Initial Results of an Intercomparison of AMS-Based Atmospheric 14CO2 Measurements

This article presents results from the first 3 rounds of an international intercomparison of measurements of Delta(CO2)-C-14 in liter-scale samples of whole air by groups using accelerator mass spectrometry (AMS). The ultimate goal of the intercomparison is to allow the merging of Delta(CO2)-C-14 data from different groups, with the confidence that differences in the data are geophysical gradients and not artifacts of calibration. Eight groups have participated in at least 1 round of the intercomparison, which has so far included 3 rounds of air distribution between 2007 and 2010. The comparison is intended to be ongoing, so that: a) the community obtains a regular assessment of differences between laboratories; and b) individual laboratories can begin to assess the long-term repeatability of their measurements of the same source air. Air used in the intercomparison was compressed into 2 high-pressure cylinders in 2005 and 2006 at Niwot Ridge, Colorado (USA), with one of the tanks "spiked" with fossil CO2, so that the 2 tanks span the range of Delta(CO2)-C-14 typically encountered when measuring air from both remote background locations and polluted urban ones. Three groups show interlaboratory comparability within 1 parts per thousand for ambient level Delta(CO2)-C-14. For high CO2/low Delta(CO2)-C-14 air, 4 laboratories showed comparability within 2 parts per thousand. This approaches the goals set out by the World Meteorological Organization (WMO) CO2 Measurements Experts Group in 2005. One important observation is that single-sample precisions typically reported by the AMS community cannot always explain the observed differences within and between laboratories. This emphasizes the need to use long-term repeatability as a metric for measurement precision, especially in the context of long-term atmospheric monitoring.