Modelling climate-ice sheet interactions during the Last and Penultimate Deglaciations (~21-9 ka and ~140-128 ka)

The Penultimate Deglaciation (PDG; ~140–128 thousand years ago; ka) is the transition from the Penultimate Glacial Maximum (PGM; ~140 ka) to the Last Interglacial (LIG; ~129–116 ka). The LIG experienced warmer temperatures than present day and it was the last time in Earth’s history that sea levels were higher than todays, due to larger contributions from the Greenland and Antarctic ice sheets. Understanding the mechanisms that led to their retreat is important for accurately projecting future sea level rise, because large uncertainties remain in how the ice sheets will respond to a warming climate. The climate and ice sheets at the LIG were still responding to changes that occurred during the PDG, yet very little is known about their evolution during this period due to incomplete and highly uncertain geological data. However, from what records do exist, it is thought that the magnitude and sequence of events differed from the more recent and better constrained Last Deglaciation (LDG; ~21–9 ka). Numerical modelling of these two periods can help fill the gaps in the empirical record and provide a deeper understanding of the complex interactions that occurred between the ice sheets and climate during glacial-interglacial cycles. This will require the improvement of complex coupled climate-ice sheet models so that they are able to simulate all time periods well.
This thesis strives to contribute to this endeavour by examining the similarities and differences between the last two deglaciations through a series of coupled climate-ice sheet simulations using the model FAMOUS-ice. First, the model is tuned to produce realistic Northern Hemisphere ice sheets during the Last and Penultimate glacial maxima through large ensemble analyses and model-data comparison. This provides a range of initial Last Glacial Maximum and Penultimate Glacial Maximum ice sheet conditions for use in subsequent simulations. Error margins resulting from uncertain model parameters are explored, and the importance of different processes and feedbacks on the configuration of the glacial maximum ice sheets is quantified through sensitivity experiments and statistical analyses. Finally, transient simulations of the Last and Penultimate deglaciations are performed in which the rates and patterns of Northern Hemisphere ice sheet retreat are compared. It is shown that modelled rates of deglaciation occurred at a quicker rate during the PDG than the LDG, and the different Eurasian ice sheet configurations led to different patterns of retreat through varied instability mechanisms. Further sensitivity tests are undertaken to investigate the role of surface, sub-shelf and dynamic processes in the ice sheet retreat, as well as the relative importance of insolation, greenhouse gases and sea surface conditions in driving deglaciation. This work highlights the high sensitivity of the surface mass balance to the albedo of the ice sheets, and its dominant role in determining the configuration of the simulated ice sheets during all stages due to the ice-albedo feedback. Accurate representation of ice dynamics becomes more important when simulating ice sheet retreat, but the ice sheet evolution is not as sensitive to the rate of sub-shelf melt, except when confined ice shelves form. This thesis also highlights the need for the tuning of coupled climate-ice sheet models across transient simulations to find sets of parameters that produce plausible ice sheet configurations at all phases of glacial-interglacial cycles, as well as for present and future scenarios, to increase the reliability of simulations of future sea level rise.