Net and gross O2 production in the southern ocean from measurements of biological O2 saturation and its triple isotope composition

Abstract

Mixed layer seawater samples from the Southern Ocean were analyzed for the triple oxygen isotope composition (δ17O and δ18O) of dissolved O2 and the ratio of [O2] to [Ar]. δ17O and δ18O together constrain the mass independent anomaly in O2, and hence the fraction of photosynthetic O2 in the dissolved O2 pool. Assuming oxygen in the mixed layer is at steady state, we calculate ratios of the gross photosynthetic O2 production rate to the O2 air–sea gas exchange rate, and ratios of net to gross O2 production rates. With estimates of the O2 gas exchange rate from a wind speed parameterization, we determine absolute in situ rates of gross and net O2 production. Based on the net/gross production ratios and δ18O of dissolved O2, we calculate a value for the 18O fractionation factor associated with marine respiration in the Southern Ocean mixed layer of 0.978 (an isotope effect of 22‰). The study regions cover latitudes between ∼45°S and the ice edge at: (1) 175°E in December 1999, (2) 145°E in December 2000, and (3) 145°E in January 2001. At both meridians, gross O2 production decreases to the south. At 145°E, rates of net O2 production follow the same pattern, while at 175°E these rates are consistently low at all latitudes. In December, gross and net O2 production rates are both higher at 145°E than at 175°E. Gross O2 production at 145°E was similar in December and January, but net O2 production decreased by ∼50%. Net/gross C production ratios, calculated by scaling O2 production rates, are lower than estimates of the 15N f-ratio, determined as the ratio of 15NO3− uptake to 15NO3−+15NH4+ uptake in 15N incubations. They are also lower than ef-ratios predicted by the model of Laws et al. (Global Biogeochem. Cycles 14 (2000a) 1231) at the temperatures of the Southern Ocean. The differences in the measurements could be due to natural interannual and spatial variability or the differences in metabolic rates measured and modeled, and the fundamental assumptions required by each technique.