An inter-comparison of autonomous in situ instruments for ocean CO₂ measurements under laboratory-controlled conditions

Observations of the marine CO₂ system are important for understanding ocean acidification, air-sea CO₂ fluxes and the marine carbon cycle in general. A variety of autonomous in situ instruments have been developed and used towards this end. There is, however, a lack of inter-comparison studies of currently available sensor technologies. In this study, a total of 10 instruments including commercially available pH, partial pressure of CO₂ (pCO₂), and total alkalinity (A_T) sensors were tested and compared in a 5000 L seawater tank located at Scripps Institution of Oceanography (SIO), California, USA. The test took place over ~12 days (August 16 to 28, 2016) where conditions in the tank were artificially varied to encompass a wide range of A_T, pH and pCO₂ as well as temperature and salinity. To assess accuracy, independent measurements of pH, pCO₂, dissolved inorganic carbon (DIC) and A_T were made using benchtop instrumentation. We also evaluated internal consistency, comparing the measured parameter from the sensors with the calculated parameter, e.g. pCO₂ calculated from A_T and pH compared with directly measured pCO₂. DIC can be precisely derived (within ±5 μmol/kg) over a wide range of conditions from measured pH or pCO₂ paired with A_T. Sensor accuracy established by discrete samples is sufficient for short-term and seasonal dynamical studies, but their ability in determining long-term (e.g. climate) variability could not be evaluated because of the brevity of the study. These results provide insights into sensor performance and strategies for data quality control for future studies of ocean acidification and carbon cycling.