The meteorological device demonstrates high performance in terms of speed and accuracy of forecasting compared to traditional solutions.
On July 5, 2023, a leading international scientific journal published an article on Pangu Weather, a revolutionary AI-based weather forecasting device. The article describes how to develop an accurate and reliable international weather forecasting system based on artificial intelligence and deep learning, using 43 years of data.
What is Pangu Weather
Pangu Weather is the first AI-based forecasting device that demonstrates better accuracy than traditional numerical methods. The invention makes it possible to speed up forecasting by 10,000 times, which will reduce the time of weather forecasting around the world to a few seconds.
Pangu Weather challenges previous assumptions that artificial intelligence-based weather forecasting is inferior in accuracy to traditional methods. Developed by the Huawei Cloud team, this first AI-based forecasting model is more accurate than previous solutions.
Thanks to the rapid development of computing power over the past 30 years, the accuracy of numerical forecasting has improved significantly. Thus, it became possible to provide warnings about natural disasters and forecasts about climate change. However, this method still remains time-consuming. To improve the speed of prediction, researchers studied how to apply deep learning methods. But the accuracy of AI-based forecasts for medium- and long-term scenarios remained lower compared to numerical forecasting. Artificial intelligence has typically been unable to predict extreme and unusual weather conditions, including typhoons.
About 80 typhoons occur in the world every year. According to China’s Ministry of Emergency Situations, direct economic losses from typhoons in China alone reached 5.42 billion yuan in 2022. The earlier you can receive warnings about natural disasters, the easier it will be to prepare for them.
Thanks to their speed, smart weather forecasting devices became popular, but they lacked accuracy for two reasons. First, existing meteorological forecasting systems based on artificial intelligence were created on the basis of 2D neural networks, which cannot effectively process non-uniform 3D meteorological data. Second, medium-term weather forecasts may suffer from cumulative forecast errors when the system is switched on too often.