Researchers at the British Antarctic Survey have developed an artificial intelligence system that can automatically track icebergs throughout their entire lifecycle, from formation to fragmentation and melting, across oceans around the world, El.kz reports citing Interesting Engineering.
The tool uses satellite imagery to identify individual icebergs, assign them unique identities, and follow them over time as they drift.
Crucially, when a large iceberg breaks apart, the system can trace the resulting fragments back to their original source, effectively reconstructing detailed iceberg “family trees.”
This marks a major shift from how iceberg tracking has traditionally been done.
Until now, scientists could only follow a small number of large, named icebergs manually, while thousands of smaller fragments went untracked once they broke off and dispersed across the ocean.
That gap has limited scientists’ understanding of how meltwater from icebergs affects the climate. As icebergs melt, they release large volumes of fresh water into the ocean, altering ocean circulation, influencing marine ecosystems, and affecting global climate patterns.
Where that fresh water enters the ocean matters, but researchers have struggled to monitor it at scale.
Ben Evans, lead author of the study and a machine learning expert at the British Antarctic Survey, said the system finally fills that blind spot.
“What’s exciting is that this finally gives us the observations we’ve been missing,” Evans said. “We’ve gone from tracking a few famous icebergs to building full family trees.”
“For the first time, we can see where each fragment came from, where it goes and why that matters for the climate,” he added.
The AI works by analyzing the unique geometric shapes of icebergs captured in satellite images.
When an iceberg fractures, the system performs what researchers describe as a digital jigsaw puzzle, matching the shapes of scattered fragments back to their original parent iceberg.
The approach was tested using real satellite observations of icebergs calved from Petermann Glacier and other locations in north-west Greenland. The system was able to consistently track how these icebergs split, drifted, and melted over time.
By mapping where iceberg meltwater is released, the tool provides data that can be fed directly into ocean and climate models.
The information will be incorporated into the NEMO ocean model, which forms part of the UK Earth System Model, improving predictions as polar ice loss accelerates with rising global temperatures.