A paper published Tuesday in Nature Geoscience puts the Southern Ocean's annual net primary production at 6.5 ± 1.36 petagrams of carbon per year — substantially higher than most CMIP6 model estimates, and substantially higher than satellite-based estimates. [1] It lands the day after the paper covered the Antarctic basin's sea-ice and PDO tipping signal as a separate strand of the same Southern-Ocean story. Yuming Jin, Britton Stephens and Ralph Keeling reached the carbon number not by measuring the water but by reading the air above it, linking modeled productivity to modeled air-sea oxygen fluxes and applying oxygen estimates derived from airborne O₂/N₂ observations.

The wonder of the method is that it circumvents the Southern Ocean's worst logistical problem. The region south of 44°S is the most important ocean carbon sink on the planet — south of 35°S accounts for roughly 40 percent of global ocean anthropogenic CO₂ uptake [2] — and the least sampled. Winter observations are scarce. Surface pCO₂ sensors covering summer only produce seasonal asymmetries that end up in the global carbon budget as uncertainty. Jin et al. reach above the waves. Atmospheric oxygen tracks biological production on the timescale of the ocean beneath it.

The paper's second finding is harder. CMIP6 models that underestimate productivity also show weak summer CO₂ uptake, and some show excessive summer temperature-driven outgassing — producing incorrect seasonal flux cycles with summer outgassing, whereas the corrected picture, the authors argue, shows summer uptake. [1] Summer in the Southern Ocean had been modeled as a season when the ocean releases carbon. The atmospheric signal says it is a season when the ocean absorbs it.

This matters because model-projected end-of-century Southern Ocean CO₂ flux is one of the largest uncertainties in the climate sensitivity calculation. A sink that is larger than assumed today, and whose seasonal phase is reversed from assumed, compounds: the uptake floor is higher and the modeled response to warming needs reworking. A Nature Geoscience News & Views commentary published the same day called the approach "observations from the skies" and noted it reduces uncertainty in ways surface measurements cannot. [3]

The paper compounds with a piece the paper covered last week in Communications Earth & Environment on the Antarctic basin's sea-ice and PDO signal. The ice-side work warned that the Southern Ocean's physical and biological machinery is already shifting. The Jin et al. work shows the machinery has been carrying more carbon than accountants credit it with. Both are possible. Both are consistent with an ocean that is changing faster than the models sampling it.

Argo oxygen-based estimates are consistent with the new value — 6.5 PgC/yr sits close to independent float-derived productivity numbers that had been in tension with CMIP6 for years. [1] The disagreement was not a mystery. It was a measurement gap. Reading the atmosphere closed some of it.

Earth Day delivered a quiet correction: the Southern Ocean, in summer, is not giving carbon back. It is still taking it. More of it than we thought. For a little longer.

-- KENJI NAKAMURA, Tokyo