Hosts
Partners
Understanding the global carbon cycle is largely based on atmospheric measurements. These data are used by modelling frameworks that derive CO2 surface-atmosphere exchange flux estimates by inversion techniques. To yield reliable results, this method requires good geographic and temporal data coverage. Discrete air samples filled in glass flasks at globally distributed sites that are later analyzed in different laboratories are a cost effective approach to achieve this goal. Merging results from different institutions does, however, require knowledge of the degree of compatibility of the measurement data. In particular, for new observational networks like ICOS in Europe or new laboratories in other parts of the world it is essential to prove the quality of their measurements. One requisite is traceability of measurements to the WMO calibration scale or another calibration scale that has a well-established scale conversion function to the WMO scale. Yet, biases (either long-term or more temporary in duration) between laboratories can result from individual strategies to propagate the scale, from differences in the calibration frequencies and detection principles as well as sporadically emerging, unnoticed problems in the individual laboratories. Here we describe results from an ongoing experiment (beginning in 2002) to assess the compatibility of CO2 flask analysis involving many international laboratories. In this so-called “sausage” flask comparison, all participants receive sets of flasks (of the type normally used within the participants program) filled with comparison air spanning the range of typical atmospheric abundance of the gases of interest. This is repeated several times per year. The large number of comparison points allows the on-going assessment of compatibility limits resulting from repeatability limitations and from systematic differences between laboratories. In addition, occasional round robin and other direct cylinder exchanges have been performed during the same period. The inter-laboratory offsets determined by the sausage flask comparison generally agree with the offset as they appear in the high pressure cylinder comparisons. However, smaller systematic differences between the respective results are observed and sporadic periods with deviating offset patterns are seen that are missed with less frequent cylinder exchanges. Hence, this evaluation provides quantitative evidence for the uncertainty of assessing inter-laboratory flask data offsets based on infrequent cylinder exchanges. This evaluation also demonstrates the potential and limits of scale conversion functions derived based on flask data comparisons.
Topic : Theme 2: Accuracy requirements for atmospheric composition measurements across economic sectors, and temporal and spatial scales.
Reference : T2-A12
Back to the list of submissions
Previous submission
·
Next submission
Comments are only accessible to participants. If you are a participant, please log in to see them.