###
本文二维码信息
国家级智能数字天气预报业务的技术进展
金荣花, 曹勇, 赵瑞霞, 代刊, 郭云谦, 徐珺, 曾晓青, 王玉, 唐健, 韦青
(国家气象中心)
Technological Advances in National-Level Intelligent Digital Weather Forecasting Operational Systems
JIN Ronghua, CAO Yong, ZHAO Ruixia, DAI Kan, GUO Yunqian, XU Jun, ZENG Xiaoqing, WANG Yu, TANG Jian, WEI Qing
(National Meteorological Centre)
摘要
相似文献
本文已被:浏览 134次   下载 153
投稿时间:2025-01-09    修订日期:2025-05-19
中文摘要: 智能数字天气预报是气象服务保障的关键手段,全球各国积极发展新一代无缝隙预报技术系统并推动人工智能在气象领域应用。我国构建了完备的智能数字预报业务系统,中央气象台实现全国 1km、全球 5km分辨率的近地面及三维0-30 d无缝隙数字预报。通过区分时效和要素构建适配性技术、以及多源预报融合的总体思路发展智能数字预报技术体系。研发了统一、规范及模块可扩展的智能数字通用技术框架,实现30多类算法,支持 “低代码” 部署,在重大活动保障和极端天气预报中发挥重要作用。实现了深度学习等人工智能技术的深度应用,在短时、短中期降水及强对流、灾害性大风预报等方面,通过构建考虑物理机理约束的深度学习模型,显著提升预报性能,精细化降尺度技术也取得进展。同时,多源预报融合集成技术不断发展,主客观智能融合预报提升了灾害性天气预警能力。中央气象台智能数字预报准确率总体较欧洲中期天气预报中心和中国气象局主流数值模式提高10%以上,在多领域应用广泛。但仍面临挑战,未来将针对灾害性和转折性天气预报、人工智能和物理机理的有效融合、专业气象和风险预报扩展、以气象“智脑”为核心的一体化平台等方面的技术进行突破。
Abstract:Intelligent digital weather forecasting is a key means to support meteorological service. Countries around the world are actively developing new-generation seamless forecasting technology systems and promoting the application of artificial intelligence in the meteorological field. China has established a relatively complete intelligent digital weather forecasting operation system, achieving seamless forecasting for the near-surface and three-dimensional meteorological elements with a resolution of 1-km across the country and 5-km globally, covering a time range of 0-30 days. By developing adaptable technologies through the strategy of implementing different strategies at different time scales and integrating multi-source forecast, remarkable results have been achieved. A unified, standardized, and modularly expandable intelligent digital general technology framework has been constructed, constructing more than 30 types of algorithms and supporting "low-code" deployment, which plays an important role in major event support and extreme weather forecasting. The in-depth application of artificial intelligence technology has significantly improved forecasting performance in short-term and short-to-medium-term precipitation forecasting, severe convective weather forecasting, and disastrous gale forecasting by leveraging deep learning models. Progress has also been made in refined downscaling technology. At the same time, the data fusion and integration technology has continued to develop, and the intelligent integration of objective and subjective forecasting has developed to enhance the ability to forecast disastrous weather. China"s intelligent digital forecasting has improved the accuracy by 10% - 31% compared with the EC_IFS and CMA models and is widely applied in many fields. However, it still faces challenges. In the future, breakthroughs will be made in technologies related to the forecasting of disastrous and transitional weather, low-altitude hundreds-meter-resolution refined downscaling forecasting, industry-specific weather and risk forecasting, and the integrated platform with the meteorological "intelligent brain" as the core.
文章编号:202501090010     中图分类号:    文献标志码:
基金项目:
Author NameAffiliationAddress
JIN Ronghua National Meteorological Centre 北京市海淀区中关村南大街46号
CAO Yong  
ZHAO Ruixia  
DAI Kan  
GUO Yunqian  
XU Jun  
ZENG Xiaoqing  
WANG Yu  
TANG Jian  
WEI Qing  
引用文本:
JIN Ronghua,CAO Yong,ZHAO Ruixia,DAI Kan,GUO Yunqian,XU Jun,ZENG Xiaoqing,WANG Yu,TANG Jian,WEI Qing,0.Technological Advances in National-Level Intelligent Digital Weather Forecasting Operational Systems[J].Meteor Mon,():-.
JIN Ronghua,CAO Yong,ZHAO Ruixia,DAI Kan,GUO Yunqian,XU Jun,ZENG Xiaoqing,WANG Yu,TANG Jian,WEI Qing,0.Technological Advances in National-Level Intelligent Digital Weather Forecasting Operational Systems[J].Meteor Mon,():-.