Google unveils AI model for tropical storm forecasting
Google DeepMind and Google Research just launched Weather Lab, an AI-powered platform that predicts tropical storms up to 15 days ahead. By running 50 different storm scenarios and analyzing details like path, size, and strength, it aims to make forecasts faster and more accurate. The system is currently being tested with the US National Hurricane Center.
Trained on a massive European weather archive
Weather Lab's AI is trained on a massive European weather archive (ERA5), learning from both real-world observations and traditional models. In tests, its five-day cyclone track predictions landed about 140km closer to actual storm paths than leading European forecasts—a big step forward for accuracy.
Better predictions can help people prepare for big storms
Better predictions mean more time for people to prepare when big storms are coming. With faster, smarter warnings, communities at risk can plan ahead and stay safer during extreme weather.
Microsoft is also working on similar tech
Google isn't alone—Microsoft's Aurora model is also making waves by beating old-school hurricane forecasts using similar AI pattern recognition tricks. The race for better weather prediction tech is definitely heating up.