Last month a new project has started, with several of our NERSC researchers involved. The "Scale-Aware Sea Ice Project" (SASIP) will run for six years, with international partners in France, the US, the UK, Italy, and Mali.

SASIP is being led by Pierre Rampal (Centre National de la Recherche Scientifique), who initiated the proposal during his primary time at NERSC. Einar Ólason (NERSC) is leading one of the work packages (Sea Ice Code Development - Optimization), and NERSC researchers Laurent Bertino, Julien Brajard, Guillaume Boutin, Anton Korosov, and Timothy Williams will contribute to several of the other work packages.

The problem: Increasingly warming polar regions and insufficient climate models

The polar regions are experiencing amplified warming, up to three times faster than anywhere else on Earth. The impact the warming has on these regions is poorly understood though. Generally, we use climate models to try to understand what will happen to the global climate. Several different models are coupled together in Earth System Models (climate models). These different models are for example ocean models, atmospheric models, and sea ice models. Coupled climate models have become realistic in many aspects, but the climate in the polar regions is generally not predicted well enough. Poorly represented are for example Arctic sea ice loss and Antarctic sea ice variability. The complex character of sea ice dynamics is not being captured well enough by current sea ice models that tie into coupled climate models. Therefore, Pierre Rampal and his team are suggesting that the sea ice models themselves are a reason for why coupled climate models do not perform well in the polar regions. To combat this problem and to improve our understanding of what warming does to the polar regions, they created SASIP.

Credit: V. Dansereau

The solution: Creating the best sea ice model out there

The aim of SASIP is to produce a paradigm-shifting sea ice model that will perform remarkably better than other sea ice models and allow for a much-increased understanding of the future of our polar regions. The base of this new and improved sea ice model is neXtSIM, a model that has been developed in recent years at NERSC. Properly addressing how sea ice dynamics are being represented in a sea ice model is critical for it producing results as close to nature as possible. In neXtSIM a novel approach is being used that includes a rheology (i.e., the laws that dictate the way sea ice breaks up and deforms) based on solid-like mechanics principles, while the rheologies of all other current sea ice models are based on fluid-like mechanics principles. In its current form, neXtSIM already shows promising results (Rampal et al. 2019; Olason et al. 2021). Further improving it will hopefully result in the best sea ice model out there. It will make predicting the climate in polar regions more reliable by better simulating coupled ocean-ice-atmosphere interactions.

The path to get there: Combining sea ice coding, remote sensing data, machine learning, and data assimilation

SASIP-Principal Investigator Pierre Rampal in the Fram Strait.: Credit: Marta Lorenz

Innovating the model requires combining several fields that we at NERSC are great at. The partners in SASIP provide excellent expertise complementing ours. The sea ice model neXtSIM will be improved by making use of the vast number of observational datasets available now and in the future. More and more satellite missions provide remote sensing data of the polar regions, and they are invaluable to SASIP. Combining machine learning and data assimilation and including it in the model setup is a clever way to objectively combine model and data, and thereby improve neXtSIM. The coming six years will be an exciting time for sea ice modelling, thanks to SASIP.

Pierre Rampal (PI), on the project: “SASIP is an exciting and extremely ambitious project, which will transform the way we understand the role of sea ice in the polar regions, and subsequently in the global climate system.”

This project is supported by Schmidt Futures, a philanthropic initiative founded by Eric and Wendy Schmidt that bets early on exceptional people making the world better, particularly through innovative breakthroughs in science and technology. It is a project of Schmidt Futures' Virtual Earth System Research Institute (VESRI). 

Partner institutions:

Centre National de la Recherche Scientifique (CNRS), France

Sorbonne University (US), France

Ecole Nationale des Ponts et Chaussées (ENPC), France

Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique (CERFACS), France

Mercator Ocean International (MOI), France

Brown University (Brown), United States

Nansen Environmental and Remote Sensing Center (NERSC), Norway

University of Reading (UoR), United Kingdom

Euro-Mediterranean Center on Climate Change (CMCC), Italy

University of Bamako (UoB), Mali

Duration and budget: 6 years, starting in April 2021. Total budget of about 10.5M$.

Project website: https://sasip-climate.github.io/

Follow SASIP on their Twitter account for future updates: https://twitter.com/sasipclimate @sasipclimate

Selected publications on sea ice modelling and neXtSIM relevant for SASIP:

Ólason, E., Rampal, P., & Dansereau, V. (2021). On the statistical properties of sea-ice lead fraction and heat fluxes in the Arctic. The Cryosphere, 15(2), 1053–1064. https://doi.org/10.5194/tc-15-1053-2021

Rampal, P., Bouillon, S., Ólason, E., & Morlighem, M. (2016). neXtSIM: a new Lagrangian sea ice model. The Cryosphere, 10(3), 1055–1073. https://doi.org/10.5194/tc-10-1055-2016

Rampal, P., Dansereau, V., Ólason, E., Bouillon, S., Williams, T., Korosov, A., & Samaké, A. (2019). On the multi-fractal scaling properties of sea ice deformation. The Cryosphere, 13(9), 2457–2474. https://doi.org/10.5194/tc-13-2457-2019