Last weekend was a mix of beautiful beach weather and heavy rain in the Netherlands. So accurate weather forecasting is a welcome development. AI models, trained using millions of hours of historical weather data, now deliver forecasts as accurate as those of supercomputers, according to an article in the journal Scientific scientific journal.
Why is this important:
Weather forecasts are getting better and better. While a six-day forecast is now as reliable as a three-day forecast was thirty years ago, AI is offering more accuracy and speed. Thanks to deep learning and pattern recognition, AI models are now able to accurately predict changes in the atmosphere.
Sudden changes
AI-powered weather models, like Google DeepMind’s GraphCast, can provide forecasts up to ten days in advance. They are able to generate accurate forecasts with less computing power and in a fraction of the time, using data from the past 40 years. This is especially important during sudden changes in weather, such as the path of a hurricane.
Big Tech Companies
Big tech companies like Google, Microsoft, and Huawei are actively contributing to the development of advanced AI-powered weather models. Google DeepMind and Huawei are competing for the most efficient AI-powered weather model, with GraphCast beating the European Meteorological Organization ECMWF by 90% of the validation targets. This competition is accelerating progress and ensuring better and faster weather forecasts.
In the Netherlands, AI companies are also active in weather forecasting. Beyond Weather, a spin-off from Vrije Universiteit (VU) Amsterdam, uses advanced AI to provide accurate weather forecasts for the energy and agriculture sectors across Europe.
Human factor
While AI can produce weather reports quickly and accurately, the role of the meteorologist remains indispensable. Meteorologists bring the intuition and expertise necessary to assess and report on risks. AI can produce forecasts automatically, but interpreting and explaining that data to the public remains a human task.
Black box
The adoption of AI in weather forecasting holds promise, but it also faces challenges. One of the biggest hurdles is the “black box” nature of AI models; it’s often unclear how these systems arrive at their conclusions. This can lead to hesitation among users and researchers. However, as long as forecasts are accurate, AI adoption in this sector is expected to continue to grow.
