Sahoo et al. (2026) Impact of satellite-derived sea ice data assimilation over Antarctica using a variational data assimilation system

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2026
• Date: 2026-01-10
• Authors: S. K. Sahoo, IM Momin, Anitha Gera, JP George
• DOI: 10.1007/s00704-025-06017-6

Research Groups

National Centre for Medium-Range Weather Forecasting, Ministry of Earth Sciences, Noida, India

Short Summary

This study assimilated satellite-derived sea ice concentration (SIC) over Antarctica using a variational data assimilation system (NEMOVAR) and found significant improvements in model performance, particularly in sea ice area (SIA) and its seasonal variability, by modulating surface thermodynamics and radiative properties.

Objective

• To assess the impact of assimilating satellite-derived Sea Ice Concentration (SIC) on model performance in simulating Antarctic Sea Ice (ASI) characteristics.
• To evaluate the impact of SSMIS-derived SIC assimilation over different regions of Antarctica.

Study Configuration

• Spatial Scale: Antarctic region, global 0.25 degree (¼ degree) grid resolution.
• Temporal Scale: Daily assimilation over a five-year period (2020–2024).

Methodology and Data

• Models used:
  • NEMO (Nucleus for European Modelling of the Ocean) ocean model (ORCA025 configuration, 75 vertical levels).
  • CICE (Los Alamos Sea ice model) configured on the same NEMO tripolar grid.
  • NEMOVAR (variational assimilation system) based on the multivariate, incremental three-dimensional variational DA scheme with the first guess at the appropriate time (3D-Var FGAT).
• Data sources:
  • Assimilation data: Daily satellite-derived SIC from the Special Sensor Microwave Imager/Sounder (SSMIS), obtained from the Ocean and Sea Ice Satellite Application Facility (OSI SAF) with a horizontal resolution of 25 km.
  • Validation data: Satellite-derived SIC data from the National Snow and Ice Data Center (NSIDC) (25 km × 25 km nominal grid cell size).
  • Forcing data: Surface boundary forcings from the global NCMRWF Unified Model (NCUM) analysis dataset.

Main Results

• The assimilation of SSMIS-derived SIC data significantly improved model performance, particularly with respect to SIA and its seasonal variability.
• The Root Mean Square Error (RMSE) against SSMIS observations for SIA was 0.49 × 10⁶ km² with assimilation (DA) compared to 0.67 × 10⁶ km² without assimilation (noDA) for the 2020–2024 period.
• The time evolution of SIC climatology in the DA experiment was more closely aligned with SSMIS observations than in the noDA experiment.
• Climatological differences (DA–noDA) in SIA, Ice Surface Temperature (IST), and near-Infrared (IR) albedo suggest that assimilation not only improves SIC representation but also modulates surface thermodynamics and radiative properties.
• In the DA experiment, higher SIA values were linked to a decrease in IST and reduced near-IR albedo, indicating a consistent thermodynamic adjustment.
• The percentage improvement in RMSE for the entire Antarctic region was 38%, with regional improvements of 55% (Weddell Sea), 155% (Western Pacific Ocean), 143% (Bellingshausen/Amundsen Seas), 20% (Indian Ocean), and 17.5% (Ross Sea).
• DA consistently reduced RMSE across all Antarctic regions, with the most pronounced improvements during the melting phase in February, indicating better capture of ice retreat processes.
• The monthly correlation between changes in SIA and total Absorbed Solar Flux (ASF) was approximately 0.52, and between SIA and IST was about 0.43, highlighting the thermodynamic sensitivity of the surface system to SIA variability.
• Correlations between SIA and Latent Heat Flux (LHF), and between SIA and Ice Velocity Magnitude (IVM), were relatively weak (< 0.3).

Contributions

• This study reinforces the importance of satellite-derived Sea Ice Concentration (SIC) assimilation for improving the representation of Antarctic sea ice in coupled ocean-sea ice models.
• It demonstrates the effectiveness of the NEMOVAR 3D-Var data assimilation system in enhancing the fidelity of Antarctic sea ice simulations, particularly in capturing seasonal transitions and reducing biases in marginal ice zones.
• The research highlights that SIC assimilation not only improves ice extent but also modulates surface thermodynamic and radiative properties, leading to a more consistent representation of the coupled ocean-ice system.
• It provides valuable diagnostics for refining coupled forecasting systems and validating assimilation strategies for polar climate modeling, suggesting potential gains in predictive skill for polar climate modeling.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Citation

@article{Sahoo2026Impact,
  author = {Sahoo, S. K. and Momin, IM and Gera, Anitha and George, JP},
  title = {Impact of satellite-derived sea ice data assimilation over Antarctica using a variational data assimilation system},
  journal = {Theoretical and Applied Climatology},
  year = {2026},
  doi = {10.1007/s00704-025-06017-6},
  url = {https://doi.org/10.1007/s00704-025-06017-6}
}