Inclusion of uncertainty in the calcium-salinity relationship improves estimates of ocean acidification monitoring data quality

Abstract

The effects of Ocean Acidification (OA) and the resulting decrease of CaCO3 saturation state (Ω) on marine organisms and biogeochemistry are observed through regionally dispersed monitoring programs. The standard of data collected by these programs is assessed based on the computed propagated uncertainties of [CO32−], with data quality regulated by thresholds defined by the Global Ocean Acidification Observing Network (GOA-ON). While these thresholds account for the adoption of lower-cost methods and technologies for carbonate parameter analysis (e.g. pH and total alkalinity), the impacts of salinity measurement and calcium concentration uncertainty on data quality are poorly understood. Currently, the publicly available marine carbonate chemistry uncertainty packages do propagate salinity uncertainty, but do not include [Ca2+] uncertainty. In this study, the uncertainty propagation methods in the R-based seacarb package were extended to include [Ca2+] uncertainty, and subsequently employed to examine the effects of uncertainty in salinity and [Ca2+] on carbonate system calculations. The results indicate that underestimation of uncertainty in [Ca2+] is of primary concern in variable coastal waters, where relatively small (<4%) deviations from the global [Ca2+]-salinity relationship leads to GOA-ON's quality standards being exceeded. In contrast, the uncertainty in salinity has a relatively minor impact on uncertainty in [CO32−] and Ω. Given the importance of Ω and its sensitivity to [Ca2+], coastal OA monitoring programs should consider whether their region conforms with the global [Ca2+]-salinity relationship, and if uncertain should directly measure [Ca2+] when calculating Ω.