Researchers are using machine learning to analyze satellite radar data to identify icebergs in the Southern Ocean around Antarctica to better understand their life cycles and environmental impacts.

Icebergs may seem exotic and distant, but as anyone who has seen the Titanic will tell you, they can have a profound effect on us when we least expect it. Just last week, the world’s largest iceberg, called A23a, which is more than twice the size of Greater London, broke free from the seafloor after remaining stranded for almost three decades and is drifting northward in the Antarctic Ocean. Meanwhile, thousands of smaller icebergs are constantly breaking off from Antarctica’s ice shelf and floating toward the sea.

The effect of all these icebergs is not just that they pose a danger to shipping. As they melt over decades, they release cold freshwater and nutrients that can alter local ecology, as well as the complex dynamics of ocean circulation, sea ice collapse, and even global sea levels.

The problem is, with all those ice chips bobbing around like giant mints, there are so many of them and they move so erratically that they’re hard to detect, let alone leave their traces. To help with this, a team of scientists funded by the Alan Turing Institute used Synthetic Aperture Radar (SAR) from ESA’s Sentinel-1 satellites, which can scan icebergs day and night in all weather conditions.

The radar data is not new, but it uses an unsupervised AI algorithm to analyze readings collected between October 2019 and September 2020. This algorithm revealed about 30,000 icebergs about 1 km² (0.4 mi²) or smaller in the western Amundsen Sea. Antarctica, on the calving front of Thwaites Glacier.

It is hoped that by accurately detecting and tracking icebergs it will be possible to create a digital replica of the Antarctic sea, enabling a better understanding of the complex physics of how the ocean, ice and atmosphere interact with each other.

“The technology we used to develop this device is already used quite widely in medical imaging, so we are excited to apply the same technology to the complex features seen in SAR satellite images of polar oceans,” said Ben Evans of the British Antarctic Survey. Survey (BAS) Artificial Intelligence Laboratory: “The method we use is as accurate as other alternative iceberg detection methods and outperforms most without the need for human intervention. This means it can easily scale beyond our research and even provide near real-time monitoring.”