Arctic AI

Program Description

Arctic AI is vital for understanding rapid environmental change, supporting sustainable development, and protecting vulnerable ecosystems in a geopolitically important area for Denmark, Norway, and the countries in the Nordics and the UK.

By enabling precise monitoring, forecasting, and decision-making in remote regions, the ambition is to strengthen climate resilience, enhance community safety, and guide responsible resource management in the Arctic.

The insights that will be gained from the Arctic AI P1 Program, aims to safeguard global environmental stability for future generations to come, starting with focusing on the concrete problem of forecasting sea ice in the Arctic region. 

The program seeks to create a high-impact collaborative platform by connecting two major Nordic initiatives: 

1. The IcyAlert project (Denmark), funded by the Novo Nordic Foundation, is connected to the Gefion supercomputer, and was recently launched as a collaboration between DMI, DTU, and the Royal Meteorological Institute of Belgium (RMI).
2. The IceNet project (UK and Norway) is a joint project conducted as part of a Memorandum of Understanding between the Visual Intelligence Research Centre in Norway and the Alan Turing Institute in the UK, anchored with the British Antarctic Survey.

Find more information on the IcyAlert project & the IceNet project.

By pioneering physics-informed ML and probabilistic AI in the climate and Earth domain, the program will leverage cutting-edge deep learning architectures and generative models trained on massive datasets to resolve fine-scale variability that existing Earth System Models often miss.

Furthermore, it will explore causal AI techniques and physics-informed modeling to identify robust predictor-predictand relationships, ensuring that the AI-driven forecasts remain both physically consistent and explainable.