China deploys first 24-hour rapid intensification forecast model for typhoons-Xinhua

Researchers at the Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences, have deployed a machine learning ensemble model that improves the accuracy of 24-hour and 12-hour typhoon rapid intensification forecasts. Now operational at China’s National Meteorological Center, the system uses inner-core symmetry and sea-land ratios to predict sudden, destructive wind speed spikes.

What is typhoon rapid intensification and why is it dangerous?

Rapid intensification occurs when a typhoon’s maximum sustained wind speed increases by more than 15 meters per second within 24 hours, or more than 10 meters per second within 12 hours. These sudden jumps in power often leave coastal cities with little time to prepare.

According to SIAT researchers, these events are highly destructive. The team cited typhoons such as Rammasun in 2014, Hato in 2017, and Yagi in 2024 as examples of storms that underwent rapid intensification before landfall, resulting in significant economic losses and casualties.

The challenge is so significant that the China Association for Science and Technology named forecasting rapid intensification one of the top 10 frontier scientific problems in 2025.

Did you know? A wind speed increase of 15 meters per second is roughly equivalent to a jump of 33 miles per hour in a single day, which can shift a storm’s category rapidly.

How does the SIAT machine learning model work?

The new “Machine Learning Ensemble Model for Tropical Cyclone Rapid Intensification Forecast” identifies physical links between a storm’s structure and its potential to strengthen. Li Qinglan, the SIAT team lead, explained that typhoon intensity is driven by interacting factors like inner-core structure, environmental backgrounds, and land-sea surface interactions.

To track these, the team created two specific quantitative indices:

The Sea-Land Ratio: This tracks how the distribution of land and sea changes along the typhoon’s path.
The Symmetric Ratio: This measures the symmetry of the inner-core convective activity.

Li noted that a typhoon typically develops a highly symmetric, ring-like structure in its inner core just before rapid intensification. The more symmetric the core, the higher the likelihood the storm will strengthen quickly.

The system integrates four different machine-learning algorithms. It only issues a rapid intensification forecast if more than half of these sub-models predict the event, a method designed to filter out noise and increase accuracy.

How does this system compare to U.S. forecasting standards?

The SIAT team validated their model by simulating 24-hour rapid intensification events in the North Atlantic between 2016 and 2020. They compared their results directly against the operational forecast system used by the U.S. National Hurricane Center (NHC).

The results showed the SIAT model achieved a higher probability of detection and a lower false alarm rate than the NHC system. This indicates the ensemble approach is more effective at spotting genuine intensification events while ignoring false signals.

Pro Tip: When tracking storms, always cross-reference local meteorological center alerts with satellite imagery of the storm’s “eye” symmetry, as this is a primary indicator of strengthening.

What happens next for China’s storm warnings?

The model is now integrated into the National Meteorological Center’s operational workflow. Lyu Xinyan, a senior engineer at the center, stated that this 24-hour rapid intensification technology now serves as a critical reference for China’s overall typhoon intensity forecasting.

By reducing false alarms and increasing the detection rate of sudden spikes, authorities can make more informed decisions regarding evacuations and infrastructure lockdowns. This shift moves forecasting from general probability toward concrete, data-driven warnings.

Frequently Asked Questions

What is the main difference between this model and older methods?

According to Li Qinglan, conventional statistical-dynamical methods often fail to capture the nonlinear characteristics of how a typhoon’s intensity changes. The SIAT model uses machine learning to identify these complex, non-linear patterns.

Which storms are mentioned as examples of rapid intensification?

The researchers highlighted Typhoon Rammasun (2014), Typhoon Hato (2017), and Typhoon Yagi (2024) as storms that intensified quickly before landfall.

Where is the model currently being used?

The model has completed real-world application testing and is now deployed at China’s National Meteorological Center.

Want to stay updated on weather technology and disaster prevention? Share your thoughts in the comments below or subscribe to our newsletter for the latest insights into climate science.