A new study published in Nature demonstrates that artificial intelligence models can produce accurate global weather forecasts faster and with less computational power than traditional physics-based systems. Researchers developed an AI system trained on decades of historical weather data that can predict key atmospheric variables, such as temperature and pressure, up to 10 days in advance. The AI forecasts were generated in minutes on a single high-performance computer, compared to the hours required by conventional supercomputer-run models. While the AI system shows competitive accuracy for short- to medium-range forecasts, experts note that further testing is needed for extreme weather events and long-term climate projections. The approach highlights a significant shift toward data-driven methods in meteorology. Read the full article at https://sciencedaily.com/releases/2024/07/240711215443.htm.