Search for: All records

Bubble-mediated gas exchange associated with wave breaking is a critical pathway for ocean–atmosphere exchange of low solubility gases such as oxygen. Yet, ocean and climate models, as well as observation-based products, usually rely on wind-only air–sea flux formulations derived from carbon constraints that ignore the asymmetric nature of the bubble flux, contributing to discrepancies between estimates of oxygen inventories and their response to climate change. Without bubbles, gas exchange is controlled by a symmetric wind-driven exchange, with the ocean–atmosphere gas partial pressure difference controlling whether outgassing or uptake occurs. Bubbles entrained by wave breaking can enhance this symmetric turbulent exchange, and contribute an additional asymmetric flux, always leading to an uptake, as they get squeezed by hydrostatic pressure (large bubbles) or collapse and fully dissolve (small bubbles). We present an observation-constrained theoretical framework of the air–sea flux accounting for air entrainment due to wave breaking and symmetric and asymmetric bubble exchange. The combined evidence from theory, laboratory, and field measurements of carbon dioxide fluxes, oxygen concentration, and noble gas supersaturation yields a universal formulation of gas exchange which we implement into a global ocean biogeochemical model. We discuss the resulting oxygen fluxes and demonstrate that our wind–wave–bubble formulation better reproduces observed in situ oxygen concentrations in water mass formation regions, where air–sea exchange is high, than a commonly used wind-only formulation. We show that the asymmetric bubble flux is essential for evaluating air–sea oxygen fluxes and estimating the magnitude of the ocean oxygen loss associated with global warming. 

Abstract Wave breaking induced bubbles contribute a significant part of air‐sea gas fluxes. Recent modeling of the sea state dependent CO₂flux found that bubbles contribute up to ∼40% of the total CO₂air‐sea fluxes (Reichl & Deike, 2020,https://doi.org/10.1029/2020gl087267). In this study, we implement the sea state dependent bubble gas transfer formulation of Deike and Melville (2018,https://doi.org/10.1029/2018gl078758) into a spectral wave model (WAVEWATCH III) incorporating the spectral modeling of the wave breaking distribution from Romero (2019,https://doi.org/10.1029/2019gl083408). We evaluate the accuracy of the sea state dependent gas transfer parameterization against available measurements of CO₂gas transfer velocity from 9 data sets (11 research cruises, see Yang et al. (2022,https://doi.org/10.3389/fmars.2022.826421)). The sea state dependent parameterization for CO₂gas transfer velocity is consistent with observations, while the traditional wind‐only parameterization used in most global models slightly underestimates the observations of gas transfer velocity. We produce a climatology of the sea state dependent gas transfer velocity using reanalysis wind and wave data spanning 1980–2017. The climatology shows that the enhanced gas transfer velocity occurs frequently in regions with developed sea states (with strong wave breaking and high significant wave height). The present study provides a general sea state dependent parameterization for gas transfer, which can be implemented in global coupled models. 

Abstract The drag coefficient under tropical cyclones and its dependence on sea states are investigated by combining upper-ocean current observations [using electromagnetic autonomous profiling explorer (EM-APEX) floats deployed under five tropical cyclones] and a coupled ocean–wave (Modular Ocean Model 6–WAVEWATCH III) model. The estimated drag coefficient averaged over all storms is around 2–3 × 10 −3 for wind speeds of 25–55 m s −1 . While the drag coefficient weakly depends on wind speed in this wind speed range, it shows stronger dependence on sea states. In particular, it is significantly reduced when the misalignment angle between the dominant wave direction and the wind direction exceeds about 45°, a feature that is underestimated by current models of sea state–dependent drag coefficient. Since the misaligned swell is more common in the far front and in the left-front quadrant of the storm (in the Northern Hemisphere), the drag coefficient also tends to be lower in these areas and shows a distinct spatial distribution. Our results therefore support ongoing efforts to develop and implement sea state–dependent parameterizations of the drag coefficient in tropical cyclone conditions.