A new study published in Science demonstrates that artificial intelligence weather models can now match or exceed the accuracy of traditional physics-based forecasting systems. The research compared Google’s GraphCast AI model against the European Centre for Medium-Range Weather Forecasts (ECMWF) system, the current global gold standard. Results showed GraphCast provided more accurate predictions for over 90% of 1,380 test variables, including critical atmospheric measurements, while being thousands of times faster and more energy-efficient. The AI model excels at predicting extreme weather events like tropical cyclones and atmospheric rivers days in advance. Researchers emphasize that AI systems should complement rather than replace conventional models, as each approach has distinct strengths. This advancement marks a significant shift toward hybrid forecasting systems that could improve early warnings for dangerous weather worldwide. Read the full article at https://technologyreview.com/2024/01/15/1086821/ai-weather-forecasting-graphcast-outperforms-traditional-models/