PhD Defense by Troels Bøgeholm Mikkelsen

Title: Ice Sheets & Ice Cores
Data Analysis & Stochastic Modelling

Abstract: Since the discovery of the Ice Ages it has been evident that Earth’s climate is liable to undergo dramatic changes. The previous climatic period known as the Last Glacial saw large oscillations in the extent of ice sheets covering the Northern hemisphere. Understanding these oscillations known as Dansgaard-Oeschger (DO) events would add to our knowledge of the climatic system and – hopefully – enable better forecasts. Likewise, to forecast possible future sea level rise it is crucial to correctly model the large ice sheets on Greenland and Antarctica. This project is divided into two parts. The first part concerns time series analysis of ice core data obtained from the Greenland Ice Sheet. We analyze parts of the time series where DO-events occur using the so-called transfer operator and compare the results with time series from a simple model capable of switching by either undergoing a bifurcation or by jumping at random. We find that the DO event time series is most consistent with the model undergoing a random transition.
In the second part, we use a simple model of an ice sheet subject to constant and fluctuating temperatures, respectively. We find that the steady state volume of the ice sheet is lower for fluctuating temperatures than for a constant temperature. This finding may have implications for future long-range ice sheet projections, as the steady state ice sheet volume could be underestimated in studies using a constant temperature.

Peter D. Ditlevsen, Centre for Ice and Climate, NBI 
Susanne Ditlevsen, Department of Mathematical Sciences 

Assessment committee
Thomas Blunier, Centre for Ice and Climate, NBI 
Roderik Van de Wal, University of Utrecht  
Poul G. Hjort, DTU

If interested please contact Troels B. Mikkelsen for a copy of the thesis (