A new artificial intelligence weather forecasting model has demonstrated superior accuracy compared to conventional physics-based models in a comprehensive global evaluation. The AI system, developed by researchers at a leading tech institute, analyzes vast amounts of historical atmospheric data to predict future conditions. In head-to-head tests over a six-month period, the AI model provided more accurate predictions for key variables like temperature and pressure at multiple altitudes up to seven days in advance. The study’s authors note that while the AI approach is computationally efficient once trained, it currently lacks the explicit physical interpretability of traditional models. This advancement signals a significant shift in meteorological science, where data-driven systems are beginning to rival and sometimes exceed established methods. Read the full article at: https://sciencedaily.com/releases/2024/05/240507150543.htm