数值模式评估系统MET及其初步应用

Numerical Weather Prediction Model Testing and Evaluation Systems MET and Application

  • 摘要: MET(Model Evaluation Tools)是美国国家大气研究中心(National Center for Atmospheric Research,NCAR)数值预报发展试验中心(Developmental Testbed Center,DTC)最新研发的数值预报检验、评估系统,其主要目的不仅在于为模式开发人员提供模式性能的测试工具,而且方便最终用户通过MET来判别模式的预报能力,进而获得预报相应的预报指标。MET系统提供了丰富的站点、格点数据接口程序,它的核心组件集成了模式经典方法评估检验,高分辨率模式诊断分析,集合预报、概率预报、台风路径等多种最先进的模式检验评估算法。为了方便用户使用,MET系统还提供基于R语言的多种绘图脚本。在对MET系统简要阐述的基础上,给出了基于MET系统的面向对象或基于目标的MODE(Method for Object-Based Diagnostic Evaluation)方法、邻域法的具体降水检验个例,为使用MET系统进行模式检验、评估提供参考。

     

    Abstract: The MET (Model Evaluation Tools) comprehensively uses many kinds of advanced model testing and evaluation algorithms, such as the classical methods, the analysis of high resolution mode diagnosis, ensemble prediction, probability forecast, and the typhoon path test. It is the newest numerical prediction test, evaluation system developed by DTC (The Developmental Tested Center) in the NCAR (National Center for Atmospheric Research, USA). The main purposes are not only to provide a mode performance test tool for mode developers, but also to give users a forecast ability of distinguishing mode and then getting the forecast index by using MET. MET provides abundant interfaces of stations and grid data; it's core component integrates many advanced classic mode test methods above mentioned. In addition, it provides various drawing scripts based on language R for user's convenience. This article briefy explains the MET system, and gives a precipitation test example based on the MET object-oriented method and neighborhood method.

     

/

返回文章
返回