A method for precise measurement of argon 40/36 and krypton/argon ratios in trapped air in polar ice with applications to past firn thickness and abrupt climate change in Greenland and at Siple Dome, Antarctica

We describe a method for measuring the Ar-40/Ar-36 ratio and the Kr-84/Ar-36 ratio in air from bubbles trapped in ice cores. These ratios can provide constraints on the past thickness of the firn layer at the ice core site and on the magnitude of past rapid temperature variations when combined with measured N-15/N-14. Both variables contribute to paleoclimatic studies and ultimately to the understanding of the controls on Earth's climate. The overall precision of the Ar-40/Ar-36 method (1 standard error of the mean) is 0.012parts per thousand for a sample analyzed in duplicate, corresponding to +/-0.6 in in reconstructed firn thickness. We use conventional dynamic isotope ratio mass spectrometry with minor modifications and special gas handling techniques designed to avoid fractionation. About 100 g of ice is used for a duplicate pair of analyses. An example of the technique applied to the GISP2 ice core yields an estimate of 11 +/- 3K of abrupt warming at the end of the last glacial period 15,000 years ago. The krypton/argon ratio can provide a diagnostic of argon leakage out of the bubbles, which may happen (naturally) during bubble close-off or (artifactually) if samples are warmed near the freezing point during core retrieval or storage. Argon leakage may fractionate the remaining Ar-40/Ar-36 ratio by +0.007parts per thousand per parts per thousand change in Kr-84/Ar-36, introducing a possible bias in reconstructed firn thickness of about +2 in if thermal diffusion is not accounted for or +6 in if thermal diffusion effects are quantified with measured N-15/N-14. Reproducibility of Kr-84/Ar-36 measured in air is about +/-0.2parts per thousand (1 standard error of the mean) but is about +/-1parts per thousand for ice core samples. Ice core samples are systematically enriched in Kr-84/Ar-36 relative to atmosphere by similar to5parts per thousand, probably reflecting preferential size-dependent exclusion of the smaller argon atom during bubble entrapment. Recent results from the Siple Dome ice core reveal two climate events during the last deglaciation, including an 18-in reduction in firn thickness associated with an abrupt warming at sometime between 18 and 22 kyr BP and a partial or total removal of the firn during an ablation event at 15.3 kyr BP. Copyright (C) 2003 Elsevier Science Ltd.