A new artificial intelligence weather forecasting model developed by Google DeepMind, called GraphCast, has demonstrated superior performance compared to the world’s leading traditional numerical weather prediction system. The AI model, described in a study published in Science, produces accurate 10-day weather predictions in under one minute on a single Google TPU v4 cloud computer. GraphCast significantly outperformed the European Centre for Medium-Range Weather Forecasts’ high-resolution forecast (HRES) system in more than 90% of over 1,300 test variables, including temperature, pressure, wind speed, and humidity at multiple atmospheric levels. The AI system uses a graph neural network architecture that learns from decades of historical weather data, representing the Earth’s atmosphere as a series of interconnected points. While traditional systems rely on complex physics equations solved on supercomputers, GraphCast’s machine learning approach enables faster, more efficient predictions that could potentially improve early warnings for extreme weather events. Researchers note that AI models like GraphCast should complement rather than replace traditional methods, as they currently depend on conventional systems for initial conditions and lack certain capabilities like ensemble forecasting. Read the full article at: https://sciencedaily.com/releases/2023/11/231114143522.htm