Validation study of air-sea gas transfer modeling

Laboratory results have demonstrated the importance of bubble plumes to air-water gas transfer (Asher et al., 1994). Bubble plumes enhance gas transfer by disrupting surface films, by directly transporting a gas, and by the creation of turbulence. Models of bubble gas transfer have been developed by different authors (Atkinson, 1973; Memery and Merlivat, 1985; Woolf and Thorpe, 1991) to determine the magnitude of gas transfer due to bubbles. Laboratory measurements of both the gas transfer rate k{sub L}, and the bubble distribution {phi} in a whitecap simulation tank (WST) have allowed these models to be validated and deficiencies in the theoretical assumptions to be explored. In the WST, each bucket tip simulates a wave breaking event. Important tests of these models include whether they can explain the experimentally determined solubility and Schmidt number dependency of k{sub L}, predict the time varying bubble concentrations, predict the evasion-invasion asymmetry, and predict the fraction of k{sub L} due to bubble plumes. Four different models were tested, a steady state model (Atkinson, 1973), a non-turbulence model with constant bubble radius (Memery and Merlivat, 1985), a turbulence model with constant bubble radius (Wolf and Thorpe, 1991), and a turbulence model with varying bubble radius. All models simulated multiple bubble tip cycles. The two turbulence models were run for sufficient tip cycles to generate statistically significant number of eddies ({number_sign}{gt}50) for bubbles affected by turbulence (V{sub B}{le}V{sub T}), found to be at least four tip cycles. The models allowed up to nine gases simultaneously and were run under different conditions of trace and major gas concentrations and partial pressures