Loading...
Thumbnail Image
Publication

An artificial intelligence and remote sensing approach to iceberg distribution around the Greenland Ice Sheet

Shankar, Siddharth
Citations
Altmetric:
Abstract
The Greenland Ice Sheet has been losing mass at an accelerated rate, from 41±17 Gt/yrin the 1980s, to 286 ± 20 Gt/yr in the 2010s, to 532 ± 58 Gt/yr in 2019 [Mouginot et al., 2019, Velicogna et al., 2020, Mohajerani, 2020]. In addition to raising sea level by 13.7 ± 1.1 mm since 1972 [Mouginot et al., 2019], roughly half of this mass loss occurs due to iceberg calving and submarine melt at outlet glacier termini [Benn et al., 2017]. As a result, icebergs act as an important link between ice sheet and ocean processes. Changes in iceberg size and abundance can reflect changes in glacier dynamics (ice velocity, rheological damage) and geometry (buoyancy condition, ice thickness). In turn, the size and abundance of icebergs can impact their melt rates and thus influence fjord circulation [Moon et al., 2018]. Despite their importance, production of an ice-sheet wide, year-round, inventory of iceberg distribution around Greenland was not possible until recently. Greenland icebergs are relatively small (1000 m2 - 2 km2) and exist in complex environmental settings (pro-glacial m´elange, sea ice, waves). Using publicly-available imagery and a cloud-based computing platform, we can efficiently produce iceberg distributions around the Greenland Ice Sheet, every 2-weeks, year-round. Through this research, we improve our understanding of iceberg patterns and trajectories,which is essential for climate models, climate projections, and shipping logistics. Importantly, we can deduce where icebergs tend to travel, how quickly they move, and whether glacier, fjord or weather conditions are more likely to modulate their trajectories. High spatial and temporal resolution imagery that can provide continuous iceberg observations will also be useful for studies focused on freshwater flux, ocean circulation, calving laws, and iceberg trajectories.
Description
Date
2022-12-31
Journal Title
Journal ISSN
Volume Title
Publisher
University of Kansas
Collections
Research Projects
Organizational Units
Journal Issue
Keywords
Geology, Deep learning, Greenland, Iceberg distribution, Iceberg tracking, Machine learning, Sentinel-1 SAR
Citation
DOI
Embedded videos