WAVE DATA ASSIMILATION WITH CLUSTERED WINDS

摘要

In the present study a method to improve wave forecasts with data assimilation is developed. To use it in operational conditions the method has to be both efficient and effective. To be effective variational data assimilation is used. To be efficient only a small number of control variables is used based on clusters of wind vectors in the computational domain. Five different methods have been investigated for the clustering. These have been tested in a series of experiments on different oceans and with different observation sets (buoys and ERS-2 and TOPEX satellite observations). The results of the tests showed little improvements in the wave forecast, which is probably caused by the high quality of wind fields in the experiments.