Variations in Atmospheric O2 and CO2 in the Southern Ocean Region from Continuous Ship-Based Measurements

Abstract

Variations in atmospheric oxygen (O2) are a sensitive indicator of biogeochemical processes involved in the global carbon cycle. Measurements of O2, along with carbon dioxide (CO2), can be used to determine the contributions of ocean-atmosphere and land-atmosphere exchanges to the atmospheric concentrations of these gases. However, observations from global O2 sampling networks differ significantly from model estimates, which underestimate concentrations in the southern mid-latitudes and under-represent meridional gradients in the Southern Ocean. To improve our understanding of the spatial and temporal variations of atmospheric O2 and the influence of biogeochemical processes, more observations are required. I helped to address this need by developing an instrument for continuous measurements of atmospheric O2 suitable for shipboard use. This instrument was based on a fuel cell O2 sensor, and with a finely tuned gas-handling scheme achieved a precision of 4 per meg (4 per meg is approximately equivalent to 0.8 ppm) for a 4-minute average of an air sample. The equipment was implemented on three voyages in the Western Pacific sector of the Southern Ocean. The first voyage, from Wellington, New Zealand to the Adelie Coast, Antarctica, in February 2003, indicated a summertime mid-latitude oceanic O2 source to the atmosphere. The second and third voyages, in March and April 2004, also showed oceanic O2 sources, which were localized to the southern coast of the South Island of New Zealand, and to the Chatham Rise, east of the South Island. Comparisons with measurements made at the NIWA atmospheric monitoring station. Baring Head, Wellington, showed that the April ship-based observations of O2 were elevated above the baseline concentrations expected for this time of the year. The ship-based observations were compared with simulations from air-sea flux estimates coupled to atmospheric transport models. In comparison, the models underestimated the summertime oceanic O2 source in the mid-latitudes. However, a comparison of the April observations with simulations from a high resolution (1.1° latitude by 2° longitude) model (PISCES-TM3) showed much better agreement most likely due to its ability to resolve spatial and temporal variability in biologically driven O2 fluxes. This result suggests that improved estimates for the distribution and short-term variability of biologically driven fluxes would produce improved estimates of the oceanic O2 signal in the Southern Ocean.