In the framework of a newly-funded project by the Swiss National Science Foundation, we are seeking to fill a 3-year Postdoctoral Researcher position in ice sheet modelling at the University of Lausanne (UNIL), within the ICE group (https://wp.unil.ch/ice/) in the Institute of Earth Surface Dynamics (IDYST). The project is a collaboration with the University of Zurich (UZH), led by Andreas Vieli.
Project Summary: The future of the Antarctic Ice Sheet and its contribution to sea level rise is a pressing concern. Despite recent advances in ice sheet modelling, projections still vary widely due to significant uncertainties. Critical challenges include model calibration and running large ensembles of high-resolution simulations spanning millennia. Recent breakthroughs in deep learning and parallel computing are transforming the field, enabling models to run faster by orders of magnitude and improving data assimilation. While these techniques have shown considerable promise in modeling mountain glaciers, they have yet to be applied to marine ice sheets. This project aims to address that gap, providing an opportunity to improve uncertainty quantification and model fidelity. Our approach will develop the Instructed Glacier Model (IGM, https://github.com/jouvetg/igm), equipping it to simulate marine ice sheets and ice shelves (calving, ocean melt, grounding line, …). By leveraging GPUs, we aim to significantly accelerate model performance compared to traditional CPU-based methods. In the second phase, we will exploit these computational gains to achieve an optimized initialization of the Antarctic Ice Sheet, ensuring alignment with observational data, followed by high-resolution ensemble simulations to reduce uncertainty in future projections.
Role of the Candidate: The Postdoctoral Researcher (based at UNIL, with close collaboration at UZH) will be responsible for co-developing and setting up the model IGM for the Antarctic Ice Sheet. Candidates should have strong expertise in numerical glacier or ice sheet modelling, preferably with experience in model development and Python programming. Experience in machine learning is a plus. Excellent writing and communication skills in English, as well as the ability to collaborate effectively in an interdisciplinary and cross-institutional research environment, are essential.
The working environment: The institute (IDYST) offers excellent working conditions, including access to state-of-the-art research facilities, a collaborative and vibrant academic community, and a strong commitment to fostering an inclusive environment that embraces equity, diversity, and inclusion in all aspects of its operations. In addition, postdoctoral positions come with attractive salary packages.
Application Details: To apply, please send a cover letter, CV (including publication list), and contact information for references as a single PDF file to Guillaume Jouvet (guillaume.jouvet@unil.ch). The first review of applications will begin on December 2, 2024, and will continue until the position is filled. The start date is flexible.
If you are interested doing an intern, a bachelor, or a Master project, check at following list of topics. The rerequisites to carry out these projects are : a keen interest at understanding main glacier processes, a few basics in python programming (or the willness to learn), a keen interest at reproducing glacier advance/retreat using numerical modeling, and to adjust the simulation to data, being able to seek for open data (e.g., topographical, climatic), and transform these data in proper format.
Our team from the IDYST ICE group has recently produced an unprecedented high-resolution model simulation of the Last Glacial Maximum of the entire European Alps (~24 thousand years ago) that is, for the first time, consistent with field-based data on former ice extent and thickness. This modelling work is currently in review for publication here. This Alps-wide model simulation opens a lot of new possibilities for further scientific projects that could analyze the model outputs (in a post-processing manner) to explore questions related to former ice dynamics and its interaction with surface geology and geomorphology. For instance, one interesting, currently open scientific question is: During major Quaternary glaciations (such as the LGM); when, where, and for how long did ice flow across some major mountain passes of the Alps, therefore not respecting hydrological basins and the present-day topography but instead following different “glaciological drainage networks”? These events, more commonly referred to as “transfluences” are interesting to better constrain as they are relatively short-lived and enable rocks of certain lithologies to end up being deposited in places that we would not expect. Over multiple glacial cycles, they also cause erosional patterns that lead to unexpected modifications of the pre-glacial Alpine landscape and catchment network. Overall, many open questions related to European glacial landscape evolution can be explored with this new simulation, and this would highly suit MSc research projects, who would work with new and exciting model data that has only been available for a few months.
Modelling the age of ice within moutain glaciers (Master level) Besides glacier evolution modeling, modeling the age of ice particles can be highly valuable for various glaciological applications (e.g., https://tc.copernicus.org/articles/14/4233/2020/). The goal of this thesis, which is strongly oriented towards numerical modeling, is to implement the ‘age’ equation within the Instructed Glacier Model to complement the current scheme that computes the age of ice in a Lagrangian manner (particle-wise). Among the possible applications, one can model the age of ice in the Aletsch Glacier (see the dedicated archived topic below).
Modelling debris-covered glaciers (Bachelor/Master level) As they melt, glaciers in the Swiss Alps become increasingly covered by debris. The goal of this thesis is to explore the capability of the Instructed Glacier Model to track individual particles transported by ice motion for modeling the formation of debris cover and its influence on the surface mass balance. Applications of this thesis could include debris-covered glaciers surrounding the Aletsch area.
Modelling the future of the Qaanaaq ice cap (Bachelor/Master level)
The Qaanaaq Ice Cap is a large coastal ice mass located in northwest Greenland. The goal of this thesis is to model the future evolution of the Qaanaaq Ice Cap (with data assimilation if this is a master’s project) and to assess its response to climate change. For that purpose, we will use the Instructed Glacier Model.
In the 1920s, artillery shells were left on the uppermost part of the Rhône Glacier with the aim of tracking its ice displacement. The goal of this project is to track the motion and current position of these artillery shells (“obus”) using historical data and – the Instructed Glacier Model, and to estimate their reappearance time, considering uncertainties in ice flow dynamics and climate parameters.
Snow avalanches play an important role in redistributing solid precipitation, contributing to the positive surface mass balance of glaciers. Our glacier evolution model, IGM, includes an avalanche module to account for this; however, it has not yet been properly tested or assessed against data. The goal of this thesis is to analyze the influence of adding this component to a glacier evolution model and determine the conditions under which it has the most impact when modeling contemporary large glaciers of the Swiss Alps.
During the Last Glacial Maximum (LGM, about 21,000 years BP) most of the Alps and wide parts of the alpine forelands were covered with ice (Bini et al., 2009). Today, only geomorphological evidence left on the landscape (such as erratic boulders and moraines) witness how the ice flowed from the highest areas to the flatlands. The former extent of glaciers in the Alps has been intensively investigated for nearly 200 years, especially in the Western Alps thanks to the involvement of prominent geologists such as Louis Agassiz. More recently, the state of our knowledge has been completed from different methods and disciplines i) dating techniques such as the surface exposure dating of erratic boulders (Ivy-Ochs et al., 2008) permits to give chronological constraints ii) glacier modelling permits to simulate former ice flow an glacial extent by computer simulations (Seguinot et al., 2018) iii) proxy data – e.g. inferred from speleothems (Luetscher et al., 2015) – permits to give paleo climate information, which in turn provide hints on glacier former extents. The goal of this work is to review interdisciplinary literature that contributed to our current understanding of the extent, dynamics, and timing of glaciers in the Western Alps during the last glacial cycle. The topic will be tackled from three different perspectives: the knowledge inferred from geomorphological evidences (Kelly et al., 2004; Ivy-Ochs et al., 2008), from glacier modelling (Seguinot et al., 2018), and from climate proxy data (Luetscher et al., 2015). The student will be asked to do a synthesis, to assess the reliability and limitations of each approach, to critically evaluate and discuss the findings from the different methods, and to identify key open questions.
Aletsch Glacier is the largest glacier in the European Alps. It consists of 3 large accumulation basins merging into a long and curved tongue. Modelling the evolution of the glacier in the past is crucial to validate prognostic models. While historical stages of the glacier have been well-documented since a few centuries, the lack of data on former fluxes (the dynamic and the precipitation prevailing in the past) is key source of model uncertainty. The goal of this work is precisely to tackle this issue by embedding new radionuclide data in an existing glacier model. These data permit to track the trajectory of ice particles that have been deposited on the glacier surface between the early 50’s and 60’s (i.e. ice that have been contaminated with Pu-239 due to fallout from nuclear weapon tests performed during the cold war.) These tracers contain therefore valuable integrated-over-time information about the ice fluxes back to the 50s, and have not been integrated to Aletsch Glacier model to date. The task of the student will be to model the Aletsch Glacier with IGM – the Instructed Glacier Model – which combine the interaction between mass balance and ice flow. The student will use IGM to reconstruct the former ice dynamics and precipitation that best reproduce the tracing of ice revealed by the radionuclide data. Last, this new model will permit to derive a calibrated map of the age of ice over the entire surface of Aletsch Glacier.
Glacial records on the Jura mountain suggest that it has not been covered by ice from the Valais Glacier, but instead has hosted its own ice cap near the Last Glacial Maximum (LGM, about 24’000 years ago). This hypothesis was supported from boulder deposition elevation along the Jura by. Unfortunately, not always such an independent Jura ice cap was modelled in any recent modelling works. The goal of this work is to reconstruct the Jura ice cap at the LGM with IGM – the Instructed Glacier Model – which combines the interaction between mass balance and ice flow. The student will use IGM to reconstruct the former ice dynamics and precipitation that best reproduce the glacial footprints left by the glaciers. This will give new insights of the local climate that has prevailed over the Jura at LGM.
Aletsch Glacier is the greatest glacier of the European Alps. While its retreat since the end of the Little Ice Age (LIA) is well-documented, its dynamical evolution prior the LIA remains uncertain, especially, the question whether Alestch glacier have been smaller than it is today remains open. The goal of this thesis is to reconstruct the evolution of Aletsch Glacier during the last 2000 years with IGM – the Instructed Glacier Model – which combines the interaction between mass balance and ice flow. The student will use IGM to reconstruct the former ice dynamics and precipitation that best reproduce the presumable former state documented by Holzhauser and al. (2005). For a Master project, the study would be extended to explore the entire Holocene period based on local climate proxies.