A new artificial intelligence weather forecasting model developed by Google DeepMind, named GraphCast, has demonstrated superior accuracy and speed compared to the world’s leading conventional system. In a head-to-head evaluation, GraphCast outperformed the European Centre for Medium-Range Weather Forecasts’ (ECMWF) high-resolution forecast (HRES) on over 90% of 1,380 test variables, including critical metrics like temperature, pressure, wind speed, and humidity. The AI model generates a 10-day forecast in under one minute on a single Google TPU v4 machine, a process that takes the traditional system hours to compute on a supercomputer. While not intended to replace established systems, GraphCast represents a significant leap in computational efficiency and could serve as a powerful complementary tool for meteorologists. The model was trained on nearly four decades of historical weather data from ECMWF’s reanalysis dataset. Researchers emphasize that AI and physics-based models will likely work in tandem to improve the reliability of weather predictions, which are crucial for agriculture, disaster preparedness, and transportation. Read the full article at: https://technologyreview.com/2023/11/14/1083729/ai-weather-forecast-graphcast-beats-traditional-methods/