PhD defense by Lisbeth Tangaa Nielsen

Title: Ice flow Modelling of the Greenland Ice Sheet
- Investigating the use of ice flow models in the interpretation of the age-structure of the Greenland ice sheet

Abstract: Models of ice flow have a range of application in glaciology, including investigating the large-scale response of ice sheets to changes in climate, assimilating data to estimate unknown conditions beneath the ice sheet, and in interpreting proxy records obtained from ice cores, among others. In this PhD project, the use of ice flow models for the interpretation of the age-structure of the Greenland ice sheet, i.e. the depth within the ice, at which ice deposited at given times are found at present day. Two different observational data sets of this archive were investigated. Further, paleo simulations of the Greenland ice sheet using ice sheet models offers the possibility of deriving reconstructions of past ice sheet topography, flow and extent, consistent with the dynamics of ice flow and the imposed climate forcing. The large-scale response of the ice sheet modelled by such approaches can potentially be used as input to more detailed analysis of the age-structure. 
The structure and depth of internal layers within an ice sheet are determined by the ice flow, which in turn depends on accumulation rates and basal conditions. The depth of a pronounced climate transition dated to 14.7kyr have previously been mapped in an extensive area in the northernl and central Greenland. Using the depth of this isochronous layer and a simple ice flow model, a map of the average accumulation rates in the area for this period was estimated. The results showed a dry central northeast, and increasing accumulation rates nearer the margins of the ice sheet. A gradient in accumulation rates were found to align with the central ice divide, suggesting that the position of the ice divide might influence the accumulation pattern in the area. The analysis further showed that in order to interpret internal layers in areas of moderate horizontal flow speeds, encountered on the flanks of the central ice divide and nearer the coast, explicitly accounting for the horizontal ice flow in the analysis was necessary, in line with the conclusions of Waddington and others, (2007). 
The counted age-scale of the high-resolution NorthGRIP ice core in northern Greenland was used to estimate the accumulation rate history at the site for the last 60 kyr. A simple ice flow model accounting for the dominating flow characteristics of the site; divergent flow across the divide and warm basal ice, was used to estimate the cumulative strain of the ice column, thereby inferring accumulation rates from the observed annual layer profile of the core. The results imply almost constant accumulation rates for the past 8kyr. In the early Holocene, accumulation rates abruptly doubles at the onset at 11.7kyr from very low accumulation rates inferred for the glacial, and afterwards slowly increases until 8kyr, where a smaller, rapid increase to the present accumulation rate is found. During the glacial, accumulation rates are found to vary with δ18O across Dansgaard-Oeschger cycles, approximately doubling at the onset of the events. A strong correlation between δ18O and accumulation rates are generally found, though in the Holocene, the correlation becomes much lower, suggesting that caution should be exerted when using δ18O as a proxy for past accumulation rates across periods with large shifts in climate and, possibly, ice sheet topography.
Ice core derived temperature and precipitation histories have a long history of being used in studies of the past evolution of the Greenland ice sheet, acting as climatic forcing of the ice sheet models. However, the conversion from the isotopic records to past temperatures remain challenging, owing to both uncertain processes and depositional histories. Using five different temperature reconstructions derived from isotope records of Greenlandic ice cores, the influence of the paleo records on the simulated ice sheet was investigated using a high-resolution, large-scale ice sheet model (PISM). The results suggest that temperature forcing records containing a period of higher than present temperatures in the early and mid-Holocene are required in order to produced a minimum in ice sheet volume, which geological evidences suggest there has been. It was further found that the ice sheet approached a steady state with respect to the reference climate at the end of the simulation and that the mass balance of the ice sheet at this time was more sensitive to recent climate fluctuations than the temperature forcing in the early or mid-Holocene.

Christine S. Hvidberg, Centre for Ice and Climate, NBI

Assessment committee
Thomas Blunier (NBI)
Robert Arthern (British Antarctic Survey, Cambridge, UK)
Ralf Greve (Hokkaido University, Sapporo, Japan)

If interested please contact Lisbeth T. Nielsen for a copy of the thesis (