The fate of the Antarctic ice sheet is the largest source of uncertainty in future sea level projections. The magnitude and sign of the Antarctic contribution indeed results from the compensation of two opposite effects: increased surface mass gain and increased ocean-induced dynamical mass loss, both of which are highly uncertain. A large part of the uncertainty on these opposite effects comes from the absence of coupling between ice sheet and ocean/atmosphere models, which has motivated the recent development of such coupled models. However, there is currently a mismatch between the coarse resolution of ocean/atmosphere models and the high resolution needed for ice sheet models near their grounding lines and at their edge.
In this project, we aim to improve the integration of the Antarctic ice sheet into an Earth System Model through the use of neural networks at the coupling interfaces. These will bring increased resolution and account for polar processes absent or poorly represented in Earth System Models (e.g., surface melt and runoff, ice-shelf basal melt). Neural networks will be trained on high-resolution polar-oriented atmospheric and oceanic simulations, including in a warmer climate and with modified ice sheet geometry. Members of our consortium have recently conducted two pilot studies on neural networks that serve as proofs of concept for this project.
The two neural networks will be used at the interface between the Elmer/Ice ice sheet model and the ocean and atmosphere components of IPSL-CM6-LR. We will assess the model’s ability to reproduce the observed evolution of the Antarctic ice sheet. We will also run projections to 2100 with uncertainty constrained by the newly developed neural network interfaces. The proposed coupling through neural network interfaces will be applicable to other climate models, including those that have recently been coupled without refined interface, as well as to standalone ice sheet models.
Our project is in line with the “Artificial Intelligence to address societal challenges” part of the French strategy regarding artificial intelligence, and will have strong societal impact (coastal management) through the improvement of sea level rise projections.
Monsieur Nicolas Jourdain (Institut des Géosciences de l'Environnement)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
IGE Institut des Géosciences de l'Environnement
IPSL Centre national de la recherche scientifique
Help of the ANR 567,301 euros
Beginning and duration of the scientific project: February 2023 - 48 Months