AI Accurately Tracks Shrinking Glaciers Using Satellite Images

A new AI-powered approach significantly improves the accuracy of tracking the world's shrinking glaciers by analyzing satellite imagery. This is crucial for monitoring climate change and predicting future sea-level rises. Glaciers flowing into the ocean are particularly important as their melting contributes to rising sea levels and can alter ocean currents. Additionally, shrinking glaciers expose darker surfaces that absorb more solar heat, exacerbating warming. While manual tracking is time-consuming and insufficient for the vast number of glaciers, AI offers a potential solution. Researchers from the Friedrich-Alexander University of Erlangen–Nuremberg (FAU) in Germany have developed a deep learning model that can be adapted to new locations with minimal data. Their method reduces the error in identifying glacier calving fronts from over a kilometer to less than 70 meters. This was achieved by providing the AI with a single hand-labeled image per glacier, unlabeled summer reference images, and a map of the underlying rock. The team has already applied this technique to monitor 145 glaciers in Norway's Svalbard archipelago from 2015 to 2024, generating monthly calving front data. They aim to expand this to an additional 1,500 Arctic glaciers, providing a more detailed understanding of glacial responses to climate change. This advancement could enable automated, long-term monitoring of glaciers globally.