深圳加速探索人工智能赋能精准气象预报，世界气象组织首个AI试点项目亚洲临近预报比对活动在深 启动。 12月7日，全国第十二届残疾人运动会暨第九届特殊奥林匹克运动会马拉松项目与2025年深圳马拉松比 赛，先后在市民中心起跑。两项赛事史无前例地共用了一条比赛路线，统一配置医疗急救、志愿者服 务、安全交通保障和气象服务等资源。 开赛前夕，竞赛指挥部MOC(赛事指挥中心)马拉松赛事气象专题页面上，赛道沿线近气象自动站和生物 舒适度仪的实时数据和精细化预报，正一组组不断向赛事竞委会同步，为决策部门提供着赛事支持。这 套经过全运会深港跨境马拉松检验的赛事气象服务"深圳方案"，正是深圳加快推动人工智能气象应用创 新生态体系建设的生动实践之一。 实际上，深圳对于人工智能在气象预报领域的应用不只在于气象服务保障，更是从2017年就开始聚焦超 大城市气象防灾减灾中，灾害性天气预报"报不早、发不快、说不清"等痛点问题，从顶层开始着手设 计，以《深圳市加快推进气象高质量发展的若干措施》为指导、业务需求为引领，构建人工智能与气象 深度融合的协同创新体系，通过联合创新研发"智霁""智瞳""阿福"等一系列AI气象预报服务核心技术， 推动气 ...

Core Insights - The launch of the first World Meteorological Organization (WMO) Artificial Intelligence Near-Term Weather Forecasting Demonstration Project (AINPP) in Asia marks a significant milestone in advancing modern early warning and forecasting technologies [3][5] - The project aims to enhance collaboration among multiple disciplines, countries, and public-private sectors to develop AI-based near-term forecasting models, particularly in regions vulnerable to extreme weather events [3][6] Group 1: Project Overview - The AINPP project is designed to set a global standard for AI applications in meteorology, focusing on improving the accuracy of early warning systems to save lives and reduce losses associated with severe weather [3][6] - The project involves participation from meteorological experts from various countries, including Japan, South Korea, Thailand, Vietnam, India, UAE, and Saudi Arabia, as well as institutions like the Hong Kong Observatory and Microsoft [3][6] Group 2: Technological Advancements - AI-driven models integrate various data sources, such as satellite images, radar networks, and ground sensors, to enhance the accuracy and delivery time of local extreme weather predictions [5] - Shenzhen's meteorological bureau has developed AI applications for early warning services, including the "Zhiji" regional forecasting model with a 3 km resolution and the "Zhitong" short-term forecasting model, which extends effective warning time from 1 hour to 3 hours [6][7] Group 3: Future Plans and Collaboration - The AINPP project has collected 13 AI-based forecasting technologies related to lightning, precipitation, and radar echoes, which will be tested during the upcoming flood season [7] - Shenzhen's meteorological bureau plans to establish a data-sharing platform and engage in cross-regional product testing and application, contributing to disaster prevention and mitigation efforts in developing countries [7]