Determination of total dissolved inorganic carbon in sea water using extraction/coulometry: The first stage of a collaborative study.

摘要

Samples of sterilized sea water were bottled and sent to a total of fifteen laboratories for analysis for total dissolved inorganic carbon. These were located in Canada (2), Germany (1), The Netherlands (1), Sweden (1), the United States (9), and the United Kingdom (1). All except one laboratory used their current implementation of an extraction/coulometric procedure to analyze the samples; the other laboratory, Dr. Keeling's laboratory at the Scripps Institution of Oceanography, analyzed the samples using their extraction/manometric procedure which is considered a ''definitive'' method. The within-laboratory standard deviation of the results reported by the various laboratories (omitting the two that had substantial difficulties) ranged from 0.56--2.75 (mu)mol(center dot)kg(sup (minus)1) with a pooled standard deviation of 1.55 (mu)mol(center dot)kg(sup (minus)1). The weighted mean calculated from the results of these twelve laboratories was 1977.8 (mu)mol(center dot)kg(sup (minus)1); almost identical to the certified value obtained by extraction/manometry (1978.8 (mu)mol(center dot)kg(sup (minus)1)). This indicates that the extraction/coulometric procedure is capable of making unbiased measurements. Unfortunately, the spread of the mean values from these twelve laboratories was substantial (17.2 (mu)mol(center dot)kg(sup (minus)1)) and only five sets of results had a mean that was equivalent to the certified value (with 95% confidence).