16 Feb 2024
 | 16 Feb 2024

Developing a deep learning forecasting system for short-term and high-resolution prediction of sea ice concentration

Are Frode Kvanum, Cyril Palerme, Malte Müller, Jean Rabault, and Nick Hughes

Abstract. There has been a steady increase of marine activity throughout the Arctic Ocean during the last decades, and maritime end users are requesting skillful high-resolution sea ice forecasts to ensure operational safety. Different studies have demonstrated the effectiveness of utilizing computationally lightweight deep learning models to predict sea ice properties in the Arctic. In this study, we utilize operational atmospheric forecasts as well as ice charts and sea ice concentration passive microwave observations as predictors to train a deep learning model with ice charts as the ground truth. The developed deep learning forecasting system can predict regional sea ice concentration at one kilometer resolution for 1 to 3-day lead time. We validate the deep learning system performance by evaluating the position of forecasted sea ice concentration contours at different concentration thresholds. It is shown that the deep learning forecasting system achieves a lower error for several sea ice concentration contours when compared against baseline-forecasts (persistence-forecasts and a linear trend), as well as two state-of-the-art dynamical sea ice forecasting systems (neXtSIM and Barents-2.5) for all considered lead times and seasons.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Are Frode Kvanum, Cyril Palerme, Malte Müller, Jean Rabault, and Nick Hughes

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-2023-3107', Anonymous Referee #1, 03 Apr 2024
    • AC1: 'Reply on RC1', Are Frode Kvanum, 03 Jun 2024
  • RC2: 'Comment on egusphere-2023-3107', Anonymous Referee #2, 25 Apr 2024
    • AC2: 'Reply on RC2', Are Frode Kvanum, 03 Jun 2024
Are Frode Kvanum, Cyril Palerme, Malte Müller, Jean Rabault, and Nick Hughes

Model code and software

Project repository Are Frode Kvanum

Are Frode Kvanum, Cyril Palerme, Malte Müller, Jean Rabault, and Nick Hughes


Total article views: 488 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
346 112 30 488 20 15
  • HTML: 346
  • PDF: 112
  • XML: 30
  • Total: 488
  • BibTeX: 20
  • EndNote: 15
Views and downloads (calculated since 16 Feb 2024)
Cumulative views and downloads (calculated since 16 Feb 2024)

Viewed (geographical distribution)

Total article views: 499 (including HTML, PDF, and XML) Thereof 499 with geography defined and 0 with unknown origin.
Country # Views %
  • 1


Latest update: 23 Jul 2024
Short summary
Recent studies have shown that machine learning models are effective at predicting sea ice concentration, yet few have explored the development of such models in an operational context. In this study, we present the development of a machine learning forecasting system which can predict sea ice concentration at 1 km resolution, up to 3 days ahead using real time operational data. The developed forecasts predict the sea ice edge position with a better accuracy than physical and baseline forecasts.