Pine Island Glacier (PIG) is one of the fastest flowing outlets in the Antarctic Ice Sheet. In recent decades the glacier has undergone substantial changes including speed up, thinning and grounding-line retreat. Between the 1970s and early 1990s, despite comprising just 1.5 % of the overall area, PIG contributed almost 13 % of the total mass loss, adding 0.34 mm to global mean sea level rise. Situated in the Amundsen Sea Embayment in West Antarctica, it has long been considered that PIG is close to an unstable state owing to its underlying bed profile. This has prompted fears about its future and the fate of the wider region. Historic records show the glacier underwent a period of mass loss and retreat after a 1940s climate anomaly, causing PIG to unground from a long term position on a subglacial ridge. The initial cause of retreat and the ongoing transition until the 1990s remains poorly understood. The aim of this thesis is to gain a better understanding of the transient evolution of PIG after it retreated from the ridge 80 years ago, and determine whether it was subjected to an instability mechanism.

In the first research chapter, an ice flow model is used to investigate how the behaviour of an idealized buttressed ice sheet is affected by changes in external forcing. It is found that ocean induced sub shelf melting leads to thinning and retreat of the ice sheet from a stable position on a retrograde slope. The choice of basal friction within the model impacts the rate of mass loss, with the commonly used Weertman law causing the least retreat. Further experiments show that the ice sheet undergoes a hysteresis in response to changes in ice softness. In the second research chapter, ice flow modelling results show that a sustained increase in basal melting beneath PIG ice shelf leads to a loss of buttressing, which is sufficient to cause a retreat from the subglacial ridge to an upstream ice plain. The formation of glacial features prior to and after the retreat, together with the evolution of mass loss are all in agreement with observations. A stability analysis indicates that an unstable region is present upstream of the ridge suggesting PIG crossed a tipping point in its retreat by the early 1970s. In the third research chapter, further modelling of PIG shows that a short period of increased basal melt is sufficient to cause irreversible retreat from the ridge to the next bed high point; a much greater decrease in forcing, compared to the initial perturbation, is needed to stop and reverse the retreat. Once there is sufficient melting upstream of the ridge, after a connection is established between the inner and outer cavities, only extreme cold conditions, or near-zero melt, can stop the retreat to the next bed high point.

The results in this thesis suggest that PIG could have experienced an unstable response to a warm anomaly in the 1940s. Periods of cool ocean conditions were unable to affect the retreat until it stabilized at an ice plain in the 1990s. We provide new insight into how future glaciers in the Amundsen Sea may respond to more prevalent warm conditions that have been predicted for the region.