27 Jun 2023
 | 27 Jun 2023

AI4SeaIce: Task Separation and Multistage Inference CNNs for Automatic Sea Ice Concentration Charting

Andreas Stokholm, Andrzej Kucik, Nicolas Longépé, and Sine Munk Hvidegaard

Abstract. We investigate how different Convolutional Neural Network (CNN) U-Net models specialised in addressing partial labelling tasks related to mapping Sea Ice Concentration (SIC) can improve performance. We use Sentinel-1 SAR images and human-labelled ice charts as the reference to train models that benefit from advantages gained from different model optimisation objectives by utilising a multistage inference scheme. We find our multistage model inference approach that apply a classification (CrossEntropy or Earth Mover's Distance squared) optimised model to separate open water, intermediate SIC and fully covered ice in conjunction with a regression (Mean Square Error or Binary CrossEntropy) optimised model, that assigns specific intermediate classes, to perform the best. To evaluate the models we introduce several specific metrics illustrating the performance in key areas, such as the separation of macro classes, intermediate class, and an accuracy metric better encapsulating uncertainties in the reference data. We achieve R2-score of ~93 %, similar to state-of-the-art in the literature (Kucik and Stokholm, 2023). However, our models exhibit significantly better open water and 100 % SIC detections. The multistage synergises high open water and fully covered sea ice accuracies achieved with classification optimised objectives with good intermediate class performance obtained by regressional loss functions. In addition, our findings indicate that the number of classes that the intermediate concentrations are compressed into does not influence the result significantly it is the loss function used to optimise the model that assigns the specific intermediate class to have the largest impact.

Andreas Stokholm et al.

Status: final response (author comments only)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2023-976', Eric Wendoloski, 02 Aug 2023
    • AC1: 'Reply to review', Andreas Stokholm, 05 Sep 2023
  • RC2: 'Comment on egusphere-2023-976', Anonymous Referee #2, 19 Aug 2023
    • AC1: 'Reply to review', Andreas Stokholm, 05 Sep 2023

Andreas Stokholm et al.

Andreas Stokholm et al.


Total article views: 289 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
219 55 15 289 11 8
  • HTML: 219
  • PDF: 55
  • XML: 15
  • Total: 289
  • BibTeX: 11
  • EndNote: 8
Views and downloads (calculated since 27 Jun 2023)
Cumulative views and downloads (calculated since 27 Jun 2023)

Viewed (geographical distribution)

Total article views: 281 (including HTML, PDF, and XML) Thereof 281 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
Latest update: 30 Sep 2023
Short summary
To navigate in the Arctic, vessels utilise maps of the local sea ice conditions, that are manually drawn daily. This is a time-consuming and labour-intensive task, which we are trying to automate using convolutional neural networks. In this article, we investigate combining the outputs of models trained with different objectives, leveraging the strengths and avoiding weaknesses to create precise and detailed sea ice concentration maps for navigation in the Arctic and climate or weather models.