Climate Research: Scientists Track Iceberg Life Cycle with AI

Scientists at the British Antarctic Survey (BAS) have developed a tool with artificial intelligence (AI) that allows them to automatically track the life cycle of icebergs as they journey through the world's oceans. This enables researchers to monitor their movements how they melt, break apart, and even reassemble. With this information, detailed “family trees” of icebergs can be created, providing insights into their impact on the ecosystem.

The BAS AI system analyzes satellite images and identifies individual icebergs. By analyzing their characteristic geometric shapes, each iceberg is assigned a unique identity. The system then regularly analyzes new satellite photos over a specific period to determine changes in the ice formation, such as those resulting from melting and calving. The broken-off pieces are then also tracked and remain associated with the original iceberg as its “offspring.”

Iceberg Tracking for Climate Research

Tracking icebergs, their calving, and melting, which releases large amounts of freshwater into the oceans, is important because it can alter ocean currents. This can have a significant impact on marine ecosystems and the global climate.

Until now, scientists could only track a few large icebergs. With the AI system, continuous observation is possible. “What's exciting about this is that we are finally getting the observations that have been missing. We've moved away from tracking just a few famous icebergs and are now creating complete family trees. For the first time, we can see where each fragment comes from, where it's going, and why it matters for the climate,” says Ben Evans, a machine learning expert at BAS.

Initial tests of the technology using observations of Arctic icebergs that had calved from the Petermann Glacier and other parts of Northwest Greenland have been promising. The system can thus show where meltwater enters the oceans. This information can then be used to improve global climate and ocean models, such as the European ocean model Nucleus for European Modeling of the Ocean (NEMO). The NEMO model is part of the UK Earth System Model (UKESM), an Earth system model used to make predictions, among other things, about how polar ice loss will affect climate change.

The determined data on icebergs can also be used in shipping, for example, to detect iceberg movements early in polar waters and react to them.

(olb)