SHENZHEN, June 15 (Xinhua) -- The deployment of a new 24-hour rapid intensification forecast model for typhoons and its 12-hour rapid intensification forecast service now offer China significantly enhanced forecasting performance, according to Shenzhen-based researchers in south China.
The model, developed at the Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences, recently completed operational deployment and real-world application testing at the country's National Meteorological Center.
In meteorology, a typhoon is defined as undergoing rapid intensification when its maximum sustained wind speed increases by more than 15 meters per second within 24 hours, or by more than 10 meters per second within 12 hours.
The weather events can be highly destructive. Typhoons such as Rammasun in 2014, Hato in 2017, and Yagi in 2024 all underwent rapid intensification before landfall, causing significant casualties and economic losses.
In 2025, forecasting typhoon rapid intensification was selected as one of the top 10 frontier scientific problems by the China Association for Science and Technology.
The new model, "Machine Learning Ensemble Model for Tropical Cyclone Rapid Intensification Forecast," was developed by the SIAT team led by Li Qinglan.
According to Li, the evolution of typhoon intensity is controlled by multiple interacting factors that include inner-core structure, environmental background and land-sea surface interactions, making accurate forecast extremely difficult.
Conventional statistical-dynamical methods also fail to capture the nonlinear characteristics of intensity changes. As a result, forecasting typhoon intensity, especially rapid intensification, remains a persistent challenge in the field.
To solve this, the team established two quantitative indices -- the sea-land ratio, which captures variations in land-sea distribution along a typhoon track, and the symmetric ratio, which describes inner-core convective symmetry. The indices reveal physical links between inner-core symmetry and rapid intensification.
"Prior to rapid intensification, a typhoon's inner core typically develops a highly symmetric ring-like structure. A more symmetric inner core indicates a higher likelihood of rapid intensification occurrence," Li explained.
As part of this innovation, the research team integrated four machine-learning algorithms into an ensemble forecast model. When more than half of the sub-models predict rapid intensification, the system issues a rapid intensification forecast, effectively improving forecast accuracy.
The team tested the new model by simulating 24-hour rapid intensification events of tropical cyclones in the North Atlantic from 2016 to 2020 and compared its performance with the operational forecast system from the U.S. National Hurricane Center. The results showed that the new model achieved a higher probability of detection and a lower false alarm rate, demonstrating enhanced forecasting performance and operational viability.
The National Meteorological Center's senior engineer Lyu Xinyan said 24-hour rapid intensification forecast technology now provides an important reference for China's typhoon intensity forecasting. ■
