The reliability of the thermodynamic constants for the dissociation of carbonic acid in seawater

Laboratory measurements of all four CO2 parameters [fCO2 ( = fugacity of CO2), pH, TCO2 ( = total dissolved inorganic carbon), and TA ( = total alkalinity)] were made on the same sample of Gulf Stream seawater (S = 35) as a function of temperature (5–35 °C) and the ratio of (X) (1.0–1.2). Overall the measurements were consistent to ±8 μ atm in fCO2, ± 0.004 in pH, ± 3 μ mol kg−1 in TCO2, and ± 3 μ mol kg−1 in TA with the thermodynamic constants of Goyet and Poisson (1989), Roy et al. (1993), and Millero (1995). Deviations between the measured pH, TCO2, TA and those calculated from various input combinations increase with increasing X when the same constants are used. This trend in the deviations indicates that the uncertainties in pK2 become important with increasing X (surface waters), but are negligible for samples with the lower X (deep waters). This trend is < 5 μ mol kg−1 when the pK2 values of Lee and Millero (1995) are used. The overall probable error of the calculated fCO2 due to uncertainties in the accuracy of the parameters (pH, TCO2, TA, pK0, pk1, and pK2) is ± 1.2%, which is similar to the differences between the measured values and those calculated using the thermodynamic constants of Millero (1995). The calculated values of pK1, (from fCO2-TCO2-TA) agree to within ± 0.004 compared to the results of Dickson and Millero (1987), Goyet and Poisson (1989), Roy et al. (1993), and Millero (1995) over the same experimental conditions. The calculated values of pK2 (from pH-TCO2-TA) are in good agreement (± 0.004) with the results of Lee and Millero (1995) and also in reasonable agreement (± 0.008) with the results of Goyet and Poisson (1989), Roy et al. (1993), and Millero (1995). The salinity dependence of our derived values of pK1 and pK2, (S = 35) can be estimated using the equations determined by Millero (1995).