DATA ASSIMILATION WITH A COUPLED WIND-WAVE MODEL

A first attempt to assimilate observations into a coupled wind-wave model system is presented. A special feature of the wind model is that the drag coefficient depends on the wave age, reflecting the dependence of the air flow on the wave state. Usually, both wind and wave data are assimilated. The assimilation is carried out by means of the adjoint technique. This coupled wind-wave data assimilation not only corrects successfully both model fields, but their dynamical connection is retained. Furthermore, the assimilation of only wave data results in enough information on the atmospheric state at the sea surface that it can correct both wave and wind fields as well. Moreover, there is still some impact of the assimilation during the forecast period. This persistence is found to be dependent on the typical relaxation time scale of the wind model, and also to be related to the swell decay time scale due to wave breaking. The computations reveal the importance of both the coupling and the coupled wind-wave data assimilation scheme that has been developed.