The European Centre for Medium-Range Weather Forecasts (ECMWF) has teamed up with Huawei to develop an AI-based forecasting system called Pangu-Weather. The model, described as the ‘quiet revolution’ of weather forecasting, has shown higher precision and faster prediction speeds than traditional numerical prediction methods. Pangu-Weather can predict global weather from one hour to seven days in advance, covering aspects like humidity, wind speed, temperature, and more.

 Indoors, Face, Head, Person, Furniture, Table, Desk, Adult, Male, Man
Source: youtube.com/@wendoverproductions

When used to forecast the movements of typhoons, cold waves, heatwaves, and other extreme weather events, the model has proven to be incredibly accurate. In order to help humanity adapt to extreme weather occurrences more effectively, the ECMWF and Huawei collaboration seeks to improve weather prediction accuracy and speed.

 Chart, Plot
ECMWF website showing weather forecasts from Pangu-Weather

Since AI models can discover novel atmospheric evolution patterns from vast amounts of data that humans cannot, the application of AI in weather forecasting is viewed as a game-changer for the sector. Due to various starting data inputs and the disregard for physical rules, the AI model may still struggle to predict the precise locations where typhoons will land. Typhoon strength forecasts are still based on numerical models.

cross-circle