Core Insights - The article discusses China's ambition to become a meteorological superpower by developing its own meteorological data sets to reduce reliance on European data sources, particularly the ERA5 data set, which is crucial for AI-driven weather forecasting [1][2]. Group 1: Strategic Importance of Meteorological Data - Meteorological data has become a core strategic resource for nations, surpassing its original purpose of weather forecasting [1]. - The ERA5 data set, which integrates global observational data, is essential for assessing climate trends and improving forecasting accuracy, with an estimated annual economic value of several hundred million dollars [1]. Group 2: Development of Domestic Data Sets - The National Data Bureau of China has initiated a global atmospheric reanalysis system project to eliminate dependence on Western reanalysis products [2]. - The newly released CMA-RA V1.5 data set is already being utilized by domestic AI weather models for training purposes [2]. Group 3: Technological Innovations - The domestic data set has achieved significant technological breakthroughs, transitioning from "catching up" to "leading in certain areas" through three main innovations: 1. Upgraded assimilation technology, improving satellite data assimilation by 13% over the first 20 years [4]. 2. Self-controllable domestic observational data, integrating data from 116 satellites, with a significant portion being domestic [4]. 3. Internationally leading resolution and timeliness, with a model resolution of 13 kilometers and hourly updates, outperforming ERA5 [4]. Group 4: International Industry Attention - The opening of the domestic data set has attracted international industry interest, with companies like Vaisala exploring potential applications for weather derivatives [5]. - The integration of multiple data sets, including CMA-RA V1.5, is seen as beneficial for researchers studying climate change and extreme weather, as well as for wind farm project developers [5][6].
气象预报事关国安,“中国要打破欧洲依赖”
Xin Lang Cai Jing·2025-12-23 04:36