Appel à candidatures | Recherche, Emploi

Postdoc: Multifractal stochastic modeling and downscaling in micro-meteorology

Du 1 février 2025 au 1 juillet 2026

18 months, from February 2025
Rennes,
Contacts : valentin.resseguier@inrae.fr

The postdoctoral fellow will contribute to the development of a stochastic, data-driven model capable of generating high-resolution 3D+1 wind velocity fields. The research will focus on coupling coarse-grid mechanistic simulations with high-frequency local data assimilation using ensemble Kalman filtering techniques. A key challenge will be to reproduce the intermittent, multifractal nature of turbulent flows, crucial for accurate dispersion models in agroecology.

 Context
The DEDECHAO (DEscente D’Échelle par CHAOs multiplicatif Gaussien) project explores a novel approach to downscaling in micro-meteorology, framed as a Bayesian inverse problem. The goal is to estimate high-resolution (10m) wind speeds based on low-resolution (10km) data, probabilistic simulations, and local measurement assimilation. This method will leverage the Gaussian Multiplicative Chaos (GMC) [1] as a generative model for wind fields, incorporating intermittency in turbulence [1,2] to capture extreme variations in wind speed at small scales. The project has significant implications for agroecological practices, such as improving the dispersion modeling of volatile organic compounds (VOCs) and spores in pest control and epidemic surveillance.

Assignment
The postdoctoral fellow will contribute to the development of a stochastic, data-driven model capable of generating high-resolution 3D+1 wind velocity fields. The research will focus on coupling coarse-grid mechanistic simulations with high-frequency local data assimilation using ensemble Kalman filtering techniques. A key challenge will be to reproduce the intermittent, multifractal nature of turbulent flows, crucial for accurate dispersion models in agroecology.

Main activities
• Adapt a wavelet-based GMC code for generating intermittent 3D+1 wind velocity fields [3].
• Couple GMC with mechanistic simulations through scale symmetries and stochastic transport
• Integrate large-scale meteorological data with in-situ measurements through data assimilation techniques.
• Validate the model's predictions against high-resolution simulation outputs (from PALM software).
• Prepare research papers and present results at scientific conferences.

Skills
• PhD in applied mathematics, fluid dynamics, atmospheric sciences, or related fields.
• Strong expertise in stochastic modeling, turbulence, and numerical simulations.
• Experience with data assimilation methods (e.g., ensemble Kalman filtering) and Bayesian inference.
• Familiarity with large-scale atmospheric simulations and downscaling techniques.
• Proficiency in Python
• Experience with high-performance computing is a plus.
• Ability to work independently and as part of a multidisciplinary research team.
References
[1] B. Castaing, Y. Gagne and E.J. Hopfinger, Velocity probability density functions of high Reynolds number turbulence, Physica D, 46 (1990) 177-200.
[2] L. Chevillard, B. Castaing, A. Arneodo, E. Leveque, J.-F. Pinton and S.G. Roux, A phenomenological theory of Eulerian and Lagrangian velocity fluctuations in turbulent flows, C.R. Physique, 13 (2012), 899-928.
[3] C. Granero-Belinchon, S.G. Roux and N.B. Garnier, Multiscale and anisotropic characterization of images based on complexity: an application to turbulence. Physica D, 459 (2024), 134027.