratory-made pure water ice (‘standard ice’) has been the focus of many previous studies, and is relatively well constrained. However, ice in nature differs from standard ice in its chemical content and microstructural characteristics, which can substantially alter its rheology in ways which are not well understood. In this thesis, I investigate the rheological properties of marine and meteoric ice from the Amery Ice Shelf, Antarctica, comparing them with standard ice. Then, I incorporate these findings into a numerical model of an ice shelf, to understand how differences in ice properties, particularly in marine ice, can influence predictions of future ice shelf dynamics. In the first chapter of this thesis, I perform a systematic study to test the viability of a new laboratory technique to substantially reduce the time required to conduct long term ice deformation experiments. I deform samples of standard ice under uniaxial compression at -2 ^◦C before lowering the temperature to either -7 ^◦C or -10 ^◦C, and compare the results to experiments run constantly at the lower temperature. The tertiary strain rate values and microstructural characteristics measured in the changing-temperature experiments are indistinguishable from those measured in constant-temperature experiments, and experiment run-time is reduced by 55%. As time is a major limitation on studies of the rheological behaviour of ice at stresses and temperatures approaching in situ conditions, this is a significant step forward.
In the second chapter, I perform compressional deformation experiments on dozens of samples of marine, meteoric and standard ice at -2 ^◦C (the approximate in situ temperature of marine ice in the Amery Ice Shelf) and -7 ^◦C (the approximate temperature of the boundary between marine and meteoric ice). These data are analysed to extract secondary and tertiary creep rates, and microstructural information including grain size, crystallographic preferred orientation (CPO) and grain sphericity distribution at a range of strains. This is the most comprehensive dataset of rheological and microstructural information ever collected from experiments in polar ice. I find that tertiary strain rates in marine ice are higher than those in standard ice deforming under the same stress and temperature conditions. This difference is greater at -2 ^◦C, likely due to the presence of liquid brine on the grain boundaries affecting the kinetics of dynamic recrystallisation and dislocation movement. As marine ice can make up a large proportion of an ice shelf (up to 40% of the total ice shelf thickness in some areas), this difference in rheology could have a significant effect on the dynamics of an ice shelf.
Finally, in the third chapter, I put my experimental results into context by testing them in a numerical ice shelf model. I use the Ice-sheet and Sea-level System Model (ISSM) to model an idealised, rectangular ice shelf with a marine ice basal layer. I perturb the thickness of the marine ice layer, the thermal profile of the shelf, and the rheological properties of the marine ice, to investigate the effects of these properties on ice shelf dynamics. Scenarios with a warm (-2 ^◦C) and soft (based on measured rheology) isothermal marine ice layer, similar to that seen in the Amery Ice Shelf, have substantially higher ice mass flux across the calving front than those without a marine ice layer. For those scenarios with a warm and soft marine ice layer, changes to the thermal profile have a factor of 10 greater effect on ice mass flux than changes to the ice rheology.
This thesis systematically examines and characterises the rheology of natural shelf ice, particularly marine ice, and contextualises the findings by incorporating them into a numerical model to assess the scale of their importance for use in ice sheet modelling. I have demonstrated that marine ice has substantially different properties to those which are currently incorporated in ice sheet models, and that overlooking those differences can result in significant underestimates of ice mass flux from the Antarctic Ice Sheet in the future.