Description du poste :
As a public expert in weather and climate, Météo-France is here to help keep you safe every day and assist you in making the best decisions in a changing climate. With dangerous weather events becoming even more intense and frequent due to climate change, our mission to keep you safe is crucial. We mobilize our expertise and scientific and technological excellence to enable you to anticipate and adapt to challenging weather and climate events.
Find us online:https://meteofrance.com/carte-didentite-de-meteo-france
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. However, a number of challenges remain to fully benefit from all the information content of spaceborne observations. IA-based models can pave the way towards a more optimal use of these observations. In particular, the use of multi-source data, incorporating both traditional NWP analysis alongside spaceborne observations, as done in Nipen et al. (2025) at Met-Norway using crowded sources observations, could improve the prediction of IA-based NWP models. Therefore, the aim of this EUMETSAT founded project is to study the impact of the next generation of spaceborne microwave and infrared satellite observations (MTG-S/IRS and EPS-Sterna) to improve the IA-based version of the kilometre-scale NWP model AROME of Météo-France.
In this project, we propose to focus on three main questions:
1. What is the added value of using satellite observations in combination to analysis data in a ML forecast system?
2. How does the usage of satellite observation in this ML context compare with the usage by traditional DA?
3. Can a Multi-Datasets approach be trained before the availability of new observations, e.g. with simulated observations, to reduce/optimize the preparation time for using new datasets shortly after launch and avoid the need to accumulate long data records, usually needed for the training of AI/ML techniques?
GMAP is a research group at CNRM (National Centre for Meteorological Research) affiliated with Météo-France and CNRS (National Centre for Scientific Research). GMAP develops numerical forecasting software packages for Météo-France.
The OBS team is in charge of the assimilation of observations in the global and regional NWP models of Météo-France.
The PREDICTABILITY team is expert in the development of ensemble prediction systems, as well as of the new data-driven weather models.
Why join us?
Embark on a stimulating adventure that serves everyone, alongside men and women who are committed every day to tackling the challenges posed to our society by weather and climate.
And enjoy the following benefits: flexible working hours, RTT (reduced working time), teleworking, staff restaurant or meal vouchers, 75% contribution to public transport costs, contribution to mutual insurance, sports and cultural associations depending on the site concerned (climbing, gym, pottery, theater, etc.).
Other benefits await you, come and discover them!
To answer the questions mentioned above, the work will be divided into three different tasks:
1. Simulation of future observing systems (EPS-Sterna and MTG-S/IRS) to progressively cover a large period of 2020-2024. This task will be performed using version 14 of RTTOV.
2. Train a new machine learning NWP model using both simulated data and analyses. This task will be performed using the open-source ANEMOI framework, which has been developed collaboratively by ECMWF and several European national meteorological services. Recently, the ability to use multi-source observations (e.g. observations and analyses for instance) has been made available within ANEMOI.
3. Impact experiments will then be performed using the new AROME-IA models, trained using either MTG-S/IRS or EPS-Sterna simulated observations on top of analyses. This task will be performed using real observations as inputs.
Application deadline : 25/05/2026
For any further information, please contact Philippe CHAMBON (philippe.chambon@meteo.fr), Laure RAYNAUD (laure.raynaud@meteo.fr), Mary BORDERIES (mary.borderies@meteo.fr) or Nadia FOURRIE (nadia.fourrie@meteo.fr)
* Experience in remote sensing and/or data assimilation is required. Strong skills in a programming language such as Fortran or Python, as well as in handling large volumes of data, are also required.
* Experience with the ANEMOI software and with the integration of new datasets into ANEMOI is recommended.
* Knowledge of meteorology or numerical weather prediction is desirable.
* 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).
Expérience demandée :
6 à 10 ans
Description du poste :
As a public expert in weather and climate, Météo-France is here to help keep you safe every day and assist you in making the best decisions in a changing climate. With dangerous weather events becoming even more intense and frequent due to climate change, our mission to keep you safe is crucial. We mobilize our expertise and scientific and technological excellence to enable you to anticipate and adapt to challenging weather and climate events.
Find us online:https://meteofrance.com/carte-didentite-de-meteo-france
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. However, a number of challenges remain to fully benefit from all the information content of spaceborne observations. IA-based models can pave the way towards a more optimal use of these observations. In particular, the use of multi-source data, incorporating both traditional NWP analysis alongside spaceborne observations, as done in Nipen et al. (2025) at Met-Norway using crowded sources observations, could improve the prediction of IA-based NWP models. Therefore, the aim of this EUMETSAT founded project is to study the impact of the next generation of spaceborne microwave and infrared satellite observations (MTG-S/IRS and EPS-Sterna) to improve the IA-based version of the kilometre-scale NWP model AROME of Météo-France.
In this project, we propose to focus on three main questions:
1. What is the added value of using satellite observations in combination to analysis data in a ML forecast system?
2. How does the usage of satellite observation in this ML context compare with the usage by traditional DA?
3. Can a Multi-Datasets approach be trained before the availability of new observations, e.g. with simulated observations, to reduce/optimize the preparation time for using new datasets shortly after launch and avoid the need to accumulate long data records, usually needed for the training of AI/ML techniques?
GMAP is a research group at CNRM (National Centre for Meteorological Research) affiliated with Météo-France and CNRS (National Centre for Scientific Research). GMAP develops numerical forecasting software packages for Météo-France.
The OBS team is in charge of the assimilation of observations in the global and regional NWP models of Météo-France.
The PREDICTABILITY team is expert in the development of ensemble prediction systems, as well as of the new data-driven weather models.
Why join us?
Embark on a stimulating adventure that serves everyone, alongside men and women who are committed every day to tackling the challenges posed to our society by weather and climate.
And enjoy the following benefits: flexible working hours, RTT (reduced working time), teleworking, staff restaurant or meal vouchers, 75% contribution to public transport costs, contribution to mutual insurance, sports and cultural associations depending on the site concerned (climbing, gym, pottery, theater, etc.).
Other benefits await you, come and discover them!
To answer the questions mentioned above, the work will be divided into three different tasks:
1. Simulation of future observing systems (EPS-Sterna and MTG-S/IRS) to progressively cover a large period of 2020-2024. This task will be performed using version 14 of RTTOV.
2. Train a new machine learning NWP model using both simulated data and analyses. This task will be performed using the open-source ANEMOI framework, which has been developed collaboratively by ECMWF and several European national meteorological services. Recently, the ability to use multi-source observations (e.g. observations and analyses for instance) has been made available within ANEMOI.
3. Impact experiments will then be performed using the new AROME-IA models, trained using either MTG-S/IRS or EPS-Sterna simulated observations on top of analyses. This task will be performed using real observations as inputs.
Application deadline : 25/05/2026
For any further information, please contact Philippe CHAMBON (philippe.chambon@meteo.fr), Laure RAYNAUD (laure.raynaud@meteo.fr), Mary BORDERIES (mary.borderies@meteo.fr) or Nadia FOURRIE (nadia.fourrie@meteo.fr)
* Experience in remote sensing and/or data assimilation is required. Strong skills in a programming language such as Fortran or Python, as well as in handling large volumes of data, are also required.
* Experience with the ANEMOI software and with the integration of new datasets into ANEMOI is recommended.
* Knowledge of meteorology or numerical weather prediction is desirable.
* 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).
Expérience demandée :
6 à 10 ans