by Sukun Cheng, February 2025
The research presented here was carried out during my postdoctoral position at the Nansen Environmental and Remote Sensing Center (NERSC) in Bergen, Norway. My work at NERSC focused on improving sea ice forecast skill by adapting advanced data assimilation techniques. I joined the University of Reading in 2024 as an NCEO strategic research scientist.
Why is Arctic Sea Ice Forecasting Important?
Arctic sea ice is a critical component of the Earth’s climate system, and accurate sea ice forecasting is becoming increasingly important.
- The evident loss of Arctic sea ice both as sea ice extent (SIE) and volume (SIV) over recent decades has been abundantly discussed [1][2].
- Predicting the Arctic sea ice conditions from near-term to decadal timescales is important due to its significance for the Earth’s climate, local ecosystem and human activities.
Data Assimilation in Sea Ice Forecasting
neXtSIM is a novel dynamic-thermodynamic sea ice model. The model employs a Lagrangian mesh, which follows the motion and deformation of the ice pack. This approach reduces computational costs in the advection process while preserving regional characteristics within the sea ice field. However, comparing variables and performing ensemble statistics becomes challenging because each ensemble member employs a different Lagrangian mesh that deforms over time. To resolve this challenge, we introduced a common, fixed reference mesh for the ensemble. The model states are interpolated onto the reference mesh during the data assimilation step. An open-source data assimilation package EnKF-C is adapted in the assimilation process.
The study [3] used the deterministic ensemble Kalman filter (DEnKF) to assimilate both satellite observations of sea ice concentration (SIC) and sea ice thickness (SIT). The ensemble was generated by perturbing the atmosphere, ocean and sea ice variables [4]. There is a relatively long history of assimilating SIC observations but SIT observations become available in recent years.

Improved Accuracy of Sea Ice Coverage Simulation Across the Arctic Basin
The authors investigated different data assimilation strategies, including assimilating SIC and SIT individually and jointly and varying the frequency of assimilation. We found that
- Daily assimilation of SIC data is essential for SIC and SIE forecasts. This high-frequency assimilation corrects the excessive ice melt caused by warm ocean forcing, particularly at the ice edge (See Fig. 2).
- Weekly assimilation of SIT observations significantly reduced biases in SIT forecasts, highlighting the importance of assimilating more observations (See Fig. 3).
Future research should focus on optimizing assimilation methods, particularly for capturing the complex, nonlinear interactions between different sea ice variables. Possible solutions include variable-based localization, where different localization radii are used for SIC and SIT. Another avenue for improvement is modifying the state vector and observation operator to include multi-categorized sea ice properties instead of total values.
References
[1] Gascard, J. C., Zhang, J., & Rafizadeh, M. (2019). Rapid decline of Arctic sea ice volume: Causes and consequences. The Cryosphere Discussions, 2019, 1-29. https://doi.org/10.5194/tc-2019-2
[2] Soriot, C., Vancoppenolle, M., Prigent, C., Jimenez, C., & Frappart, F. (2024). Winter arctic sea ice volume decline: uncertainties reduced using passive microwave-based sea ice thickness. Scientific Reports, 14(1), 21000. https://doi.org/10.1038/s41598-024-77046-w
[3] Cheng, S., Chen, Y., Aydoğdu, A., Bertino, L., Carrassi, A., Rampal, P., & Jones, C. K. (2023). Arctic sea ice data assimilation combining an ensemble Kalman filter with a novel Lagrangian sea ice model for the winter 2019–2020. The Cryosphere, 17(4), 1735-1754. https://tc.copernicus.org/articles/17/1735/2023/
[4] Cheng, S., Aydoğdu, A., Rampal, P., Carrassi, A., & Bertino, L. (2020, December). Probabilistic forecasts of sea ice trajectories in the Arctic: impact of uncertainties in surface wind and ice cohesion. In Oceans (Vol. 1, No. 4, pp. 326-342). MDPI. https://www.mdpi.com/2673-1924/1/4/22