Evaluation of the temperature dependence of dissociation constants for the marine carbon system using pH and certified reference materials

Abstract

Due to the uptake of anthropogenic carbon dioxide (CO2) from the atmosphere, and the resulting ocean acidification, long term monitoring of the marine carbon cycle is of utmost importance and requires high precision and accuracy across many laboratories over several decades. Despite this, many uncertainties still remain in carbon system measurements and calculations. Calculations of the carbon system are facilitated by publicly available software, but the large number of options for various constants make data and study inter-comparability challenging. Here, the carbon system is evaluated using internal consistency calculations and spectrophotometric pH measurements on two batches of Certified Reference Material over the full oceanic range of temperatures (−1.7–40 °C). The choice of formulation for the bisulfate dissociation constant is insignificant over at least the salinity range of the CRM (S = 33.4–33.8). The choice of formulation of the hydrogen fluoride dissociation constant adds a small amount of uncertainty, but the best option is unclear. The total boron concentration significantly impacts the calculated pH, with the value of Lee et al. (2010) being more internally consistent and, thus, recommended. Of the 8 carbonic acid dissociation constant sets evaluated, nearly all remain internally consistent at near-surface open-ocean salinities (~31–36) within the accuracy of pH measurements. Nevertheless, the values of Mehrbach et al. (1973), as refit by Dickson and Millero (1987), and Lueker et al. (2000) were the most internally consistent and therefore are preferred for current surface ocean studies (S ≈ 31–36, T = −1.7–40 °C, fCO2 < 500 μatm). The pH was more internally consistent at lower temperatures than higher temperatures.