Antarctic iceberg freeboard retrieval via deep learning from optical satellite imagery
Year: 2025
Type: article
Abstract: Antarctic icebergs are critical components of the Antarctic ice sheet-ice shelf-ocean system and play a key role in understanding the impacts of climate change, particularly in estimating iceberg volume through the freeboard measurements. This study presents an innovative deep learning-based approach (U-Net) for automatically measuring iceberg freeboard from optical remote sensing data in the Antarctic coastal region, following data acquisition. Landsat 8 imagery from September 2022 was used to test the method, successfully extracting 85,083 icebergs, the maximum freeboard reaching 95.16 m at (66.44°S, 91.28°E). Results revealed significant regional disparities in iceberg distribution, with the West Antarctic region exhibiting a higher iceberg density than the East Antarctica region. Iceberg coverage was notably sparse around large ice shelves. Freeboard analysis categorized icebergs into five height levels, with small to medium-sized icebergs (less than 25 m) comprising 80.07% of the total. A decreasing trend in freeboard was observed with increasing distance from the coastline. The study also identified the Bellingshausen-Amundsen Sea region as having the highest iceberg concentration, with consistent density patterns across height categories in regions between 90°W–160°W and 15°W–30°W. This research provides valuable data for understanding iceberg formation and its potential impacts on oceanic and climatic systems.
Institution: Wuhan University
Language: English
Subfield: Atmospheric Science
Field: Earth and Planetary Sciences
Domain: Physical Sciences
SDG: Climate action
Open Access status: gold
Estimated APC paid: $2,390
Awards: G5237616405, G6882039737
OpenAlex ID: W4413275590