An inter-laboratory comparison assessing the quality of seawater carbon dioxide measurements

Seawater CO2 measurements are being made with increasing frequency as interest grows in the ocean's response to changing atmospheric CO2 levels and to climate change. The ultimate usefulness of these measurements depends on the data quality and consistency. An inter-laboratory comparison was undertaken to help evaluate and understand the current reliability of seawater CO2 measurements. Two seawater test samples of different CO2 content were prepared according to the usual method for the creation of seawater reference materials in the Dickson Laboratory at Scripps Institution of Oceanography. These two test samples were distributed in duplicate to more than 60 laboratories around the world. The laboratories returned their measurement results for one or more of the following parameters: total alkalinity (AT), total dissolved inorganic carbon (CT), and pH, together with information about the methods used and the expected uncertainty of the measurements. The majority of laboratories reported AT and CT values for all their measurements that were within 10 μmol kg−1 of the assigned values (i.e. within ± 0.5%), however few achieved results within 2 μmol kg−1 (i.e. within ± 0.1%), especially for CT. Results for the analysis of pH were quite scattered, with little suggestion of a consensus value. The high-CO2 test sample produced results for both CT and pH that suggested in many cases that CO2 was lost during analysis of these parameters. This study thus documents the current quality of seawater CO2 measurements in the various participating laboratories, and helps provide a better understanding of the likely magnitude of uncertainties in these measurements within the marine science community at the present time. Further improvements will necessarily hinge on adoption of an improved level of training in both measurement technique and of suitable quality control procedures for these measurements.