Two decades of pHT measurements along the GO-SHIP A25 section

Abstract. The North Atlantic (NA) GO-SHIP A25 OVIDE-BOCATS section is a long-term repeat hydrographic transect extending from Portugal to Greenland. Since 2002, physical and biogeochemical measurements have been carried out biennially along the OVIDE-BOCATS section, contributing to a better understanding of water mass properties, mixing, circulation, carbon storage, and climate change impacts such as ocean acidification (OA) in the NA. In particular, the high-precision pH measurements on the total hydrogen ion scale (pH_T) from the OVIDE-BOCATS program represent a key milestone in monitoring OA in this particularly climate sensitive region. The method used for pH_T determination relies on adding meta-cresol purple (mCP) dye to the seawater sample and spectrophotometrically measuring its absorbances at specific wavelengths. The OVIDE-BOCATS program has used unpurified mCP dye, which impurities have been proven to bias pH_T values. Here we quantified the bias induced by these impurities in pH_T measurements. We found that measurements carried out using the unpurified mCP dye tend to be, on average, 0.011 ± 0.002 pH_T units higher than those obtained using the purified mCP dye, with this difference slightly decreasing at higher pH_T values. Moreover, we tested independent methods to correct the effect of impurities in both the historical and recent OVIDE-BOCATS pH_T data, demonstrating that the correction is consistent across methods. The long-term pH_T dataset has been updated to include newly acquired data and absorbance measurements, and to standardize corrections for mCP dye impurities. This effort results in a twenty-year dataset of pH_T corrected for mCP dye impurities, that demonstrates the possibility of a global effort to improve the reliability and coherency of spectrophotometric pH_T measurements made with unpurified mCP dye. The corrections applied to our pH_T dataset have negligible implications for the OA rates previously reported, but they do affect the depth of the aragonite saturation horizon, implying a shoaling of approximately 150 m.