As the British Antarctic Survey (BAS) reports, scientists have developed a novel AI system that can, for the first time, automatically track the entire life cycle of icebergs, from their formation through drifting and melting to their breakup into numerous fragments.
Using satellite imagery, the artificial intelligence identifies individual icebergs, assigns unique identifiers, and can re-associate the resulting pieces with the original iceberg even after a breakup. This creates detailed digital “family trees” showing where each ice fragment originated and how it continues to evolve.
This information is crucial for climate research, as melting icebergs release large amounts of freshwater into the sea, which can influence ocean currents, ecosystems, and global climate patterns. Until now, tracking smaller ice fragments has been a major source of uncertainty in climate and ocean models.
“For the first time, we can trace where every single ice fragment comes from and why that matters for the climate,” says Ben Evans, lead author of the study and a machine-learning expert at BAS.
The data are incorporated into the NEMO ocean model, which is part of the UK Earth System Model, improving predictions of oceanic and climatic processes. In addition, the approach offers practical benefits for navigation in polar waters.
The research was funded by the EPSRC grant EP/Y028880/1, the Alan Turing Institute, and the British Antarctic Survey’s Polar Science for a Sustainable Planet program.
Heiner Kubny, PolarJournal