University of Manchester Deploys NVIDIA Earth-2 AI for UK Air Pollution Forecasting
Researchers retrained NVIDIA's weather AI models to generate high-resolution UK pollution forecasts.
AI-generated editorial illustration.
The University of Manchester trained NVIDIA Earth-2 CorrDiff, a generative downscaling model, on air pollution data using Isambard-AI, the UK's national AI supercomputer. According to NVIDIA, training completed in two days on a single eight-GPU node, producing a UK-wide pollution model at 2-3 square kilometer resolution. The team has since added Earth-2 StormCast for time-dependent forecasting. Inference now runs on an NVIDIA DGX Spark desktop system. Researchers plan to release open-source training data and workflows so other countries can build similar models. Potential applications cited by the team include proactive health alerts for patients with respiratory conditions and real-time wildfire response.
Source: NVIDIA News. Original announcement: September 16, 2026.