● Covered by 1 source · 1 reportMedium impact1 neutral
University of Manchester uses NVIDIA Earth-2 models for UK air pollution forecasting
Key points
- University of Manchester adapted NVIDIA Earth-2 AI models for air pollution forecasting.
- Traditional chemistry-based models are computationally expensive and slow.
- Earth-2 CorrDiff and StormCast models were trained on UK's Isambard-AI supercomputer.
- The models provide detailed, time-dependent air quality forecasts for the UK.
AI for Air Quality Forecasting
Researchers at the University of Manchester, led by Professor David Topping, have successfully applied NVIDIA Earth-2 open AI models and tools to forecast air pollution across the United Kingdom. This project aims to overcome the limitations of traditional chemistry-based air quality models, which are computationally intensive and restrict the detail and frequency of forecasts.
Adapting Generative Frameworks
Topping identified that the generative frameworks used in NVIDIA Earth-2 for weather forecasting could be adapted for pollution fields. The team generated training data from existing chemistry-climate simulations and then trained Earth-2 CorrDiff, a generative downscaling model, on Isambard-AI, the UK’s national AI supercomputer. The initial model proved effective on its first attempt.
Enhanced Forecasting Capabilities
The project further integrated Earth-2 StormCast, a model that enables time-dependent forecasts by directly incorporating air quality observations. The test-training and inference workflows were demonstrated on an NVIDIA DGX Spark personal AI supercomputer. This development allows for more dynamic and responsive air quality predictions.
Potential Impact on Public Health and Policy
The UK-wide pollution model offers the ability to simulate future scenarios, such as the impact of different government policies related to pollution. This could provide proactive insights for healthcare organizations, allowing them to alert patients with conditions like asthma about impending high air pollution levels. The team is also exploring integration with edge AI devices for real-time air quality data and decision-making, particularly in events like wildfires.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
The University of Manchester, in collaboration with NVIDIA, has adapted NVIDIA Earth-2 AI models to forecast air pollution across the UK. This initiative addresses the computational expense of traditional chemistry-based models, enabling more detailed and frequent air quality predictions. The new AI-driven approach could inform public health interventions and policy decisions related to air quality.