Deciphering the Fine Structure of Methane Change Around the Time Interval Encompassing DO event 21 in the GISP2 Ice Core.

We redeveloped the melt-refreeze extraction technique for measuring methane in ice cores allowing precision higher than previously possible with that method. The standard deviation of methane values achieved on replicate pieces of ice (~60 g) is typically within the 1-10 ppb range on the new technique. The improvement was partly due to adopting analytical approaches employed by laboratories working on modern air. The second component of the improvement had to do with subtle refinements in the handling of ice and extracted air. Equipped with the ultra-precise method of measuring methane we are currently focusing on the time interval during the early glacial, at ca. 75-85 kyrs BP. This interval was punctuated by the Dansgaard-Oeschger event 21, and previous workers have shown that methane experienced change in accord with DO-21. However, it is the fine structure of this methane change that may be particularly illuminating for supplying additional clues for how mechanisms of climate and methane change are interlinked. The approach of Severinghaus and Brook (Science, vol. 286, p. 930, 1999) that is in making a direct comparison between methane and gas-recorded temperature change should offer new perspectives when pushed to next level of precision. The methane dataset presented here can be exemplified with depth interval 2670 to 2640 meters (~77 to 82.5 ka, presently analyzed every 2 meters). There is a general decline in methane (~670 ppb to ~520 ppb) that matches the pattern of declining δ18Oice (temperature proxy). The decline in methane is punctuated by two reversals at 2665 and 2650 meters. The dataset allows identifying the timing and magnitude of these reversals (e.g. a magnitude of 18 ± 7 ppb for the earlier reversal).