The possibility of a temporary increase of 1.5 degrees Celsius in the global average temperature over the next five years compared to pre industrial levels is 66%. "The warning issued by the World Meteorological Organization (WMO) in May is still in the ears, causing people to suffer from the cycle of summer exposure and cooking patterns, and the losses caused by Typhoon Dussuri's landfall are even more worrying. The basic situation of natural disasters in China in 2022 released by the Emergency Management Department shows that approximately 80 typhoons are generated globally each year, with an average of 7 typhoons landing in China. The direct economic losses caused by typhoon disasters in 2022 are as high as 5.42 billion yuan. Typhoon generation is frequent, and path prediction is difficult. The earlier humans obtain accurate information about extreme weather events such as typhoon movement paths, the more conducive it is to cope with typhoon attacks. Two independent studies by Chinese and American scientists recently published in the top international academic journal Nature have revealed the potential of artificial intelligence to assist in weather forecasting. AI weather forecasting has surpassed traditional numerical forecasting in terms of accuracy. In recent years, numerical weather forecasting methods have achieved great success in daily weather forecasting, extreme disaster warning, climate change prediction, and other fields. However, as the growth of computing power slows down and physical models become increasingly complex, the bottleneck of traditional numerical forecasting is becoming increasingly prominent, and researchers are starting to explore new meteorological forecasting paradigms to predict future weather. On July 6th, the official issue of Nature published the research results of the Huawei Cloud Pan Gu Big Model R&D team - "Accurate medium range global weather forecasting with 3D neural networks". This is the first paper in recent years published by a Chinese technology company as the only signatory unit in the journal, Breaking through the global challenge of AI predicting weather with lower accuracy than traditional numerical forecasting. The reviewers of "Nature" highly praised the achievement: "The Huawei Cloud Pan Ancient Meteorological Model has prompted people to re-examine the future of meteorological forecasting models, and the openness of the model will drive the development of this field." The research and development team of Huawei Cloud Pan Ancient Model introduced that in the most widely used fields of numerical methods, such as medium and long-term forecasting, the accuracy of existing AI forecasting methods is still significantly lower than that of numerical forecasting methods, and is subject to a lack of interpretability Constraints such as inaccurate extreme weather predictions. The main reasons for the insufficient accuracy of AI meteorological prediction models are: firstly, the original AI meteorological prediction models are based on 2D neural networks, which cannot handle uneven 3D meteorological data well. Secondly, the AI methods lack mathematical and physical mechanism constraints, which will continuously accumulate iteration errors during the iteration process. To this end, the team creatively proposed a 3D Earth Specific Transformer adapted to the Earth coordinate system to process complex and non-uniform 3D meteorological data, and used a hierarchical time-domain aggregation strategy to reduce the number of prediction iterations, thereby reducing iteration errors. Huawei's rotating chairman Hu Houkun stated that the Huawei Cloud Pan Gu model released in 2021 has penetrated into finance, manufacturing, government affairs, coal mines, railways, and other fields