短临降水精准预报大模型

采用“气象雷达监测数据+人工智能算法、深度学习、雷达去噪”等技术开发实现，可接入国家气象局的全国多普勒天气雷达拼图产品、中国地面2170站6要素逐小时数据、华东区域数值预报产品，通过模型实现预报未来2小时逐分钟、逐公里降雨情况，可接入私有气象雷达数据，实现未来2小时短临降水预报。This dataset is developed by adopting technologies including meteorological radar monitoring data, artificial intelligence algorithms, deep learning, and radar denoising. It supports integration with national Doppler weather radar mosaic products issued by the China Meteorological Administration, hourly 6-element observation data from 2170 national ground meteorological stations across China, and numerical weather prediction products for the East China region, and can generate minute-by-minute, kilometer-by-kilometer rainfall forecasts for the next 2 hours via the trained model. Moreover, it allows access to private meteorological radar data to produce short-term nowcasting precipitation forecasts for the upcoming 2 hours.