IceAnatomy: a benchmark dataset and methodology for automatic ice boundary extraction from radio-echo sounding data

Dreier, Marcel; Koch, Moritz; Gourmelon, Nora; Blindow, Norbert; Steinhage, Daniel; Wu, Fei; Seehaus, Thorsten; Braun, Matthias; Maier, Andreas; Christlein, Vincent

doi:https://doi.org/10.5194/tc-19-5337-2025

Articles | Volume 19, issue 11

https://doi.org/10.5194/tc-19-5337-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/tc-19-5337-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 19, issue 11

Research article

|

04 Nov 2025

Research article |

| 04 Nov 2025

IceAnatomy: a benchmark dataset and methodology for automatic ice boundary extraction from radio-echo sounding data

Marcel Dreier, Moritz Koch, Nora Gourmelon, Norbert Blindow, Daniel Steinhage, Fei Wu, Thorsten Seehaus, Matthias Braun, Andreas Maier, and Vincent Christlein

Data sets

A Benchmark Dataset and Methodology for Automatic Ice Boundary Extraction from Radio-Echo Sounding Data (h5-v3) Marcel Dreier et al. https://doi.org/10.5281/zenodo.14036897

Model code and software

Implementation of Ice Anatomy: A Benchmark Dataset and Methodology for Automatic Ice Boundary Extraction from Radio-Echo Sounding Data Marcel Dreier https://doi.org/10.5281/zenodo.14038570

Short summary

In this paper, we present a ready-to-use benchmark dataset to train machine learning approaches for detecting ice thickness from radar data. It includes radargrams of glaciers and ice sheets alongside annotations for their air–ice and ice–bedrock boundary. Furthermore, we introduce a baseline model and evaluate the influence of several geographical and glaciological factors on the performance of our model.