University of Manchester Uses NVIDIA Earth-2 to Forecast Air Pollution Across the UK

The University of Manchester has deployed NVIDIA’s Earth‑2 large language model to generate high‑resolution air‑pollution forecasts for the entire UK, dramatically cutting the computational cost compared to traditional chemistry‑based simulators. By ingesting satellite imagery and ground‑sensor data, Earth‑2 produces hourly particulate‑matter maps that can be refreshed every 15 minutes, offering a tenfold increase in temporal granularity. The approach trades a modest drop in physical‑model fidelity for orders‑of‑magnitude speed, enabling near‑real

Want the full story? Read the original article.