Rongcheng.Kuiniu Model Launched for Short-range Meteorological Forecast

The Chengdu Institute of Plateau Meteorology (CIPM), University of Electronic Science and Technology of China (UESTC), and Huawei jointly built the Ascend AI-powered Rongcheng.Kuiniu, China's first foundation model to predict short-range precipitation, thereby applying AI technologies to the meteorological field for natural disaster pre-warning.

Short-range weather forecast is closely related to our daily lives, as sudden extreme weather may cause loss of lives and properties. Timely and accurate short-range weather forecast can guide disaster prevention and mitigation. To meet this demand, three parties cooperated and agreed to build a new short-range weather forecast system by combining AI and big data technologies with radar and satellite resources. This will help better understand the scientific rules of extreme weather and implement numerical weather prediction with 1-kilometer resolution and 1-hour forecast frequency.

At the signing ceremony was Mr. Jiang Xingwen, Director of CIPM, who first explained the impact of disasters caused by short-range extreme weather in recent years, before listing the common problems of conventional numerical forecast models in predicting the frequency and arrival time of bad weather. He ended his speech stating that CIPM was and would always be committed to combining meteorology with AI to solve these problems.

Built on the superb AI-enabled computing power provided by the Chengdu Intelligent Computing Center, the MindSpore-based Rongcheng.Kuiniu significantly improves the frequency and accuracy of prediction, to slash time in disaster prevention. In addition to precipitation forecast, foundation models can also predict temperature, pressure, humidity, and wind speed and direction, signaling a promising future in fields such as energy, manufacturing, transportation, and urban disaster management.

Representing UESTC was Mr. Shao Jie, Professor of School of Compute Science and Engineering, who said that the advantages of AI and talented personnel in university were perfectly complemented by the other two parties: extensive meteorological data provided by the CIPM, and the strong technical capabilities of the Ascend AI base software and hardware platform at Huawei. Shao predicted that AI technologies would play a crucial role in meteorology, and help reduce the loss of life and property caused by short-range disastrous weather.

Another speaker was Mr. Tian Kunyang, Product Director of MindSpore, who discussed the evolution of Rongcheng.Kuiniu. It was, he explained, a flagship of the three parties and a pioneer in exploring AI application in meteorological field, which would directly benefit the public. Fully utilizing the advantages of the three parties, Rongcheng.Kuiniu would be the catalyst for developing new technologies that break performance issues typically found in current weather prediction models.

Ascend AI has entered a new phase of development. Together with other partners, Huawei will continue to innovate AI technologies, to drive the digital economy and promote the intelligent upgrade for all industries.