The air-water exchange of gases can be substantially enhanced by wave breaking and specifically by bubble-mediated transfer. A feature of bubble-mediated transfer is the additional pressure on bubbles resulting from the hydrostatic forces on a submerged bubble and from surface tension and curvature. This peculiarity results in asymmetry of bubble-mediated gas transfer and equilibrium supersaturations of dissolved gases in a bubbly ocean. A second peculiarity is the finite capacity of bubbles, so that the composition of a bubble may change during the exchange. The result is that gas transfer mediated by bubbles is characterized by an altered dependence on the molecular properties of the dissolved gas compared to direct transfer across the main air-water interface. A related phenomenon for bubble plumes with a high void fraction (air volume to total volume ratio) is that the composition of the dissolved gas within the interstitial water of a plume may alter during the exchange process and only mix into the full water reservoir later. Three asymptotes are identified for gas exchange mediated by high-void-fraction bubble plumes and a semi-empirical parameterization of bubble-mediated gas transfer is devised on the basis of these asymptotes, which describes the dependence of the overall transfer velocity on plume properties and molecular properties of the gas.
These models are confronted with data from laboratory experiments. The experiments use artificial aeration with the gas source switched during each run. Measurements of the bubble distribution enable calculation of the theoretical transfer of the gases. A parameterization fits the theoretical transfer satisfactorily. Gas measurements are used to test if the actual transfer of gases is similar to the theoretical transfer. The experimental method enables separation of bubble-mediated transfer from transfer directly across the main air-water interface. The agreement between gas and bubble-derived values of transfer velocity is sufficient to generally validate the theory, but is imprecise. The results suggest that the interstitial water plays a significant role in limiting gas transfer-in particular, limiting transfer of helium-despite the fact that typical void fractions were low (< 0.1%). It should be possible to predict gas transfer velocities in the field by simulating oceanic bubble plumes sufficient to constrain that part of the transfer, but targets of 10% or 20% may be beyond reach especially for the most poorly soluble gases (for which the bubble-mediated mechanism is particularly important). These simulations require accurate bubble distributions, void fractions and a good description of the entire plume dynamics. Such simulations are particularly important for interpreting dual tracer and nitrogen/oxygen experiments in stormy conditions, where the relative transfer of different gases is a non-trivial problem.