postdoctoral researcher in climate modelling (M/F)

Job Description

Missions

Calibrate climate models and quantify uncertainties

Activity

The project aims to support the implementation of major evolutions in the atmospheric convection scheme of the IPSL coupled model in order to:
1. tune the new parameterizations to achieve a better representation of rainfall,
2. understand the link between changes in parameterizations and climate changes, both in simulations forced by sea surface temperatures and in coupled simulations, and
3. quantify uncertainties in future projections.
The study will focus on two specific aspects of the simulated climate, due to their climatic importance and the known biases or uncertainties in climate models in general—or in the IPSL model in particular:
· Tropical convective storms over the ocean: The project will examine the extent to which changes in atmospheric convection parameterizations control sea surface temperatures and help explain endemic model biases, such as the "double ITCZ" or more specific structural biases in the IPSL model.
· The African monsoon: Here, the focus will first be on the influence of convection, as well as surface processes and the varying presence of desert dust, on the northward extension of monsoon rainfall over the Sahel. The project will then assess how these changes can improve the representation of recent variations in Sahel rainfall and its evolution in response to global warming—one of the aspects of climate change where models diverge the most.
In the first phase, a Perturbed Physics Ensemble (PPE)—a set of simulations differing in their choice of certain free parameters—will be used to initiate analyses and explore the links between hydrological aspects and parameters. These analyses will provide an initial insight into the scientific questions and inform the design of a second phase using History Matching to generate a PPE with metrics targeted at these specific issues. Tuning will be performed using the HMIR method (Hourdin et al. 2021) with the LMDZ atmospheric model coupled to the ORCHIDEE land surface model (focusing on convection and surface parameters), particularly targeting improvements in tropical rainfall. Finally, a PPE will be extracted to conduct joint simulations with both the SST-forced configuration and the coupled model, potentially adding ocean model parameters related to evaporation.
Ultimately, a subset of this PPE will be selected to study uncertainties in climate projections, both in terms of global warming (climate sensitivity, Hourdin et al. 2023) and the evolution of tropical rainfall, with a particular focus on the Sahelian monsoon.

Your Profil

Skills

The candidate must hold a PhD in climate science, geosciences, atmospheric physics, or a related field, with expertise in statistical approaches applied to the calibration of numerical models and uncertainty quantification.
The candidate must demonstrate:
· Experience in parameter calibration (tuning) or uncertainty quantification in environmental or climate models;
· Proficiency in advanced statistical methods (inference, Bayesian methods, or approaches such as history matching, emulation, or related techniques);
· The ability to exploit and analyze ensembles of numerical simulations (PPE or equivalent);
· Prior experience with land surface, atmospheric, oceanic, or coupled models will be an asset;
· A strong interest in convection processes, hydrological processes, tropical precipitation, and/or climate variability (particularly in the Sahel) is expected.
The candidate must also:
· Have scientific programming skills (Python, Fortran, or equivalent) and data processing experience (CDO, NCO, etc.);
· Demonstrate scientific autonomy, analytical skills, and initiative;
· Be able to work in a collaborative environment;
· Have a level of scientific English sufficient for publishing in peer-reviewed international journals and presenting at international conferences.

The selection committee will take into account the gender balance of the research team.

Your Work Environment

The IPSL develops a global climate model used to understand climate and anticipate its future changes. The ability of such models to accurately represent past and future climate evolution depends both on their physical content—which improves with ongoing research—and on the calibration of free parameters, known as climate model tuning. The recent advent of machine learning techniques, particularly the HMIR approach (History Matching with Iterative Refocusing, Hourdin et al. 2021), has transformed the adjustment of parameters and the quantification of associated uncertainties into a new field of scientific research. The capacity to perform ensembles of simulations under observational constraints now allows these questions to be addressed much more efficiently and objectively than in the past. These issues are at the heart of project QUINTET of PEPR TRACCS.
While climate models can reproduce the warming caused by the increase in greenhouse gases, as well as the distribution of this heating and some of its major consequences, significant uncertainties remain regarding certain key aspects, such as the future evolution of the water cycle in the tropics—particularly changes in precipitation over the Sahel (Roehrig et al. 2013, Wang et al. 2020). Understanding the extent to which these future changes are linked to model design choices (structural uncertainty) or parameter choices (parametric uncertainty) is a major challenge in climate modeling.
The research project is based within the team developing the physical component of the IPSL coupled model, spanning both the LMD (scientific lead: Frédéric Hourdin) and LOCEAN (co-scientific lead: Juliette Mignot). Both leads are also specialists in Sahelian climate.
Furthermore, the project is part of the preparation phase for the IPSL CM7 model for the upcoming CMIP7 intercomparison exercise. The model is expected to include numerous improvements in the representation of atmospheric physics compared to the version finalized in spring 2026 for the FastTrack exercise. In particular, it aims to refine the parameterizations of deep convection (or thunderstorm activity) and improve the representation of the transition between shallow convection (such as trade wind cumulus or fair-weather convection over land) and deep convection. Changes to convection parameterizations are expected to significantly impact rainfall over both oceans and monsoon regions. These new parameterizations will require extensive calibration. Combined with the implementation of new objective tuning methods and uncertainty quantification, these developments should enable a major leap forward in the representation of tropical rainfall—over oceans, monsoon regions, and the Sahel—and provide a better understanding of the link between rainfall representation and past/future changes in Sahelian rainfall.

Compensation and benefits

Compensation

From 3132,83

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions