Laboratory simulation of ocean-atmosphere CO2 exchange

At EGU2024 we presented initial laboratory results from bubbling a simulated pre-industrial atmosphere through samples of freshwater and seawater across a range of temperatures, making comparison with literature values for the CO₂/water partition equilibrium as determined at a higher partial pressures of the gas as reviewed by Carroll et al 1991.

Two changes have been made. Our 2024 results were based on a temperature range of 0.1’C to 16.5’C, and following valued discussion with Raphael Hebert we have brought that range closer to the global average ocean temperature range since the 1940s hockey-stick, i.e. 15’C to 16.5’C. At the same time, in addressing whether last year’s `paradox’ and `slow-release’ were artefacts of laboratory simulation, we test whether changes in CO₂ fraction as measured in the headspace have a reciprocal effect in the liquid phase, measured by a continuous-reading conductivity probe in each flask.

Two recent papers are of relevance within this temperature range. Firstly the Universities of Exeter and Plymouth, UK, report transects in the Atlantic Ocean and note that temperature gradients near the ocean surface will affect the proportion of atmospheric CO₂ taken into solution (D Ford et al `Enhanced ocean CO<sub>2</sub> uptake due to near-surface temperature gradients’, Nature Geoscience (Sept 2024).  They conclude that `accounting for near-surface temperature gradients would increase estimates of global ocean CO₂ uptake.’  In parallel the University of East Anglia, UK, finds ‘that process-based models underestimate the amplitude of the decadal variability in the ocean CO₂ sink, but that observation-based products on average overestimate the decadal trend in the 2010s’. (N Mayot et al `Constraining the trend in the ocean CO₂ sink during 2000–2022’ Nature Communications, September 2024)

We understand from Raphael Hebert (pers. comm.) that the Alfred Wegener Institute, Germany, is also investigating this issue using a different approach, hence our interest in confirming the partition constant at relevant partial pressures, as a fourth contribution.