Weather forecasters in Taipei, Taiwan are using a new AI powered method to help track the path of tropical storm Bebinca, according to a report from Reuters.

The new technology includes a range of software such as Nvidia's FourCastNet, Google's GraphCast and Huawei's Pangu-Weather, as well as a deep learning-based system by European Centre for Medium-Range Weather Forecasts.

The AI-based software uses historical weather data to decipher the cause-and-effect relationships of meteorological systems and is able to predict hundreds of weather variables days in a few minutes.

The forecasting method represents a significant leap forward for Taiwan, offering improved precision in tracking typhoons compared to traditional methods.

Examining data related to Bebinca, forecaster at Taiwan's Central Weather Administration (CWA) Lin Ping-yu said AI predictions confirmed with a high degree of certainty there will not be a direct hit.

"This (AI) is a good thing for us,” he said. “It is like having one more useful tool to use.”

The new model is also being used in other regions to help predict storms, cyclones and hurricanes as it offers higher accuracy compared to traditional methods, forecasters and academics told Reuters.

According to research by CWA, AI accuracy has been over 20 per cent higher than that of traditional methods in predicting typhoons formations and movements in the Western Pacific this year.

In July, an AI-based weather model confirmed previsions of a direct hit of typhoon Gaemi eight days before it made landfall, overpassing traditional forecasting methods.

Typhoon Gaemi was the strongest storm to hit Taiwan in eight years and killed three people, injuring hundreds and causing widespread flooding. It also caused power outages that reportedly affected half a million households.

Chia Hsin-sing, director at the weather service provider Taiwan Integrated Disaster Prevention of Technology Engineering Consulting Company Ltd, told Reuters: "People are starting to realise AI indeed delivered some stunning performances compared to conventional models."

"It is a hotly watched competition. We will know soon who is winning," he added.

According to some experts, AI technology is still unable to provide more accurate data forecasts regarding the impact of typhoons, including its strengths and winds.

"Was it just good luck?" said Chia, referring to the success of AI in predicting Gaemi. "We need to give AI a bit more time. It is something to look forward to."

---