PhD Position F/M Data assimilation of satellite data in oceanic models, learning of oceanic dynamics

Social network you want to login/join with:

PhD Position F/M Data assimilation of satellite data in oceanic models, learning of oceanic dynamics, Ille-et-Vilaine

Client: INRIA

Location: Ille-et-Vilaine

Job Category: Other

EU work permit required: Yes

Job Reference: a0f1c54081a7

Job Views: 4

Posted: 02.03.2025

Expiry Date: 16.04.2025

Job Description:

The Odyssey team is offering a PhD position on numerical ocean dynamics simulation, machine learning, and data assimilation.

Odyssey (for Ocean DYnamicS obSErvation analYsis) is a recently created team involving researchers from Inria (Rennes, France), Ifremer (Brest), and IMT Atlantique (Brest).

Inria is one of the leading research institutes in Computer Sciences in France, and Odyssey is also affiliated with the mathematics research institute of the Rennes University (IRMAR).

The team expertise encompasses mathematical (stochastic) and numerical modelling of ocean flows, observational and physical oceanography, data assimilation, and machine learning.

Gathering this large panel of skills, the team aims at improving our understanding, reconstruction, and forecasting of ocean dynamics, specifically to bridge model-driven and observation-driven paradigms to develop and learn novel representations of the coupled ocean-atmosphere dynamics ocean models.

For accurate climatic predictions, plausible forecasts of the future ocean state are essential. Ideally, high-resolution ocean simulations would be used for this purpose. However, due to their associated computational costs, this approach is currently infeasible, and we must rely only on large-scale ocean representations.

To address this challenge and the urgent need to generate various likely scenarios, there has been a growing interest in geophysical sciences and climate studies in developing flow models that incorporate noise to account for modelling uncertainties or errors.

The introduction of noise into ocean dynamics models must be done on a theoretically rigorous ground. Ad-hoc choices for model noise can fundamentally disrupt the corresponding fluid dynamics models, leading to unrealistic properties. Rigorously justified methodologies for deriving stochastic dynamics models have been recently introduced in the Odyssey team within the ERC STUOD and a longstanding collaboration with Imperial College and Ifremer.

The theoretical framework on which we rely, referred to as "modelling under location uncertainty", decomposes the flow in terms of a resolved smooth component and a rapidly oscillating random component. The stochastic dynamics is then defined from a stochastic representation of the Reynolds transport theorem. From this modelling principle, stochastic equivalents of the classical geophysical flow models can be defined.

A set of models ranging from multi-layers quasi-geostrophic models to primitive equations have been defined and numerically implemented. Ensemble data assimilation is currently under development, as well as simplified ocean-atmosphere coupled models.

The present PhD position aims to explore data-driven dynamics specification and learning from high-resolution data, as well as devising hierarchical data assimilation ensemble strategies to couple stochastic ocean models and high-resolution satellite data such as the SWOT data.

The PhD student will collaborate directly with the Odyssey group in Rennes (Noé Lahaye, E. Mémin, Gilles Tissot) and in Brest (B. Chapron and R. Fablet). He will be supervised by Etienne Mémin and co-supervised by Bertrand Chapron and Ronan Fablet to cover different aspects, including stochastic modelling of ocean dynamics, satellite observation, and machine learning for ocean dynamics and data assimilation.

He/She will be part of a small group devoted to ensemble methods for forecast, learning, and data assimilation of ocean dynamics. His/Her work will also undergo strong collaborations with the Odyssey group at IMT Atlantique (R. Fablet) as well as with the other PI of the ERC Stuod group (Bertrand Chapron, Dan Crisan, Darryl Holm). During this PhD position, we will explore in particular ensemble methods and kernel representation and data assimilation.

The candidate should have a solid background in applied mathematics and/or in fluid mechanics and/or in geophysical dynamics.

She/he must have a good knowledge of Fortran, C/C++, Python, and Pytorch.

Avantages

• Partial reimbursement of public transport costs
• Possibility of teleworking (90 days per year) and flexible organization of working hours
• Partial payment of insurance costs

Monthly gross salary amounting to 2100 euros for the first and second years and 2200 euros for the third year.