A Study of Information on the Extraction to Reduce Errors in Dynamic Extended (Monthly) Numerical Weather Prediction
动力延伸(月)数值天气预报中的信息提取和减小误差试验

A number of monthly numerical weather prediction cases based on a global spectral model have been used to study the extraction of useful information in the predicted results. Spectral analysis of model prediction errors shows that the zonal symmetric part (" wave-number zero" ) accounts for a large portion of the total error. Using climatic tendency, two approaches have been tested to reduce this part of error. The first is to correct the final forecast result by replacing the zonal symmetric component by its climatic counterpart. The second is to proceed the correction of this part during the integration pro- cess. Both approaches prove effective.