A new artificial intelligence weather forecasting model developed by researchers at the University of California, Berkeley, has demonstrated superior accuracy in predicting global weather patterns compared to established numerical models. The system, named ClimaNet, uses a deep learning architecture trained on decades of historical atmospheric data to generate high-resolution forecasts up to ten days in advance. In benchmark tests against the European Centre for Medium-Range Weather Forecasts (ECMWF) system, ClimaNet showed a 15% improvement in predicting key variables like precipitation and temperature anomalies, particularly in complex mid-latitude storm systems. The researchers highlight the model’s computational efficiency, as it can produce a forecast in minutes on a standard server cluster, a task that typically requires hours on a supercomputer using traditional physics-based methods. While promising, the team notes that AI models like ClimaNet must be continuously validated against real-world events and integrated with conventional approaches for the most reliable predictions. The full study is published in the journal *Nature Geoscience*. Read the full article at: https://sciencedaily.com/releases/2024/05/240521123456.htm