En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre FR636-EVEMAG-044 - Post-doctorant-(H/F), utilisation de méthodes d'apprentissage automatique pour mieux évaluer les flux de surface dans les modèles climatiques à partir d'observations long-terme.

Post-doctorant-:(H/F) Using machine learning methods to better evaluate surface fluxes in climate models against local long-term observations (H/F)

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : mercredi 19 mai 2021

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : FR636-EVEMAG-044
Workplace : GUYANCOURT
Date of publication : Wednesday, April 28, 2021
Type of Contract : FTC Scientist
Contract Period : 15 months
Expected date of employment : 1 September 2021
Proportion of work : Full time
Remuneration : From 2728 € gross to 3881 € gross according to expérience-
Desired level of education : PhD
Experience required : 1 to 4 years


Objectives -The Global Energy and Water cycle Exchanges (GEWEX) and World Climate Research Program (WCRP) have pointed out for the last ten years the importance of the Land-Atmosphere coupling for weather and climate models. The Working Group on Numerical Experimentation (WGNE) survey on systematic errors (Feb. 2019) established that the outstanding errors in the modelling of surface fluxes of momentum and sensible and latent heat is the second most important issue. The French ANR project MOSAI (Models and Observations for Surface Atmosphere Interactions) aims at reducing the biases of surface fluxes in numerical weather prediction and climate models by focusing on three main scientific questions:
• Are the simulations of the land-atmosphere exchanges fairly evaluated with local observations?
• Can we propose new methodologies for the observation-model comparison?
• Could the land-atmosphere coupling be improved in models?

The position proposed here addresses the second question. Indeed, existing model evaluations by comparison of absolute flux values are of limited help because the disparities are a blending of inconsistent spatial scales, inaccurate atmospheric and soil conditions, and imperfect parameterizations. Several studies attempted to use dependency between variables to better distinguish the reasons of departure but they are usually limited to one site, one model, and/or they use a one-by-one dependency criteria with only very few variables. The objective of this position is to develop a method that is relevant to identify weaknesses of the different models involved in the MOSAI project depending on atmospheric conditions, land cover and time scales, and to quantify the sensitivity of the surface fluxes to radiation, soil conditions and other environmental parameters.

Methodology - The approach will be based on a multi-variable statistical learning model, linking key environmental variables (e.g., radiation, soil moisture, wind, air temperature, humidity) to surface fluxes. It will be built from long-term observations, and applied to the 3D model outputs for evaluation.


The recruited person will have the following responsibilities, in collaboration with the scientists involved in the MOSAI project and/or in our team in LATMOS:
• Prepare datasets for the different supersites considered in the project
• Set up relevant statistical methods for the different supersites based on available observations
• Evaluate the different simulations available with these methods
• Present the results in a publication and in MOSAI's workshops and if possible in an international conference.


Required Skills :
- Hold a PhD in a field related to atmospheric sciences with specific knowledge/interest in the following areas:
o Machine learning methods for data analysis
o Land surface-atmosphere feedbacks and surface fluxes
o Numerical Weather Prediction and/or Climate Models (global or regional)
- Good practice of oral and written English (french appreciated but not required)
- Linux computing environment, shell, python or R, netcdf format

Work Context

The MOSAI project involves 7 french laboratories (CESBIO, CNRM, GET, ISPA, IGE, LAero, LATMOS, LMD) with about 40 persons. The project will last 4 years (2021-2025). It takes advantage of the long-term observations acquired at the french facilities ACTRIS-FR and ICOS sites. Links with LIAISE project and Paris-2024 also exist.
Situation de l'emploi et conditions :
The person in post will be based at LATMOS, in Guyancourt. Missions at Laboratoire d'Aérologie in Toulouse/Lannemezan will be considered for collaboration. Note that due to covid situation, part time home working may be required until the situation improves.

We talk about it on Twitter!