Long-term trends in surface ocean pH in the North Atlantic

Presently available direct pH measurements do not have a sufficient data density in space or time in order to determine long-term trends across wider geographic regions, limiting our ability to assess the magnitude and impacts of ocean acidification. We overcome this limitation by using the much more frequently measured fugacity of CO2 (fCO2), as synthesized in the SOCAT data product, from which we calculate pH using algorithms for alkalinity based on temperature and salinity. The estimated pH at 25 °C, i.e., pHsws25 °C has a calculation error of 0.0033 ± 0.0003, and evaluation using co-located pH observations yields a RMSE of 0.010 and a non-significant bias of 0.004. The estimated pHsws25 °C is rather sensitive to uncertainties and biases in fCO2, while uncertainties in alkalinity, temperature, and salinity matter much less. The high precision and low bias of the computed pH permit us to apply this method to data from the North Atlantic Subpolar Gyre, for which we find a statistically significant trend in surface ocean pHswsinsitu of − 0.0022 ± 0.0004 yr− 1 over the period 1981 to 2007. This long-term trend in pH is nearly entirely driven by the long-term trend in surface ocean fCO2, while the impact of temperature is negligible. This pH trend is very close to that expected based on the assumption of thermodynamic equilibrium of CO2 between the atmosphere and the surface ocean.