Text size

ADDS-Reax from Professor at the University of Manchester

A new artificial intelligence-based weather model can deliver 15-day forecasts with unmatched accuracy and speed, a Google lab said, with potentially life-saving applications as climate change worsens.

GenCast, invented by London-based AI research lab Google DeepMind, “demonstrated better predictive capabilities” than the current world-leading model, the company said on Wednesday.

The European Center for Medium-Range Weather Forecasts (ECMWF) produces forecasts for 35 countries and is considered the global benchmark for meteorological accuracy.

However, according to DeepMind, GenCast exceeded the precision of the center’s predictions in more than 97 percent of the 1,320 real-world scenarios from 2019 on which both were tested.

Details of the results were published in Nature, a leading scientific journal.

ECMWF chief Florence Rabier told AFP the project was a “first step” towards integrating AI into weather forecasting, but “it is actually a leap forward”.

At this point, it can be used as a complement to current models, she said.

“We are making progress every year,” she added. “Any new method that can strengthen and accelerate this progress is extremely welcome given the extreme societal pressures of climate change.”

The model was trained on four decades of temperature, wind speed and barometric pressure data from 1979 to 2018 and can produce a 15-day forecast in just eight minutes – compared to the hours it currently takes.

“GenCast provides better predictions of both daily weather and extreme events than the top operating system… up to 15 days in advance,” DeepMind said in a statement.

DeepMind said GenCast has “consistently outperformed” the current leading forecast model in predicting extreme heat, extreme cold and high wind speeds.

“More accurate predictions of the risks of extreme weather can help authorities protect more lives, prevent damage and save money,” DeepMind said.

Due to human-caused climate change, extreme weather events are becoming more frequent and more severe.

In August 2023, around 100 people died in a series of forest fires in Hawaii. Authorities faced criticism from locals who said they were not warned about the impending fire.

This summer, a sudden heat wave in Morocco killed at least 21 people in 24 hours. And in September, Hurricane Helene killed 237 people in Florida and other southeastern U.S. states.

“I am confident that AI-based weather forecasting systems will continue to improve incrementally in the future, including better prediction of extreme events and their intensity, where there is still much room for improvement,” said David Schultz, Professor of Synoptic Meteorology at the University of Manchester, who was not involved in the research.

However, he said these forecasting systems rely on weather forecast models already running, such as the model operated by ECMWF.

jts/klm/bjt