Sea spray exchanging momentum, heat, and moisture is one of the major uncertainties in modeling air–sea surface heat fluxes under high wind speeds. As a result of several untested assumptions in existing models and low fidelity in the measurements, questions regarding the appropriate method for modeling the effects of spray on air–sea fluxes still exist. In this study, we implement idealized direct numerical simulations (DNS) via an Eulerian–Lagrangian model to simulate spray droplets in turbulent flows. Then, we verify the bulk spray models of Fairall et al. and Andreas et al. with the detailed physics from DNS. We find that the quality of the underlying assumptions of bulk models is sensitive to the time scales governing spray microphysics and lifetime. While both models assume that spray experiences a uniform and steady ambient condition, our results show that this assumption only works well for droplets with long thermodynamic time scales and relatively short lifetime. When the thermodynamic time scales are short, the models fail to predict the correct temperature and radius change of spray (e.g., condensation), thus spray-mediated heat fluxes, which in turn overestimates the total heat fluxes. Moreover, using our two-way coupled simulations, we find a negative feedback induced by the spray evaporation that may be missing in the bulk models, which could lead to further overestimates of the total heat flux when the spray-mediated flux is treated as an add-on to the corresponding interfacial flux. We further illustrate that the feedback effects are consistent under different flow Reynolds numbers, which suggests that the findings are relevant at practical scales.

No abstract

We improved the overestimation issue in bulk model for the total heat flux by con-8 sidering spray timescales. 9 • Using the volume-weighted mean size as a representative size could simplify the 10 quantifying of air-sea fluxes.

No abstract