Impact of physical and biological processes on temporal variations of the ocean carbon sink in the mid-latitude North Atlantic (2002–2016)

The ocean is currently a significant net sink for anthropogenically remobilised CO2, taking up around 24% of global emissions. Numerical models predict a diversity of responses of the ocean carbon sink to increased atmospheric concentrations in a warmer world. Here, we tested the hypothesis that increased atmospheric forcing is causing a change in the ocean carbon sink using a high frequency observational dataset derived from underway pCO2 (carbon dioxide partial pressure) instruments on ships of opportunity (SOO) and a fixed-point mooring between 2002 and 2016. We calculated an average carbon flux of 0.013 Pg yr−1 into the ocean at the Porcupine Abyssal Plain (PAP) site, consistent with past estimates. In spite of the increase in atmospheric pCO2, monthly average seawater pCO2 did not show a statistically significant increasing trend, but a higher annual variability, likely due to the decreasing buffer capacity of the system. The increasing led to an increasing trend in the estimated CO2 flux into the ocean of 0.19 ± 0.03 mmol m−2 day−1 per year across the entire 15 year time series, making the study area a stronger carbon sink. Seawater pCO2 variability is mostly influenced by temperature, alkalinity and dissolved inorganic carbon (DIC) changes, with 77% of the annual seawater pCO2 changes explained by these terms. DIC is in turn influenced by gas exchange and biological production. In an average year, the DIC drawdown by biological production, as determined from nitrate uptake, was higher than the DIC increase due to atmospheric CO2 dissolution into the surface ocean. This effect was enhanced in years with high nutrient input or shallow mixed layers. Using the rate of change of DIC and nitrate, we observed Redfieldian carbon consumption during the spring bloom at a C:N ratio of 6.2 ± 1.6. A comparison between SOO and PAP sustained observatory data revealed a strong agreement for pCO2 and DIC. This work demonstrates that the study area has continued to absorb atmospheric CO2 in recent years with this sink enhancing over time. Furthermore, the change in pCO2 per unit nitrate became larger as surface buffer capacity changed