中国部署首个台风24小时快速增强预报模型

中国国家气象中心已部署由深圳先进技术研究院开发的机器学习集合模型,用于台风快速增强预报。

该模型名为“热带气旋快速增强预报机器学习集合模型”,由李清兰领导的团队开发。它集成了四种机器学习算法,并在超过一半子模型预测时发布快速增强预报。

模型引入了海陆比和对称比两项定量指数,以揭示内核对称性与快速增强之间的物理联系。测试显示,该模型在2016至2020年北大西洋热带气旋模拟中,探测概率高于美国国家飓风中心,误报率更低。

国家气象中心高级工程师吕心妍表示,24小时快速增强预报技术为中国台风强度预报提供了重要参考。此前,2014年威马逊、2017年天鸽和2024年潭美等台风均经历快速增强并造成损失。

相关文章

Illustration of Typhoon Francisco bringing heavy rains and strong winds to Cebu
AI 生成的图像

Typhoon Francisco intensifies as habagat brings rains to Cebu

由 AI 报道 AI 生成的图像

Typhoon Francisco strengthened to maximum sustained winds of 140 kilometers per hour on Monday morning while enhancing the southwest monsoon.

Hong Kong authorities have launched an AI-driven flood forecasting and alert system capable of 3D simulations to enable timelier pre-emptive measures and emergency responses. It helps plan faster responses and preventive actions in high-risk areas. The technology adds to the territory-wide flood risk visualisation system deployed after last year's Super Typhoon Ragasa.

由 AI 报道

The Hong Kong Observatory said the developing El Nino could intensify four to seven tropical cyclones into super typhoons this year. It also forecast temperatures in the city could reach 35 degrees Celsius on Friday.

Hong Kong emergency crews used robotic water-pumping dragons to tackle flooding in the northern New Territories after intense downpours triggered a red rainstorm warning early Thursday.

由 AI 报道

The Shanghai Advanced Research Institute of the Chinese Academy of Sciences (CAS) has developed an AI model that calculates China’s 2022 greenhouse gas emissions as 17.7% lower than a widely used United Nations equivalent, while raising the US total by 15.2%. The system factors in both production and consumption, lessening the responsibility of exporters such as China.

此网站使用 cookie

我们使用 cookie 进行分析以改进我们的网站。阅读我们的 隐私政策 以获取更多信息。
拒绝