Skip to main content
Go to the home page of the European Commission (opens in new window)
English en
CORDIS - EU research results
CORDIS

Advancing Projections of Antarctic Ice-Ocean Interactions with Physics-Informed Neural Networks

Project description

A new framework to understand Antarctic ice-ocean interactions

Warm ocean currents that melt ice shelves from below are threatening the stability of the Antarctic ice sheet, which is a regulator of global sea levels. Monitoring these processes is important but not easy to do because there is a lack of subsurface data and traditional ocean models are costly. Supported by the Marie Skłodowska-Curie Actions programme, the CAPTSA project aims to use the first physics-informed neural network framework dedicated to Antarctic ice–ocean interactions. It will merge satellite data with fundamental physical laws, and reconstruct daily full-depth ocean temperatures and circulation patterns. This approach allows researchers to link climate forcing to specific melt rates. Ultimately, CAPTSA provides the data needed to accurately predict Antarctic contributions to future sea-level rise.

Objective

The Antarctic ice sheet, a critical regulator of global sea level, is losing mass mainly through basal melting driven by warm Circumpolar Deep Water (CDW) intrusions. These are controlled by the Antarctic Slope Front (ASF), whose variability dictates when and where CDW can access ice-shelf cavities. Understanding these processes is limited by sparse in situ data, satellites’ inability to observe subsurface conditions, and the coarse or computationally expensive nature of existing ocean models.
This project will develop CAPTSA (Circum-Antarctic Physics-informed model for Temperature, Salinity, and Advection), the first Physics-Informed Neural Network (PINN) framework to enhance our understanding of Antarctic ice-ocean interactions. CAPTSA will integrate satellite-derived data, limited in situ observations, and governing physical equations to reconstruct the daily full-depth structure of temperature, salinity, and circulation across the Southern Ocean and Antarctic shelves.
By combining the flexibility of machine learning with physical consistency, CAPTSA will overcome challenges of data sparsity, coarse resolution, poorly constrained parametrization schemes, and computational cost. Its time series predictions will be employed in the analysis of ASF dynamics, CDW intrusions, and cross-shelf exchanges across fresh, warm, and dense shelf regimes. Further, by linking the CAPTSA predictions with satellite- and radar-derived melt estimates, CAPTSA will allow causal attribution between external climate forcing, oceanic variability, and ice-shelf melt rates. This integrative approach will advance process-based understanding of ice-ocean feedbacks and provide a robust framework for regime-specific projections of basal melting and Antarctic contributions to sea-level rise.

Fields of science (EuroSciVoc)

Keywords

Programme(s)

Topic(s)

Funding Scheme

HORIZON-TMA-MSCA-PF-EF - HORIZON TMA MSCA Postdoctoral Fellowships - European Fellowships

See all projects funded under this funding scheme

Call for proposal

(opens in new window) HORIZON-MSCA-2025-PF

See all projects funded under this call

Coordinator

NORSK POLARINSTITUTT
Net EU contribution
€ 251 578,56
Address
HJALMAR JOHANNESENS 14
9296 Tromsoe
Norway
Activity type
Research Organisations
Links
Total cost
No data

Partners (1)