Preprints
https://doi.org/10.5194/essd-2022-139
https://doi.org/10.5194/essd-2022-139
 
02 May 2022
02 May 2022
Status: this preprint is currently under review for the journal ESSD.

Calving Fronts and Where to Find Them: A Benchmark Dataset and Methodology for Automatic Glacier Calving Front Extraction from SAR Imagery

Nora Gourmelon1, Thorsten Seehaus2, Matthias Braun2, Andreas Maier1, and Vincent Christlein1 Nora Gourmelon et al.
  • 1Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany
  • 2Institut für Geographie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany

Abstract. Exact information on calving front positions of marine- or lake-terminating glaciers is a fundamental glacier variable for analyzing ongoing glacier change processes and assessing other variables like frontal ablation rates. In recent years, researchers started implementing algorithms that could automatically detect the calving fronts on satellite imagery. Most studies use optical images, as in these images, calving fronts are often easy to distinguish due to sufficient spatial resolution and the presence of different spectral bands, allowing the separation of ice features. However, detecting calving fronts on SAR images is highly desirable, as SAR images can also be acquired during the polar night and are independent of weather conditions, e.g., cloud cover, facilitating all-year monitoring worldwide. In this paper, we present a benchmark dataset of SAR images from multiple regions of the globe with corresponding manually defined labels to train and test approaches for the detection of glacier calving fronts. The dataset is the first to provide long-term glacier calving front information from multi-mission data. As the dataset includes glaciers from Antarctica, Greenland and Alaska, the wide applicability of models trained and tested on this dataset is ensured. The test set is independent of the training set so that the generalization capabilities of the models can be evaluated. We provide two sets of labels: one binary segmentation label to discern the calving front from the background and one for multi-class segmentation of different landscape classes. Unlike other calving front datasets, the presented dataset contains not only the labels but also the corresponding preprocessed and geo-referenced SAR images as PNG files. The ease of access to the dataset will allow scientists from other fields, such as data science, to contribute their expertise. With this benchmark dataset, we enable comparability between different front detection algorithms and improve the reproducibility of front detection studies. Moreover, we present one baseline model for each kind of label type. Both models are based on the U-Net, one of the most popular deep learning segmentation architectures. Additionally, we introduce Atrous Spatial Pyramid Pooling to the bottleneck layer. In the following two post-processing procedures, the segmentation results are converted into one-pixel-wide front delineations. By providing both types of labels, both approaches can be used to address the problem. To assess the performance of the models, we first review the segmentation results using the recall, precision, F1-score, and the Jaccard Index. Second, we evaluate the front delineation by calculating the mean distance error to the labeled front. The presented vanilla models provide a baseline of 150 m ± 24 m mean distance error for the Mapple Glacier in Antarctica and 840 m ± 84 m for the Columbia Glacier in Alaska, which has a more complex calving front, consisting of multiple sections, as compared to a laterally well constrained, single calving front of Mapple Glacier.

Nora Gourmelon et al.

Status: open (until 27 Jun 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Nora Gourmelon et al.

Data sets

CaFFe (CAlving Fronts and where to Find thEm: a benchmark dataset and methodology for automatic glacier calving front extraction from sar imagery) Gourmelon, Nora; Seehaus, Thorsten; Braun, Matthias Holger; Maier, Andreas; Christlein, Vincent https://doi.pangaea.de/10.1594/PANGAEA.940950

Model code and software

Calving Fronts and Where to Find Them Gourmelon, Nora https://doi.org/10.5281/zenodo.6469519

Nora Gourmelon et al.

Viewed

Total article views: 249 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
190 52 7 249 5 4
  • HTML: 190
  • PDF: 52
  • XML: 7
  • Total: 249
  • BibTeX: 5
  • EndNote: 4
Views and downloads (calculated since 02 May 2022)
Cumulative views and downloads (calculated since 02 May 2022)

Viewed (geographical distribution)

Total article views: 206 (including HTML, PDF, and XML) Thereof 206 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 26 May 2022
Download
Short summary
Ice loss of glaciers shows in retreating calving fronts, i.e. the position where icebergs break off the glacier and drift into the ocean. This paper presents a benchmark dataset for calving front delineation in Synthetic Aperture Radar (SAR) images. The dataset can be used to train and test deep learning techniques, which automate the monitoring of the calving front. Provided example models achieve front delineations with an average distance of 887 meters to the correct calving front.