唐健, 代刊, 宗志平, 曹勇, 刘凑华, 高嵩, 于超
Methods and Platform Realization of the National QPF Master Blender
(National Meteorological Center of CMA)
本文已被:浏览 199次   下载 53
投稿时间:2017-11-08    修订日期:2017-12-20
中文摘要: 随着天气预报业务现代化的发展,预报员面临气象数据爆发增长、服务前端需求不断提高以及客观预报技术广泛应用带来的挑战。传统以手工绘制降水落区为主的定量降水预报(QPF)技术流程已经不能帮助预报员在更高层面体现附加值。为支撑预报员在QPF流程中的核心作用,设计和开发了主客观融合QPF平台。该平台从海量预报数据选取、多源QPF集成、QPF调整和订正、格点化处理和服务产品制作五个方面帮助预报员控制数字化预报全流程。发展了多项关键技术支持平台的智能化,包括多模式QPF数据集构建技术、多模式QPF集成技术、QPF预报场调整和订正技术以及格点场后处理技术。基于MICAPS4系统,实现了主客观融合QPF平台的主要功能,发布了“QPF Master Blender 1.0”版本,并在2017年5月投入业务应用,取得良好反馈和效果。最后,对平台的未来发展进行了展望,包括发展数值模式检验评估工具支持预报员做出最优判断,研究多尺度模式信息的融合技术等。
Abstract:With the development of the weather forecast modernization, the forecasters are facing challenges brought by meteorological data explosion, the increasing demand of the service front end as well as the wide use of objective forecasting technology. Tradition Quantitative Precipitation Forecast (QPF) routine, which mainly based on manually plotting precipitation areas, can no longer assist forecasters in demonstrating added value at a higher level. To support the forecasters’ central role in the QPF routine, a subjective and objective QPF blender was designed and developed. This platform helps forecasters to take control of the whole process of numerical forecast from the following five aspects: selection from mass forecast data, integration of multi-source QPF, adjustment and correction of QPF, grid processing and service product production. The intelligence of the platform is secured by the development of a number of key supporting techniques, including multi-model QPF dataset construction technology, multi-model QPF integration technology, QPF field adjustment and correction techniques and gridded QPF post-processing technology. Based on MICAPS4, the main functions of this QPF platform were realized. The "QPF Master Blender 1.0" version was released and put into operational use in May 2017, which has yielded good feedback and effectiveness. By the end of this paper, the future development of the platform is prospected, including the development of numerical model verification tools to support forecasters to make the best judgments, and research on the fusion technologies of multi-scale model information.
文章编号:201711080484     中图分类号:    文献标志码:
唐健,代刊,宗志平,曹勇,刘凑华,高嵩,于超,0.[en_title][J].Meteor Mon,():-.
TANGJIAN,DAIKAN,ZONGXHIPING,CAOYONG,LIUCOUHUA,GAOSONG,YUCHAO,0.Methods and Platform Realization of the National QPF Master Blender[J].Meteor Mon,():-.