T213 降水预报订正系统的建立与研究

目前T213降水预报存在一定程度上的系统性误差,为了更好地使用T213降水预报产品,减小系统性误差对主观预报的影响,利用一种统计学方法可以对T213降水预报进行订正,减小T213降水预报的系统性误差。通过对2004年6—11月订正前后的T213降水预报进行统计学和天气学检验分析,检验该订正系统的订正效果。结果表明,订正后的降水预报的预报偏差B值有了显著改善,其他统计检验量也有了不同程度的提高;订正后雨带的位置和轮廓更加接近降水实况。

The Establishment and Research of T213 Precipitation Calibration System

With the development of the numerical forecast model,the forecasters rely more and more on the numerical models.At the same time,many products are derived from the numerical model and are used in the operational system so the numerical forecast models face the reality of improving the numerical prediction's quality and level.Among the weather elements such as the wind,the temperature,the precipitation,the air pressure and so on,without doubt the precipitation is the focus.As a result,the development of verifying precipitation is very important. On one hand,it can evaluate the level of the numerical model,on the other hand,it is benifit to realize the model's capability and features with which the model can be understood thoroughly so that to improve the capability of precipitation forecast. T213 model is the main global numerical model in China.On the basis of a long term(from 2002 to 2005) verification and evaluation towards the T213 precipitation forecast,a conclusion can be drawn that there are systemic errors in the T213 precipitation forecasting.The errors of the bias is bigger than 1 in most regions and in most seasons.In order to reduce the error's influence towards the subjective prediction in the use of numerical forecasting,a new statistical method is proposed in order to reduce T213 precipitation systemic errors with which the bias is adjusted.The method's first goal is to adjust the statistical verification bias with statistical means with which bias is much closer to 1 than before.In the precipitation forecasting,there exists over-forecast,less-forecast and accurate-forecast cases.This method is appropriate to whichever situation mentioned above. The calibration of T213 precipitation forecast is done from the summer of 2003 to the winter of 2005.In order to verify the calibration system's effectiveness,the precipitation is verified after calibration both from statistical method and synoptic method.Compared with the verification towards the precipitation before calibration,the conclusion can be reached that this calibration system is effective.The first goal of calibration is achieved because most bias of different seasons is near 1.The figure of regions where there have precipitations in term of the calibration is similar to the figure of that before the calibration.At last,the characteristics of the calibration system aimed at T213 precipitation forecast are proposed,for example,this calibration system is more effective under large-scale situation than the meso-scale and small-scale synoptic system which triggers the precipitation,also the causes are discussed.