Missions

The general mission is to improve the integration of ice sheets into Earth System Models through the use of neural network emulators at the interface between an Antarctic ice sheet model (Elmer/Ice) and a global ocean model (NEMO). The selected candidate will contribute to the ANR-AIAI project (https://anr-aiai.github.io).

Activités

• Develop and test neural networks producing ice-shelf basal melt rates.
• Test and use the developments in global 1° ocean simulations.
• Present results at international conferences.
• Participate in the AIAI project and IGE activities (meetings, seminars, etc.).
• Regularly monitor publications on the subject and write scientific articles.

Compétences

The selection will be based on the following scientific and technical criteria:
• Proven experience in deep learning methods.
• Proven coding experience in Python.
• Proven experience in writing scientific articles.
• General knowledge in physical oceanography or climate dynamics.
The selection committee will take into account the gender balance of the research team.

Contexte de travail

The Institute for Environmental Geosciences (IGE) is a public research institute under the affiliation of CNRS, IRD, University Grenoble Alpes, Grenoble-INP, and INRAE It brings together about 330 people, including 190 permanent members (researchers, teacher-researchers, engineers) and about 140 contractual agents (doctoral students, postdocs, engineers and technicians). The institute also welcomes several dozen trainees and scientific visitors every year. It is located on three sites of the Grenoble University Campus that are 5 minutes away from each other. IGE is one of the main institutes within the Observatoire des Sciences de l'Univers de Grenoble (OSUG) which is a federative structure of INSU.
The selected candidate will join the CryoDyn team which has a focus on ice dynamics and connections to the climate system. The person will be supervised by Nicolas Jourdain (IGE), Romain Millan (IGE), and Clara Burgard (IPSL-LOCEAN, Paris). There will be strong collaborations with IPSL-LSCE (Cécile Agosta) where a postdoc will be hired on a similar approach at the atmosphere–ice-sheet interface.

Scientific background :
The melting of Antarctic ice shelves by the ocean is a major source of uncertainty on the evolution of the global climate and sea level. Adequate representation of ice-ocean interactions beneath ice shelves is challenging at the coarse resolution of climate models. Simple parameterisations exist to represent melting under non- or partly- resolved ice shelves, but they generally fail to capture the complexity of its response to evolving ocean conditions (Burgard et al., 2022). The application of deep learning to this problem has led to promising results (Rosier et al., 2023; Burgard et al. 2023). Further development and refinement of these techniques is therefore expected to significantly improve the representation of basal melt in simulations.

Contraintes et risques

Nothing to report