A new study published in Nature demonstrates that artificial intelligence models can produce medium-range weather forecasts with accuracy comparable to traditional physics-based systems, but significantly faster and at lower computational cost. The research team developed a deep learning model trained on decades of historical global weather data. In head-to-head comparisons, the AI model matched the European Centre for Medium-Range Weather Forecasts’ high-resolution system in predicting key atmospheric variables like geopotential height and specific humidity for forecasts up to seven days ahead. The AI system completed its forecasts in under a minute on a single graphics processing unit, a task that takes conventional supercomputers hours. Researchers caution that while promising for certain applications, AI models must still be rigorously tested for extreme events and integrated with physical knowledge for full operational reliability. Read the full article at: https://sciencedaily.com/releases/2024/05/240521123456.htm