I am a research scientist at the National Center for Climate (NCKF) of the Danish Meteorological Institute DMI (København, Denmark) and an associate researcher at IMT Atlantique (Brest, France) and Lab-STICC, UMR CNRS 6285 (French CNRS laboratory). I am also a member of the Odyssey joint research team (with INRIA and Ifremer). My main research interest is how to build data-driven models for solving inverse problems and emulating large numerical models in geosciences.
News
2025. Fast and scalable Air Quality neural emulator for high-resolution predictions on very large domains. Maxime Beauchamp., Vincent Lécluse, Nicolas Thorr, Anas Oubida, Charles Schillinger, Florent Vasbien, Jérôme Le Paih. Submitted to JGRL.
2025. Enhancing Air Quality simulations with neural downscaling architectures. Beauchamp M., Bessagnet B., Pisoni E., Rey Pommier A., De Meij A. and Thunis P. Atmospheric Sciences Letters., https://doi.org/10.1002/asl.70003
2025. Multiscale neural assimilation scheme for learning Near-Real-Time Level 4 SST products from satellite observations. Beauchamp M.., Karagali I., Gacitúa G., Høyer J.L., Ballarotta M. and Fablet R. Nature Scientific Report, https://doi.org/10.1038/s41598-025-23682-9
2025. Neural variational Data Assimilation with Uncertainty Quantification using SPDE priors. Beauchamp M., Fablet R., Benaı̈chouche S., Tandeo P. and Desassis N. Artificial Intelligence for the Earth Systems. American Meteorological Society. https://doi.org/10.1175/AIES-D-24-0060.1
Timeline
Since 2024.Research scientist, Danisk Meteorological Institute, Copenhague
Since 2019.Postdoctoral researcher*, Institut Mines-Télécom Atlantique, MEE department and & LAB-STICC
TOMS team & Odyssey, Brest
Since 2019.Lecturer. Centre National des Arts et Métiers (CNAM), Paris
Since 2019. Self-employed as a scientific consultant.
2017-2018.Research Engineer in the Data Assimilation Group, Nansen Environmental and Remote Sensing Center (NERSC), Bergen (Norway)
2016-2018.PhD in Geosciences, Ecole des Mines de Paris, University PSL, Geosciences Center, Fontainebleau. Geostatistical contributions to improve the estimation in air quality. Development of geostatistical methods applied to air quality spatial estimation and spatio-temporal forecast. Responses to operational constraints and regulatory requirements for the exceedances of limit values.
2010-2017Research Engineer. Geostatistical and deterministic air quality modelling, INERIS (French National Institute for Industrial Risks and Environmental Safety), Chronic Risks Department, MOCA unit (Modelling
and Mapping), Verneuil-en-Halatte.
Research Topics
Data-driven data assimilation
Reconstruction (left) and SSH observation datasets (right)
Model emulation
Coarse resolution (left) and high resolution (right) of daily NO2 in France (2015)
Risk assessment
Probability of exceeding annual PM10 regulatory thresholds in France (2009)
