Researchers from the University of the Philippines Diliman (UPD) have developed an artificial intelligence (AI) model capable of predicting rainfall from tropical cyclones (TCs) more quickly and efficiently than traditional forecasting methods.

The study, conducted by Cris Gino Mesias and Dr. Gerry Bagtasa from the UP Diliman College of Science’s Institute of Environmental Science and Meteorology (IESM), uses machine learning to link past typhoon tracks with recorded rainfall.

The model identifies patterns in typhoon behavior to forecast potential rainfall distribution.

“Most predictions of TC rainfall rely on dynamic models, which are very difficult to run as they take a lot of computational resources and require high-performance computing,” Bagtasa said in an article published on UPD’s College of Science website.

In contrast, the AI model can run on a standard laptop within minutes.

“When we assessed the AI model, its predictive skill was comparable to a dynamic model that we regularly use. The AI model had better skills for extreme rainfall from tropical cyclones,” he added.

Bagtasa said the model’s forecasts are influenced primarily by a typhoon’s proximity and duration.

For example, typhoons closer to a region or those that move slowly tend to produce heavier rainfall in affected areas.

“This AI model, admittedly, is not perfect. But it can add to the suite of rainfall forecast models available to equip our disaster managers with more information on impending hazards,” he said.

While the AI model is not intended to replace existing systems, it adds to the range of forecasting tools available.

It can also be updated with new data, allowing it to improve over time.

Bagtasa also noted the importance of AI literacy, saying not all AI systems function in the same way.

Unlike large language models such as ChatGPT or Gemini, which require significant computing power, the weather-focused AI developed by UP scientists is more energy-efficient.

The study, titled “AI-Based Tropical Cyclone Rainfall Forecasting in the Philippines Using Machine Learning,” is published in Meteorological Applications.

It was supported by the Department of Science and Technology–Accelerated Science and Technology Human Resource Development Program and DOST-Philippine Council for Industry, Energy, and Emerging Technology Research and Development.