The AI model, developed in collaboration with university scientists and tech companies like Nvidia, Huawei, Microsoft, and Google, is part of a broader shift towards integrating AI into weather forecasting. While the AI model shows improved accuracy for certain parameters, experts like Rob Hutchinson from Meteomatics AG caution that it won't replace conventional forecasts soon. Instead, a hybrid approach combining AI and traditional methods is anticipated. The European center plans to enhance the AI model further by connecting it directly with satellite and weather station data and exploring new data sources from everyday devices. These advancements are expected to benefit energy markets by increasing forecast frequency and accuracy.
Key takeaways:
- A new AI weather model developed by the European forecasting center is improving the accuracy and speed of weather predictions, aiding energy traders in making quicker market decisions.
- The AI model uses historical data in addition to satellite and sensor information, providing more accurate forecasts for temperature, precipitation, wind, and tropical cyclones with less computing energy.
- The AI model can generate a raw forecast in three minutes, significantly faster than the 30 minutes required by conventional supercomputer models.
- Despite its advancements, the AI model is currently used alongside conventional forecasts, with future plans to integrate more data sources and improve accuracy for cloud cover, dust, and weather extremes.