CMA数值预报模式在线交互式试验管理系统  

An Online Interactive Experiment Management System for CMA Numerical Weather Prediction Models

作　　者：王彬 胡江凯[1,2,3] 崔应杰 周斌[1,2,3] 孙婧 赵春燕[4] 娄盼星 Wang Bin;Hu Jiangkai;Cui Yingjie;Zhou Bin;Sun Jing;Zhao Chunyan;Lou Panxing(CMA Earth System Modeling and Prediction Center,Beijing 100081;State Key Laboratory of Severe Weather,Chinese Academy of Meteorological Sciences,Beijing 100081;CMA Key Laboratory of Earth System Modeling and Prediction,Beijing 100081;National Meteorological Information Center,Beijing 100081;Shaanxi Institute of Meteorological Sciences,Xi'an 710016)

机构地区：[1]中国气象局地球系统数值预报中心,北京100081 [2]中国气象科学研究院灾害天气国家重点实验室,北京100081 [3]中国气象局地球系统数值预报重点开放实验室,北京100081 [4]国家气象信息中心,北京100081 [5]陕西省气象科学研究所,西安710016

出　　处：《应用气象学报》2025年第1期110-120,共11页Journal of Applied Meteorological Science

基　　金：国家重点研发计划(2017YFA0604500);数值预报关键技术研发——国省统筹团队科创平台组,气象软科学研究(2024MSXM18)。

摘　　要：数值预报研发试验是改进数值预报模式性能的主要技术手段。模式试验管理系统是支持研发试验不可或缺的工具平台。面向数值预报统筹研发工作机制需要,并借鉴国内外相关研究工作,设计数值预报模式在线交互式试验管理系统。该系统采用浏览器/服务器软件部署架构,提供模板编辑-流程编排-实例构建功能,支持灵活快捷设计模式试验,所有模块完全打通、一站式访问使用,通过个例数据治理、程序代码制品缓存等手段提升用户体验。系统实现对模式试验全过程的支持,包括模式试验的设计定义、试验实例的创建运行、作业监控管理、结果可视化与检验分析。系统已于2023年上线业务运行,多场景支持数值预报国省统筹研发。Numerical weather prediction research and development experiments are the key technical means to improve the model performance.The model experiment management system is an essential platform that enables model experts to conduct experiments effectively.To address the need for a coordinated development mechanism in numerical weather forecasting and drawing from relevant domestic and international research,an online interactive experiment management system for numerical weather prediction models has been designed.The system utilizes a browser/server(B/S)software deployment architecture,which consists of four layers on the server side:Application layer,support layer,data layer,and resource layer.The application layer offers management functions at the user application level for modeling experiments.Software modules of the support layer are designed to address the needs of various scenarios and are integrated to provide services that can be accessed by the application layer.The data layer offers a comprehensive data management platform service for the entire system.The resource layer provides foundational support in terms of computing power,data,and code.The system offers functions for template editing,process arrangement,and instance construction,facilitating the flexible and rapid design of model experiments.All modules are fully interconnected and can be accessed in a single,convenient location.The system has enhanced the operational efficiency of model experiments and improved user experience through methods such as case data governance and program code artifact caching.The system provides full support for the entire process of model experiments,including the design and definition of model experiments,data acquisition for experiment instances,creation and operation,job monitoring and management,and the visualization and analysis of results.The system has been developed using a framework that separates the front end from the back end.The front end incorporates load balancing and horizontal scaling,while the back-e

关 键 词：数值预报模式 模式试验 试验管理系统 在线 交互式 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]