Master Thesis Defense by Majbritt Kristin Eckert

Title: Modelling the Greenland Ice Sheet in Past Warmer Climates - Investigating the Impact of Domain Choice and Oceanic Forcing on the Last Two Deglaciations

Abstract: 
The ice sheets on Earth are critical components of the climate system at present day and in the future. Mass loss by melt due to climate warming could contribute a large amount to global sea level rise. With this thesis, we aim to explore the dependency on ice flow parameters and climatology that become important when trying to model a realistic last deglaciation and present day ice sheet using the Parallel Ice Sheet Model (PISM). This could help to improve future climate change simulations. Furthermore, we want to explore the response of the Greenland Ice Sheet to rapid climate warming events with high temperatures in the past like the Eemian deglaciation.

During the Holocene, surface elevation changes derived from ice cores suggest a large thinning in the North of Greenland which is where the Greenland Ice Sheet and the Innuitian Ice Sheet were connected at the end of the Last Glacial Maximum. The first research question was therefore to include Canada in the domain and compare the difference this makes to only modelling Greenland. In a second experiment, we tested different oceanic forcing parameterizations while tuning to present day observations. The initial glacial ice sheets for these transient runs were simulated using equilibrium runs approximating the glacial climate.

We have found three main influences on the modelled interglacial ice sheet, the climate forcing and regional climate, ice flow parameterization and, in our case the larges contributor to enhanced retreat, the bedrock depression. Our results show that including Canada alters the ice extent and velocity patterns in northern Greenland due to enhanced bedrock depression and delayed isostatic rebound from the additional glacial ice mass. While we can say with high confidence that the domain choice affects the present day outcome, our results are not precise in what the exact impact is. This is because tuning a model like PISM with a large amount of variables requires much accuracy and we have not managed to match present day observations of the Greenland Ice Sheet well enough. We found that including oceanic forcing shifts the deglaciation onset by a few thousand years into the past. This means that the choice of melt rate and scaling in this forcing are important. A similar sensitivity on temperature and precipitation forcing was found in simulations of the Eemian. We used the same parameterization as for the Holocene deglaciation and see that a delayed bedrock response during the retreat and the simplifications we made to approximate the regional climate in the past lead to an almost full retreat of the Greenland Ice Sheet.

Supervisor: Christine Schøtt Hvidberg
Censor: Peter Lang Langen (AU)