Hybrid DNS/LES of high Schmidt number mass transfer across turbulent air–water interface

Abstract

Numerical simulation of a coupled air–water turbulent flow and associated high Schmidt number mass transfer is carried out via a hybrid scheme of direct and large-eddy simulations (DNS/LES). Due to the large density ratio of water and air, the dynamical coupling between the air and water turbulent flows is found to be weak at the low wind velocity considered here. Instead, the self-sustaining mechanisms due to the mean shear, which are similar to those near a solid wall, are dominant even close to the air–water interface. The spatio-temporal correlations between the local mass transfer rate and velocity fluctuations around the interface reveal that impingement of fresh water on the interface governs the interfacial mass transfer. It is found that the local mass transfer rate can be predicted from the surface divergence by the Chan and Scriven’s stagnation flow model. This explains why the mass transfer rate is well correlated with the intensity of the surface divergence under a variety of flow conditions.