Assimilation of land surface temperature in the global PNT model ARPEGE with a bidimentional ensemble variational method 2DEnVar F/M

The position is opened at the Centre National de Recherches Météorologiques in Toulouse, UMR Météo-France and CNRS. The candidate will be recruited in the Observations team in the Groupe de Modélisation et d'Assimilation pour la Prévision (GMAP). This group is in charge of the development of future Météo-France operational numerical weather prediction models. The Observations team develops the integration of meteorological observations into model to define the initial state of forecasts.

Data assimilation methods have made significant progress in the past decades and are now able to ingest a large variety of observations from different observing systems with the aim of improving Numerical Weather forecasts at different scales. At Météo-France as well as in other NWP centres, the use of ensemble information in data assimilation algorithms is now central. ARPEGE global model uses a hybrid 4DVar data assimilation algorithm for upper-air assimilation with the use of 3D ensemble covariances. Near the surface an ensemble variational method has been recently developed to initialise soil variables in the AROME limited-area NWP model (Marimbordes et al., 2024). With this system, the introduction of new types of observations from satellites is now made easier (Marimbordes et al., 2025). This system was recently extended to the ARPEGE global model.

The objective of this work is to study the possibility of assimilating a new satellite observation, the land surface temperature from IASI-NG and/or IRS sounder into the ARPEGE global model using an advanced surface data assimilation system based on an ensemble variational approach.

This study happens in the framework of CHERRI European project (Climate Hazards Enhanced Reanalyses & Reasoning with AI) lead by ECMWF (European Centre for Medium range Weather Forecast).

 Assimilation of land surface temperature in the global NWP model ARPEGE with a bidimensional ensemble variational method (2DEnVar) in the framework of CHERRI Horizon Europe project.

The study will be divided into four steps

1. Assessement of land surface temperature observations from IASI-NG and/or IRS quality to prepare their assimilation. 

2.Assimilation of these observations into a surface data assimilation system based on a variational ensemble approach (2DEnVar). These developments will take place in the ARPEGE global model.

3.Evaluation of results and sensitivity tests to find the configuration that improves low atmospheric forecasts.

4.Preparation of a report and a scientific publication. Presentation of results at conferences in the framework of CHERRI project.

Additional information

This position is open to people with disabilities.

Office hours.

Fixed-term contract of 24 months.

Remote work available: 1 day a week maximum.

• You hold a master's degree or a doctorate

• Experience in data assimilation is required.

• Solid knowledge in programation langages as Fortran/C++/Python is required.

• Knowledge in meteorology or numerical weather prediction is a recommended.

• The candidate must demonstrate scientific curiosity, autonomy, teamwork skills, responsiveness, analytical abilities, and rigor in interpreting and presenting results. They must be able to report regularly on their work to the project team.

• The candidate must also be proficient in English (both written and spoken).

Application deadline: 16/11/2026

Email addresses for further information: camille.birman@meteo.fr, sophie.marimbordes@meteo.fr