Huawei Cloud, in collaboration with the Meteorological Bureau of Shenzhen Municipality, has launched a first-of-its-kind regional artificial intelligence (AI) model to facilitate weather forecasting. The officially introduced model allows short-term forecasts spanning a five-day period to be created for an accurate 3 km range, a significantly improved scope compared to traditional global models, which operate within a 25 km range.

Named the Zhiji Regional Model, this system leverages Huawei Cloud's Pangu-Weather Model and taps into high-quality regional datasets to generate five-day forecasts with a resolution of 3 km for Shenzhen and its adjacent areas. Departing from the broad perspective of typical global models with a 25km spatial resolution, the Zhiji model delivers forecasts including temperature, precipitation, and wind speed, covering a variety of meteorological elements.

The regional model has shown a marked improvement in capturing smaller-scale meteorological details compared to global weather models. This was evidenced by the regional model's high precision in predicting multiple cold-temperature periods in Shenzhen during its trial run launched in February this year.

"Weather forecasting is crucial to disaster prevention and mitigation," stated Lan Hongping, Deputy Director of the Meteorological Bureau of Shenzhen Municipality. He added that the Bureau has been consistently exploring ways to fine-tune weather forecasting and warning services, viewing AI as a key tool in achieving this. Lan is optimistic about this innovation with Huawei Cloud and anticipates that it will lead to superior services that will better address the needs of Shenzhen's residents and the city's future development.

William Dong, President of Huawei Cloud Marketing Dept, shared that the regional AI weather model's introduction for Shenzhen epitomises how AI can facilitate precise, smaller-scale weather forecasts. He voiced the importance of this scientific development following the publication of a paper on the Pangu-Weather Model in the established scientific journal Nature last July, showing that these scientific discoveries can be adapted to provide tangible benefits beyond the laboratory realm.

Going forward, Huawei Cloud has plans to continue innovating AI weather forecasting methods and broaden their applications to benefit more sectors, refine weather services and enhance disaster preparedness. Bearing in mind that weather, climate, and water-related hazards have created nearly 12,000 disasters between 1970 and 2021 globally, resulting in the deaths of over two million people and $US 4.3 trillion in economic damages, early warnings can significantly decrease casualties and prevent economic losses. The precision of weather prediction directly relates to the effectiveness of these early warnings. Thus, AI's ability to process substantial amounts of information and identify patterns therein can markedly enhance the precision of weather forecasting and the speed of prediction generation.

In July 2023, a research paper on Huawei Cloud's Pangu-Weather Model was featured in the leading science magazine Nature. This regional AI model amplifies Pangu-Weather's scientific advancements. As the Southern China monsoon season draws near, Huawei Cloud and the Meteorological Bureau of Shenzhen Municipality plan to further verify and comprehensively evaluate the regional AI weather forecasting model during this season and continue refining the model for weather forecasters' utility.