Taiwan uses AI to predict typhoon trajectories amid Bebinca's approach
As tropical storm Bebinca approaches northern Taiwan, the country's meteorologists are turning to artificial intelligence (AI) for accurate predictions. This innovative approach, supported by software from tech giants like NVIDIA, has proven more effective than traditional methods in forecasting typhoon trajectories. The chips used in these AI systems are manufactured by TSMC, a leading semiconductor company based in Taiwan.
Conventional methods in typhoon prediction outperformed
In July, Taiwan successfully used AI-based weather models to predict the path and impact of Typhoon Gaemi. This storm was the most powerful to hit the island in eight years, bringing record-breaking rainfall. The AI technology accurately predicted a direct hit eight days before Gaemi made landfall, surpassing traditional prediction methods. Chia Hsin-sing from Taiwan Integrated Disaster Prevention of Technology Engineering Consulting Company Ltd., stated "people are starting to realize AI indeed delivered some stunning performances compared to conventional models."
AI boosts confidence in weather forecasting
The same AI tools are now being used to track Bebinca. Lin Ping-yu from Taiwan's Central Weather Administration (CWA), stated that "AI has given them a higher degree of confidence there will not be a direct hit." The AI weather programs include NVIDIA's FourCastNet, Google's GraphCast, Huawei's Pangu-Weather, and a deep learning-based system by the European Centre for Medium-Range Weather Forecasts. These models have also been successfully used to predict storms and hurricanes in other regions with good accuracy.
Predictive power in weather forecasting
The AI-based system is trained using historical weather data. It can predict hundreds of weather variables days in advance, a process that takes only a few minutes to complete. For all typhoons in the Western Pacific this year up until this month, AI's accuracy in predicting storm tracks over a three-day window was nearly 20% higher than that of traditional models, according to data compiled by the CWA.
Role in early warning and future prospects
Ahead of Gaemi, AI helped the administration foresee an unusual loop in its path that prolonged its impact on Taiwan. This led them to swiftly issue a rare warning for rainfall of 5.9-feet, which was later proven accurate, according to CWA's deputy head Lu Kuo-Chen. However, experts note that while AI tools have shown promise in predicting typhoon paths, they are yet to deliver quality forecasts for more detailed impacts such as strength and winds.
