The Arctic region is currently experiencing rapid changes that are mainly driven by global warming. Over the last decade, we have observed a drastic shrinkage of the Arctic sea ice cover in summer accompanied by a spectacular thinning of mean thickness and an increased export of sea ice through the Fram Strait. These changes perturb the interactions between the atmosphere and the ocean and modify the Arctic climate.

To understand these changes, predict the state of the sea ice cover in the future and also produce short-term sea ice forecasts, numerical models implemented in realistic settings are necessary. These models are usually taken to be a 3-layered system where the sea ice cover, the ocean below and the atmosphere above are physically coupled. Several studies have reported that the current models poorly represent the sea ice dynamics, and in particular miss the observed acceleration of the sea ice motions over the last 30 years, as well as their seasonal cycle.

Sea ice dynamics are very complex and show strong similarities with plate tectonics and earthquake dynamics. Failure to represent these dynamics partly explains why current coupled models do not reproduce the recent changes in the Arctic, and presents one of the main challenges faced today by the Arctic modelling community at large.

The aim of this project is to demonstrate the validity of a completely new modelling approach to reproduce the dynamics of the Arctic ice pack. The model, which is inspired by solid mechanics, will be further developed, implemented in a coupled model framework on a parallelized supercomputer, and validated against satellite observations. Thanks to the development and implementation of our new elasto-brittle rheology, a model will be capable of simulating the drift and deformation of sea ice using the appropriate physics for the first time.