Uusinoka et al. (2025) Scale invariance in kilometer-scale sea ice deformation

Identification

• Journal: ˜The œcryosphere
• Year: 2025
• Date: 2025-12-03
• Authors: Matias Uusinoka, Jari Haapala, Jan Åström, Mikko Lensu, Arttu Polojärvi
• DOI: 10.5194/tc-19-6493-2025

Research Groups

• Aalto University, School of Engineering, Department of Energy and Mechanical Engineering, Finland
• Finnish Meteorological Institute, Marine Research Unit, Finland
• CSC – IT Center for Science Ltd., Finland

Short Summary

This study uses high-resolution MOSAiC ship radar data and a deep learning optical flow technique to investigate kilometer-scale sea ice deformation, revealing a lower limit of approximately 100 meters for scale-invariance in winter pack ice that disappears in summer.

Objective

• To identify the lower spatial limit of scale invariance in kilometer-scale sea ice deformation and understand its seasonal variability, particularly in relation to the validity of continuum models.

Study Configuration

• Spatial Scale: 10 km × 10 km domain, with a 20 m spatial resolution for deformation estimates. Nominal spatial scales for analysis ranged from 20 meters to 800 meters.
• Temporal Scale: Nine-month period (mid-October 2019 to end of September 2020), with a 10 minute temporal resolution for deformation estimates. Nominal temporal scales for analysis ranged from 10 minutes to 12 hours.

Methodology and Data

• Models used:
  • Novel deep learning-based optical flow tool built on Recurrent All-pairs Field Transforms (RAFT) architecture.
  • Green-Lagrange strains for deformation estimates, accounting for large strains and rigid body rotations.
  • Dispersion method for spatiotemporal scaling analysis.
• Data sources:
  • High-resolution ship radar imagery collected during the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) expedition on RV Polarstern.
  • Original data: 2.4 second intervals, 20 km × 20 km area, 8.33 m resolution.
  • Processed data: 1 minute temporal resolution, cropped to 10 km × 10 km, spatially averaged to 10 m resolution.

Main Results

• Sea ice within the 10 km domain experienced extended periods of quiescence during winter, punctuated by intermittent, localized deformation events.
• A lower limit for scale-invariance (L_c) of approximately 100 meters was observed for winter pack ice.
• This 100 meter threshold remained stable throughout the winter season, suggesting an intrinsic mechanical constraint, even as deformation features became more localized.
• The 100 meter limit was found to be independent of the temporal scale, while the magnitude of error in scale invariance increased with spatial and temporal scale.
• In summer, as the ice transitioned to a floe-dominated configuration in the marginal ice zone, no comparable scaling signature emerged, with deformation becoming more continuous and viscous-like.
• The spatial scaling exponent (β) varied between 0.4 and 0.6 for winter cases at L ≥ 150 meters and T = 10 minutes, decreasing with increasing T. For L ≤ 150 meters, β was approximately constant.
• The observed ~100 meter length scale coincides with the typical width of major leads and shear zones in the radar data.

Contributions

• Provides the first empirical observation of a lower spatial limit for scale invariance in sea ice deformation at approximately 100 meters, extending previous studies that typically reached 250 meters.
• Challenges conventional assumptions regarding the validity of continuum models at very small scales (e.g., comparable to ice thickness) for winter pack ice.
• Highlights the significant seasonal variability of scale invariance, demonstrating its breakdown in summer marginal ice zone conditions.
• Leverages a novel deep learning-based optical flow technique on unique high-resolution MOSAiC ship radar data, enabling unprecedented fine-scale analysis over a large domain for an entire season.
• Offers crucial insights for the development of hybrid modeling frameworks that integrate continuum modeling with discrete element methods to better account for discontinuities in sea ice dynamics.

Funding

• Research Council of Finland (project 347802, DEMFLO: Discrete Element Modeling of Continuous Ice Floes and Their Interaction)
• European Union's Horizon 2020 research and innovation programme (grant agreement no. 101003826, project CRiceS)
• Nordic Council of Ministers (NOCOS DT project)
• CSC – IT Center for Science, Finland (computational resources under project 2006428, DEMFLO)
• Multidisciplinary drifting Observatory for the Study of the Arctic Climate (MOSAiC) project (tag MOSAiC20192020, ProjectID:AWIPS122_00)

Citation

@article{Uusinoka2025Scale,
  author = {Uusinoka, Matias and Haapala, Jari and Åström, Jan and Lensu, Mikko and Polojärvi, Arttu},
  title = {Scale invariance in kilometer-scale sea ice deformation},
  journal = {˜The œcryosphere},
  year = {2025},
  doi = {10.5194/tc-19-6493-2025},
  url = {https://doi.org/10.5194/tc-19-6493-2025}
}