Researchers have launched a system that dramatically reduces the time it takes to predict a storm
Weather forecasts are also changing in the age of Artificial Intelligence. Photo: Cherezoff | Dreamstime.com
A team of Hong Kong scientists has developed an artificial intelligence (AI) weather forecasting system capable of predicting thunderstorms and torrential rain up to four hours before they hit, compared to the current range of 20 minutes to two hours, Reuters reports.
The team from the Hong Kong University of Science and Technology said on Wednesday that the system will help governments and emergency service providers respond more effectively to the increasingly frequent extreme weather associated with climate change.
“We hope to use artificial intelligence and satellite data to improve forecasting of extreme weather events so that we can be better prepared,” said Su Hui, coordinating professor of the university's department of civil and environmental engineering, who led the project.
In a press conference to present the research published in December in the journal Proceedings of the National Academy of Sciences, Su pointed out that the system was designed to predict heavy rainfall.
How the 'AI weatherman' developed by Hong Kong researchers works
The developed model applies generative artificial intelligence techniques, introducing “noise” into the training data so that the system learns to reverse the process in an attempt to produce more accurate forecasts.
The team of Hong Kong scientists said it was developed in collaboration with China's meteorological authorities and that the system updates forecasts every 15 minutes and has increased accuracy by more than 15 percent.
Such work is crucial because the number of typhoons and rainy weather episodes experienced by Hong Kong and much of southern China in 2025 far exceeded seasonal norms, researchers say.
Last year, Hong Kong issued the highest rainstorm alert five times and the second most severe 16 times, setting new records, according to its weather watchdog.
Both the China Meteorological Administration and the Hong Kong Observatory are working to integrate the model into official forecasts. The team's new AI framework, called the Deep Diffusion Model based on Satellite Data (DDMS), was trained using temperature data collected between 2018 and 2021 by China's Fengyun-4 satellite.
Satellites can detect cloud formation earlier than other forecasting systems such as radar. The data were combined with meteorological expertise to capture the evolution of convective cloud systems and were subsequently validated with samples from the spring and summer of 2022 and 2023.
PHOTO article: Cherezoff / Dreamstime.com.
