An alternative approach to the parameterization of the momentum flux over the sea

In this study we search for the optimal wave parameters characterizing the effect of waves on the air-sea momentum flux. When an attempt using traditional statistics failed to give an unequivocal answer, we applied a neural network approach. This led to a clear choice for the optimal parameters viz. the wind speed at a level equal to the wave height and the wavelength at the peak of the spectrum. With this set we could find a dimensionally correct parameterization that gave a close fit to our stress data, obtained in the southern North Sea. A comparison with open ocean data gave comparable results, allowing a generally applicable formulation. Wave breaking could explain the remaining small differences between the two sets of stress data.