Spectrophotometric Determination of Carbonate Ion Concentrations: Elimination of Instrument-Dependent Offsets and Calculation of In Situ Saturation States

This work describes an improved algorithm for spectrophotometric determinations of seawater carbonate ion concentrations ([CO32–]spec) derived from observations of ultraviolet absorbance spectra in lead-enriched seawater. Quality-control assessments of [CO32–]spec data obtained on two NOAA research cruises (2012 and 2016) revealed a substantial intercruise difference in average Δ[CO32–] (the difference between a sample’s [CO32–]spec value and the corresponding [CO32–] value calculated from paired measurements of pH and dissolved inorganic carbon). Follow-up investigation determined that this discordance was due to the use of two different spectrophotometers, even though both had been properly calibrated. Here we present an essential methodological refinement to correct [CO32–]spec absorbance data for small but significant instrumental differences. After applying the correction (which, notably, is not necessary for pH determinations from sulfonephthalein dye absorbances) to the shipboard absorbance data, we fit the combined-cruise data set to produce empirically updated parameters for use in processing future (and historical) [CO32–]spec absorbance measurements. With the new procedure, the average Δ[CO32–] offset between the two aforementioned cruises was reduced from 3.7 μmol kg–1 to 0.7 μmol kg–1, which is well within the standard deviation of the measurements (1.9 μmol kg–1). We also introduce an empirical model to calculate in situ carbonate ion concentrations from [CO32–]spec. We demonstrate that these in situ values can be used to determine calcium carbonate saturation states that are in good agreement with those determined by more laborious and expensive conventional methods