The ocean and climate change : relations with marine chemistry and biology (4)

Anthropogenic _O2 disturbance is very low compared to the partial pressure of _O2 in air (21 %). This is why we had to wait until the 1990s to detect anthropogenic oxygen depletion, whereas the first reliable measurements of atmospheric_CO2 date back to the ^19th century. A few decades after Alexander von Humboldt's first attempts to measure_CO2, it was the Frenchman Jules Reiset who published the first correct value for the atmosphere in 1872. At the time, the atmosphere was already anthropized, with 294 ppm, i.e. some fifteen ppm above the natural level. Humboldt and Gay-Lussac also measured _O2 content by chemical means, but were unable to identify any significant variations (" according to all our experiments, there are no variations of more than one thousandth in the quantity of oxygen contained in the air ").

The first precise measurements of _O2 (with an uncertainty < 1 ppm) were obtained in the early 90s by refractometry (San Diego) and mass spectrometry (Princeton and Hobart). The first technique is based on the fact that the refractive index of air depends (weakly !) on the _O2/N2 ratio. At present, atmospheric _O2 content is measured at some twenty stations around the world. All stations show the same long-term trend of decreasing _O2 and increasing_CO2.

However, seasonal variations between the two gases differ according to latitude. Seasonal_CO2 and _O2 cycles depend on the relative importance of continental and oceanic biosphere effects, as well as thermal (e.g. solubility) and dynamic (oceanic mixed layer) effects. Seasonal pumping by the continental biosphere is predominant in the northern hemisphere. In the Southern Hemisphere, the oceanic biosphere plays an important role, as illustrated by the seasonality of _O2. Nevertheless, the amplitude of variation in_CO2 is low because it is smoothed out, as this gas equilibrates more slowly with total dissolved_CO2, composed essentially of bicarbonate ions.

Multi-year trends in_CO2 and _O2 can be introduced into the two balance equations described above, to derive the system's two unknowns, the net flux from the oceanic pump (for_CO2) and the flux from the biospheric pump (for _O2 and_CO2). In fact, the data show that the system cannot be closed. A third unknown must be introduced, equivalent to a small flow of _O2 from the ocean to the atmosphere. Further research has shown that this flow of _O2 is linked to the degassing of the ocean in response to the warming of surface waters (approximately + 1 °C over the last century). In addition to this direct effect on _O2 solubility, other recent disturbances to ocean oxygen (stratification of surface waters, reduced ventilation, coastal eutrophication) need to be taken into account. Numerous measurements of recent changes in _O2 levels enable us to quantify the third unknown in the geochemical balance. In the end, the oceanic carbon sink deduced by this approach is 2.2 ± 0.6 GtC/year for the decade from 1993 to 2003, compared with fossil carbon emissions of around 7 GtC/year (8 GtC/year with deforestation). We therefore independently find that the oceanic sink accounts for around 30 % of fossil carbon emissions, which is compatible with the three air-sea flux mapping techniques and inventories of anthropogenic_CO2 and its isotopes^(13C, ^14C).