\n
<\/p>\n
AI is being touted as the future of weather forecasting\u2014faster and more precise. But new research shows a major blind spot: it often fails at predicting extreme weather. Traditional physics-based models still do better.<\/p>\n<\/div>\n
\u201cThey do perform well on a lot of tasks, but for very extreme events\u2014that are the most important for society\u2014they still struggle,\u201d says Sebastian Engelke, a statistics professor at the University of Geneva and one of the authors of a new study<\/a> in Science<\/em> that pitted some of the leading AI weather models, including GraphCast<\/a> and Pangu-Weather<\/a>, against a database of recent extreme events.<\/p>\n For record-breaking heat, like a heat wave in Siberia in early 2020 that led to wildfires and melting permafrost, AI predictions tend to underestimate high temperatures. (The heat wave would have been almost impossible without climate change; another study found that global warming made it 600 times<\/a> more likely to occur.) They\u2019re also less accurate than older models at predicting extreme wind or record-breaking cold.<\/p>\n That\u2019s because they\u2019re trained using decades of past data. \u201cThey try to empirically understand, if I see a certain type of weather today, what is the weather tomorrow?\u201d says Engelke. \u201cEssentially, they are reproducing what has happened in the past. If we\u2019re looking at extreme weather, and especially record-breaking events, then this has not been observed in the past. It\u2019s really the lack of information in their training data that makes it almost impossible for them to forecast it.\u201d<\/p>\n<\/div>\n The study looked at models a year ago, so they\u2019ve already improved; some have added probabilistic models that predict multiple outcomes to try to become more accurate. But the fundamental problem still exists, because they\u2019re still based on training data from the past. Traditional physics-based forecasting uses complex mathematical models to represent the physical world instead, and can more readily adapt to new conditions. (Traditional models aren\u2019t perfect at predicting extreme weather, either, but still do a better job.)<\/p>\n For more typical weather forecasting, or extreme weather that isn\u2019t wildly outside the range of past events, AI can outperform traditional models. When Nvidia released its AI forecasting model Atlas earlier this year, it ran a study showing how well it performed on an extreme event it had not been trained on: Storm Dennis, a rapidly intensifying cyclone that impacted the U.K.<\/p>\n \u201cYou can see just clearly by visualizing the magnitude of the wind and the magnitude of the pressure gradient that the model was able to capture realistically intense wind events and really intense cyclones that cause damage,\u201d says Mike Pritchard, director of climate simulation research at Nvidia. The models can also accurately predict the path of hurricanes. They\u2019re already used alongside traditional models by weather agencies, weather data companies like the Weather Company, and insurance companies.<\/p>\n<\/div>\n