Towards inferring the variability in oceanic CO2 fluxes at high latitudes using atmospheric O2 observations

The oceanic CO₂ sink displays year-to-year to decadal variabilities which are not fully reproduced by global ocean biogeochemistry models, especially in the high-latitude oceans. Oceanic CO₂ is influenced by the same climate variability and the same ecosystem processes as oceanic oxygen (O₂), although in different proportions. Unlike for CO₂, oceanic O₂ flux is not influenced directly by the rise in atmospheric CO₂, and therefore its variability reflects purely climatic and biogeochemical variability and trends. Therefore, natural climate variability and changes in oceanic processes controlling air-sea exchanges of CO₂ can be studied by focusing on oxygen (O₂), where the signal is unencumbered by direct anthropogenic influence. A global time series of oceanic O₂ flux was obtained by building a global O₂ budget, with an approach similar to the one used for the global carbon budget. The global O₂ budget is based on atmospheric O₂ observations and fossil fuel statistics, and infers the partitioning of the land and ocean fluxes using constant C:O₂ ratios for land processes. One key result of this analysis is that air-sea O₂ exchange induced significant year-to-year variability in observed atmospheric O₂. Estimates of regional oceanic O₂ fluxes were obtained from an atmospheric transport inversion analysis that inferred air-sea O₂ exchange based on global atmospheric O₂ observations and a global atmospheric transport model. For the Southern Ocean, a comparison was made between time series of winter oceanic O₂ fluxes from this inversion method and winter mixed layer depths from Argo floats. Results from this comparison confirmed the previously suggested relationship between the winter ocean mixing and air-sea O₂ exchange, which might be controlled by the climate variability induced by the Southern Annular Mode. Finally, these global and regional air-sea O₂ fluxes were compared with outputs from six global ocean biogeochemistry models to examine their current skills in simulating O₂ variability. Preliminary results suggested that all models underestimated the interannual variability in oceanic O₂ fluxes, however they were able to simulate some of the observed multi-annual variability in O₂ fluxes at high latitudes. We discuss the implications for the model’s representation of the variability in CO₂ fluxes.