月尺度降水的客观预测方法研究

Study on Objective Prediction Method of Monthly Precipitation

从短期气候预测业务的实际出发,提出了一个以单站作为预报量的客观预测模型,分析了利用该模型制作预测产品的目的、实现途径以及流程设计。在气候预测业务中,利用该模型制作气候预测产品,能在实际业务中实现动态优化,方便地更换因子,降低业务工作强度。通过实例,详细说明了利用该模型制作月尺度降水产品的全过程,其中包括标准化预测模型数据文件的结构,动态建模过程及在动态建模过程中输出的各项参数的含义,把动态建模过程中的输出参数用于改进模拟质量,以及如何把试报结果分站Ps评分用于业务决策等。针对月动力延伸预报产品样本较少的情况,文中还进行了交叉预报试验。该模型作为气候预测业务工具,具有使用简单、预测质量稳定的特点。

Based on the practical application of short-term climate prediction,an objective prediction model taking the single station as predictand is built.The process of the used model making the monthly precipitation prediction is given in detail.Prediction model includes the standardized data structure,dynamic modeling process,the meaning of the output parameters in the dynamic modeling process,the way of the output parameters from the dynamic modeling process improving simulation quality,and how to use Ps score of the predicted results from the sub-station test for business decision-making.For the lack samples from the dynamical extended range forecast,the cross-validation is carried out in the paper.In short,as a climate prediction tool the forecast model has the characteristics with easy to use and stable forecast quality.