Artificial intelligence for mapping glacier calving front positions

Although the Greenland Ice Sheet is often divided into individual catchments due to its spatially differentiated dynamics, its totality embodies a connected glacier system.
Ice masses are discharged into the ocean through a multitude of outlet glaciers. Mapping the frontal positions of these outlet glaciers and their temporal variations contributes significantly to the understanding of ice dynamics and thus to more reliable modeling. The increasing availability and quality of remote sensing data enables an accurate and continuous mapping of calving front locations. However, the huge amount of data also accentuates the necessity for automated analysis strategies. Most current calving front data products are based on manual delineation – a very laborious and time-consuming process.

In the recently published study "Extracting Glacier Calving Fronts by Deep Learning: The Benefit of Multispectral, Topographic and Textural Input Features", we present a method for automated calving front delineation. For this we apply artificial intelligence methods, specifically deep learning, to optical satellite imagery. We also evaluate the benefit of multi-spectral, topographic and textural input features on the prediction performance of our algorithm.

The research is part of the multidisciplinary research project "Artificial Intelligence for Cold Regions" (AI-CORE), which we are working on in collaboration with partners from the German Aerospace Center (DLR) and the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI).

E. Loebel, M. Scheinert, M. Horwath, K. Heidler, J. Christmann, L. D. Phan, A. Humbert, X. X. Zhu. "Extracting Glacier Calving Fronts by Deep Learning: The Benefit of Multispectral, Topographic and Textural Input Features," in IEEE Transactions on Geoscience and Remote Sensing, 2022, doi: 10.1109/TGRS.2022.3208454.