As Tropical Storm Bebinca barrels toward waters off northern Taiwan gathering strength into a possible typhoon, weather forecasters in Taipei are using a new and so far successful method to help track its path: artificial intelligence (AI).
AI-generated forecasts, some powered by software from tech giants including Nvidia, whose chips are made by Taiwan’s homegrown semiconductor champion Taiwan Semiconductor Manufacturing Co, have so far outperformed traditional methods in predicting typhoon tracks.
In July, it was AI-based weather models, used for the first time, that helped Taiwan better predict the path and impact of Typhoon Gaemi, the strongest to strike the nation in eight years that brought record-breaking rainfall.
Photo: Liao Chen-huei, Taipei Times
The new technology impressed Taiwanese forecasters by predicting a direct hit as early as eight days before Gaemi made landfall, handily outperforming conventional methods, which remain the mainstay of prediction planning.
“People are starting to realize AI indeed delivered some stunning performances compared to conventional models,” Taiwan Integrated Disaster Prevention of Technology Engineering Consulting Co director Chia Hsin-sing (賈新興) said.
Bebinca is being tracked using the same AI tools by people including Lin Ping-yu (林秉煜), a forecaster at the Central Weather Administration (CWA), who said that AI has given the agency a higher degree of confidence there would not be a direct hit.
“This [AI] is a good thing for us. It is like having one more useful tool to use,” Lin said.
The AI weather programs on offer include Nvidia’s FourCastNet, Google’s GraphCast and Huawei’s Pangu-Weather, as well as a deep learning-based system by the European Centre for Medium-Range Weather Forecasts.
“It is a hotly watched competition. We will know soon who is winning,” Chia said.
Such AI models have also begun to be used to predict storms and hurricanes in other regions with good accuracy, forecasters and academics say.
The AI-based software is trained using historical weather data to learn the cause and effect relationships of meteorological systems and can predict hundreds of weather variables days in advance — a process that requires only a few minutes to complete.
For all the typhoons in the western Pacific this year until the middle of this month, AI’s accuracy in predicting storm tracks over a three-day window was nearly 20 percent higher than that of conventional models, data compiled by the CWA showed.
Ahead of Gaemi, AI helped the CWA foresee an unusual loop in its path that prolonged its impact on Taiwan and prompted it to swiftly issue a rare warning for rainfall of 1.8m, which was later proven accurate, CWA Deputy Administrator Lu Kuo-chen (呂國臣) said.
AI “boosted the confidence for forecasters to make that prediction,” Lu said, adding that the early warning gave extra time for authorities to carry out preparations.
Lu is also pinning hopes on a partnership with Nvidia, which this year announced a generative AI tool called CorrDiff that aims to forecast more precise locations of typhoon landfall and provide higher resolution images inside a storm.
“We are seeing the potential,” Lu said.
However, for now experts say the AI tools were not able to deliver quality forecasts for more detailed impact of a typhoon, such as its strength and winds, and more time is needed for the new technology to solidify its lead over more traditional ways.
“Was it just good luck?” Chia asked, pointing to AI’s stellar performance on Gaemi. “We need to give AI a bit more time. It is something to look forward to.”
AI is also being used to forecast wave patterns, with the system expected to be fully operational by 2026 in 16 of the 22 administrative regions in Taiwan, the CWA said.
Twelve monitoring stations and several forecasting systems have been established nationwide to improve safety for recreational water activities, the agency said.
Pilot programs are under way along the north coast and at the Northeast and Yilan Coast National Scenic Area, it said.
The CWA said it has partnered with local research institutions to develop rip current monitoring and forecasting tools.
Using AI, they are producing detailed wave forecasts to enhance the safety of coastal recreation and other operations, it added.
Additional reporting by CNA
Actor Darren Wang (王大陸) was sentenced to six months in prison, commutable to a fine, by the New Taipei District Court today for contravening the Personal Data Protection Act (個人資料保護法) in a case linked to an alleged draft-dodging scheme. Wang allegedly paid NT$3.6 million (US$114,380) to an illegal group to help him evade mandatory military service through falsified medical documents, prosecutors said. He transferred the funds to Chen Chih-ming (陳志明), the alleged mastermind of a draft-evasion ring, although he lost contact with him as he was already in detention on fraud charges, they said. Chen is accused of helping a
UNREASONABLE SURVEILLANCE: A camera targeted on an road by a neighbor captured a man’s habitual unsignaled turn into home, netting him dozens of tickets The Taichung High Administrative Court has canceled all 45 tickets given to a man for failing to use a turn signal while driving, as it considered long-term surveillance of his privacy more problematic than the traffic violations. The man, surnamed Tseng (曾), lives in Changhua County and was reported 45 times within a month for failing to signal while driving when he turned into the alley where his residence is. The reports were filed by his neighbor, who set up security cameras that constantly monitored not only the alley but also the door and yard of Tseng’s house. The surveillance occurred from July
SECURITY: Starlink owner Elon Musk has taken pro-Beijing positions, and allowing pro-China companies to control Taiwan’s critical infrastructure is risky, a legislator said Starlink was reluctant to offer services in Taiwan because of the nation’s extremely high penetration rates in 4G and 5G services, the Ministry of Digital Affairs said yesterday. The ministry made the comments at a meeting of the legislature’s Transportation Committee, which reviewed amendments to Article 36 of the Telecommunications Management Act (電信管理法). Article 36 bans foreigners from holding more than 49 percent of shares in public telecommunications networks, while shares foreigners directly and indirectly hold are also capped at 60 percent of the total, unless specified otherwise by law. The amendments, sponsored by Chinese Nationalist Party (KMT) Legislator Ko
A Japan Self-Defense Forces vessel entered the Taiwan Strait yesterday, Japanese media reported. After passing through the Taiwan Strait, the Ikazuchi was to proceed to the South China Sea to take part in a joint military exercise with the US and the Philippines, the reports said. Japan Self-Defense Force vessels were first reported to have passed through the strait in September, 2024, with two further transits taking place in February and June last year, the Asahi Shimbun reported. Yesterday’s transit also marked the first time since Japanese Prime Minister Sanae Takaichi took office that a Japanese warship has been sent through the Taiwan
RSS