The EGU interactive community platform
The EGU interactive community platform
EGUsphere
Preprints
Preprints
https://doi.org/10.5194/egusphere-2026-5855
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/egusphere-2026-5855
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
Preprints
30 Oct 2026
 | 30 Oct 2026
Status: this preprint is open for discussion and under review for Geoscientific Model Development (GMD).

EMEPai 1.0.0: A fast regional data-driven air-quality forecast model for Europe

Erlend Lima, Halvard Sutterud, Lewis Blake, Magnus Ulimoen, and Erik Askov Mousing

Abstract. Air-quality forecasting over Europe relies largely on chemical transport models, whose computational cost limits forecast resolution and frequency. This study presents EMEPai, a regional data-driven forecasting system designed to emulate atmospheric composition fields produced by the EMEP/MSC-W modeling and data-assimilation framework. The model uses a graph-based encoder–processor–decoder architecture to generate three-hourly forecasts for NO2, O3, PM10, PM2.5, and SO2 over Europe at 0.1° resolution, with lead times up to 96 hours.

Against surface observations, its performance exceeds that of the corresponding EMEP hindcast while approaching the skill of the assimilated reanalysis. Reduced skill is found for SO2 and during high-concentration events. In an operational forecast setting, EMEPai performs comparably to or better than the CAMS regional ensemble across the evaluated pollutants and forecast lead times. A 96-hour forecast requires only seconds on a single GPU, with the full forecast workflow approximately 230–460 times faster than the corresponding EMEP/MSC-W worfklow.

How to cite. Lima, E., Sutterud, H., Blake, L., Ulimoen, M., and Mousing, E. A.: EMEPai 1.0.0: A fast regional data-driven air-quality forecast model for Europe, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-5855, 2026.
Received: 25 Sep 2026 – Discussion started: 30 Oct 2026
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Share
Erlend Lima, Halvard Sutterud, Lewis Blake, Magnus Ulimoen, and Erik Askov Mousing

Status: open (until 25 Dec 2026)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
Post a comment Subscribe to comment alert
Erlend Lima, Halvard Sutterud, Lewis Blake, Magnus Ulimoen, and Erik Askov Mousing

Data sets

EMEPai v1.0.0: Archive of source code for training and inference, and checkpoint and input data for inference Halvard Sutterud et al. https://doi.org/10.5281/zenodo.22955196

Erlend Lima, Halvard Sutterud, Lewis Blake, Magnus Ulimoen, and Erik Askov Mousing
Metrics will be available soon.
Latest update: 31 Oct 2026
Download
Short summary
Air-quality forecasts are important but often expensive to produce. In this study, we present a data-driven air quality forecast model, EMEPai, trained on model output from an existing physics-based system that combines modeling and data assimilation. The model predicts major pollutants across Europe up to four days ahead while preserving most of the skill of the more expensive traditional approach. It shows that accurate and reliable forecasts can be produced at a much lower computational cost.
Read more
Share