A new artificial intelligence weather forecasting model developed by researchers has demonstrated superior accuracy compared to conventional physics-based systems in global medium-range weather predictions. The AI model, trained on decades of historical weather data, can generate a 10-day forecast in under a minute on a single high-performance computer, a task that typically requires hours on supercomputers using traditional methods. In benchmark tests against the European Centre for Medium-Range Weather Forecasts (ECMWF) system—considered the global gold standard—the AI model showed equal or better accuracy in over 90% of over 1,300 atmospheric variables measured, including critical metrics like atmospheric pressure and temperature. While the AI system excels at identifying large-scale weather patterns, researchers note that traditional models still hold an edge in predicting extreme, localized weather events. The development signals a potential shift in meteorological science, where AI could handle routine forecasts, freeing supercomputers to run more detailed simulations for high-impact events. Read the full article at: https://sciencedaily.com/releases/2024/07/240715140735.htm