A new study has suggested that AI weather forecasts can produce predictions of similar accuracy, faster and cheaper than traditional methods.
The University of Reading study, published in npj Climate and Atmospheric Science, highlights the rapid progress of AI weather forecasts.
Professor Andrew Charlton-Perez, who led the study, said: “AI is transforming weather forecasting before our eyes. Two years ago, modern machine learning techniques were rarely being applied to make weather forecasts. Now we have multiple models that can produce ten-day global forecasts in minutes.
“There is a great deal we can learn about AI weather forecasts by stress-testing them on extreme events like Storm Ciarán. We can identify their strengths and weaknesses and guide the development of even better AI forecasting technology to help protect people and property.”
Comparing AI and physics-based forecasts
The scientists compared AI and physics-based weather forecasts for Storm Ciarán, which hit northern and central Europe in November 2023. The storm claimed 16 lives in northern Europe and left millions of homes without power in France.
The researchers used four AI models and compared their results with traditional physics-based models.
The AI models were able to predict the storm’s rapid intensification and track 48 hours in advance.
The AI weather forecasts were said to be indistinguishable from the performance of conventional forecasting models.
They also captured the large-scale atmospheric conditions that fuelled Ciarán’s explosive development, such as its position relative to the jet stream.
Underestimation of the storm
However, the machine learning technology underestimated the storm’s damaging winds.
All four AI forecasting systems underestimated Ciarán’s maximum wind speeds, which, in reality, gusted at speeds of up to 111 knots at Pointe du Raz, Brittany.
The team showed that the underestimation was linked to some of the storm’s features that the AI systems could not predict well.
Further investigation of the use of AI is needed
The researchers argue that further investigation of the use of AI in weather production is urgently needed to protect people from extreme weather events.
The development of machine learning models could mean that AI weather forecasts will be routinely used in the future, saving forecasters time and money.
