Processes controlling the calving regime of the S‚àö‚àèrsdal ice shelf, East Antarctica

The calving of icebergs from Antarctic ice shelves is one of the primary ways in which mass is transported from the continent to the ocean. As a result, the mechanisms that determine how and when icebergs are produced also determine the rate at which this mass transfer occurs. These mechanisms also influence the buttressing potential of ice shelves and in doing so can contribute to ice shelf change, which is likely the largest and most uncertain aspect of Antarctica's contribution to sea level rise. The mass lost from ice shelves over time due to the calving of icebergs is known as the calving flux. Baseline observations of ice shelf calving flux can be used in conjunction with an understanding of the processes involved to identify changes in ice shelf behaviour. Relatively little research has focused on the calving regimes of the smaller ice shelves that fringe the Antarctic continent, and how they may differ from larger ice shelves. In an attempt to address this knowledge gap, the work presented in this thesis focuses on the calving regime of the small ice shelf that is fed by the Sørsdal Glacier, in East Antarctica. To derive a comprehensive understanding of the Sørsdal ice shelf's calving regime this work quantifies the mass loss due to calving, identifies the processes involved in calving and investigates the applicability of a number of calving laws to this particular ice shelf. In order to determine whether the calving flux of an ice shelf is changing, it is first essential to determine the baseline calving flux. Methods used to estimate the calving flux either take into account non-steady-state behaviour by capturing movement of the calving front location (e.g. using satellite observations), or they assume the calving front is stationary and that the ice is in steady state (e.g. flux gate methods). Non-steady-state methods are hampered by the issue of temporal aliasing, i.e. when the satellite observation frequency is insufficient to capture the cyclic nature of the calving front position. Methods that assume a steady state to estimate the calving flux accrue errors if the ice shelf is not in steady state. In order to overcome these limitations a new approach is proposed and implemented here that combines a time-series of calving front locations with a flux gate method. The approach involves the creation of a unique semi-temporal domain as a mechanism to overcome the issue of temporal aliasing, and only requires readily available datasets of ice thickness (when using ice thickness estimates derived from surface elevation values and the assumption of hydrostatic equilibrium), as well as surface velocity estimates. This approach allows for complex calving front geometries and captures calving events of all sizes that are visible within the satellite images. Application of the approach to the Sørsdal ice shelf revealed the long-term average calving flux to be 0.74 ± 0.105 Gt/yr. The temporal baseline separating the satellite images has a strong influence on the resulting calving flux estimate. It is critical that an appropriate temporal baseline is used to estimate the long-term mean calving flux: if the baseline is too short the estimated calving flux may not be representative of the long-term average. The estimated minimum temporal baseline needed to produce a representative estimate of the long-term average calving flux for the Sørsdal ice shelf is 7.6 years. The variation in the location of the calving front suggests that between 1972 and 2019, the pattern of calving from the Sørsdal ice shelf has changed. To further examine the variation in the calving front location, the following section of this thesis then focuses on the mechanisms involved in the production of icebergs from the Sørsdal ice shelf. Surface elevation data were used in conjunction with a time-series of visible satellite imagery to perform a glaciological structural analysis, identifying the presence of rifts, surface and basal crevasses as well as basal channels in the Sørsdal ice shelf. Analysis of the characteristics of these features, how they vary over time and how they interact, allowed the development of a conceptual calving model. This model shows that glaciological features, in particular basal crevasses, are the primary driver of the pattern of calving at the Sørsdal ice shelf and that there exists an important relationship between rifts and basal crevasses on this ice shelf. These glacial features were analysed for change and compared to the environmental drivers of air temperature observations (as a proxy for surface melt) and sea ice observations (as a proxy for mechanical oceanic drivers). This analysis was performed in order to determine how these glacial features may have affected the calving front position of the Sørsdal ice shelf. However, the proxies were found to have no significant correlation with the calving front location. Instead, analysis of the glaciological structures over time identified a change in basal channel shape, which was linked to the propagation of basal crevasses and hence the stable calving front position. To determine how well the glaciological interactions and calving front variability are represented by calving laws in a numerical model, three commonly-used approaches were applied to the Sørsdal ice shelf. Calving laws can be used within numerical models to either predict the calving rate or to predict the calving front location. The three laws that were investigated include eigencalving, a continuum damage mechanics-based calving law and a modified crevasse depth criterion. The eigencalving law was used to investigate the calving rate, whereas the damagebased calving law and the crevasse depth calving criterion were investigated for their ability to predict the calving front location. A domain was created for the Sørsdal ice shelf within the Ice Sheet System Model running Shallow Shelf Approximation flow equations. Application of the eigencalving law to the Sørsdal ice shelf using a calving rate that was equal to the observed velocity field produced a median proportionality factor of 1.04 x 10\\(^7\\) m/yr, which is well within the range of published values for Antarctic ice shelves. The proportionality factor was found to be very sensitive to the region of the ice shelf from which it was obtained. Regions of high damage aligned with regions of high fracture depth derived from the modified crevasse depth model, which match the spatial pattern of surface crevasses visible in satellite imagery. However, surface crevasses present in two along-flow crevasse bands were not captured, nor were basal crevasses present at the calving front. The results of the crevasse depth calving law may have been adversely influenced by pinning points that were unaccounted for in the model geometry, and the consequent incorrect assumption of hydrostatic equilibrium. The damagebased calving law results may have been influenced by the advection of damage. Neither the fracture nor the damage maps identified a new position of the calving front, suggesting that the ice shelf front is stable and unlikely to shift up-glacier from its current position. However, these findings require the assumption that the modelled fracture and damage distributions are representative of the fracture distribution across the Sørsdal ice shelf. The work outlined in this thesis highlights the important role that basal crevasses play in the process of calving. Application of three calving laws found that the results were very sensitive to the input datasets (particularly the ice thickness and ice temperature fields), as well as to the constants used within the rigidity inversion. Neither the crevasse depth calving law nor the damage-based calving relation explicitly captured the influence of basal crevasses, however this was likely a result of model geometry issues and deficiencies of the ice flow model used. Another important finding of this work is that the temporally dynamic interactions between glacial features show that an ice shelf calving regime is not static, but can change even without obvious climatic drivers. The outcomes of this work emphasise the need to include smaller ice shelves in studies of Antarctic calving mechanisms as they offer the opportunity to expand the observed range of calving mechanisms, and are a useful tool for assessing calving model performance.