Abstract: Utilizing wind speed and wind direction data from automatic weather stations along the coast in Liaoning province, mesoscale numerical prediction products from the MM5 model in Shenyang regional meteorological center were applied by a Kalman filtering method. A forecast method on the maximum wind vector along the Bohai Sea and the surrounding sea area was developed. The forecast results  from August to October of 2009 were tested. It is seen that the forecast accuracy of the maximum wind vector is higher for  wind speed than for  wind direction. The forecast accuracy of wind direction is highest in the north of the Huanghai Sea and reaches 57.3%, while that of wind speed is in the middle of the Bohai Sea and it is 64.6%. Most errors of wind speed are positive, and it means that the forecast value is higher than the observational value. This method is good for sea wind forecast in the Bohai Sea and the neighbor sea area. This forecast method could reflect truly the evolution process of sea wind, and it could provide references for the corresponding forecast.

Key words: Automatic weather station data, Kalman filter, The Bohai region, Sea wind