Master thesis defense by David Soestmeyer

Title: Local temperature reconstruction from gas trapping processes in the Dye 3 ice core from Greenland

Abstract: Ancient air trapped in ice cores can reveal information about past atmospheres and climate conditions. The inert gas nitrogen and noble gas argon are gases with long atmospheric lifetimes – thus any alteration in their isotopic composition observed in ice cores is indicative of the drill site’s past conditions. Using a combination of laboratory isotopic measurements and an inversion model, it is possible to reconstruct the site’s past atmospheric temperature. This master thesis project developed a Continuous Flow Analysis (CFA) system to measure δ15N of nitrogen and δ40Ar of argon continuously from gas bubbles trapped in the Dye 3 ice core from South Greenland. The system gives an unprecedentedly high data resolution and generates results faster than previously established discrete measurements. A perovskite membrane was used to remove oxygen from the sample stream to obtain an improved δ40Ar data quality. In total, 108 m of ice core were melted in 1 m sections continuously and then analyzed by mass spectrometry. The δ15N data result in a realistic surface temperature reconstruction, however unfortunately it seems likely the δ40Ar data suffer from analytical issues. A fitting algorithm coupled with a firn model were used to reconstruct the Dye 3 surface temperature. The reconstructed age section dating from 28,000 years to 44,000 years b2k includes Dansgaard‑Oeschger (D-O) climatic events 4 to 11. It indicates an average warming of 22.41°C between the onset and peak temperature of the D-O events. In comparison with the NGRIP ice core from Central Greenland, the temperature reconstruction shows on average 10.35°C warmer temperatures for Dye 3 which is comparable to present day temperature differences. Compared to the discrete samples taken from Dye 3, the CFA δ15N are on average 0.09 ‰ lower. Finally, as a result of the experience gained and the analytical challenges faced, suggestions are given to improve the data quality for future measurements.