| 03 Aug 2026
Sea-ice concentration estimated from microwave radiances in a coupled ocean-atmosphere weather forecasting system
Abstract. Weather forecasts are increasingly made by earth system models that couple the atmosphere, ocean, sea ice and land surface. These coupled forecasts are best initialised from a coupled data assimilation system, especially if this uses satellite radiances that are simultaneously sensitive to surface and atmosphere. This is more consistent and optimal than using separate geophysical retrievals. Working towards this, microwave imager radiances sensitive to the ocean, sea ice and atmosphere have previously been used in atmospheric data assimilation, with retrieval of sea ice concentration (SIC) at observation locations as a by-product. The current work further activates the assimilation of these SIC retrievals, derived from Advanced Microwave Scanning Radiometer 2 (AMSR2), into a coupled ocean and sea ice model, using an outer loop coupled data assimilation approach. Improvements in modelled SIC are demonstrated indirectly in the marginal ice zones (MIZ) by the 12 h forecast fits to scatterometer wind and radar wave height measurements, where errors are locally reduced by up to 10 %. In the Antarctic summer, comparison is made to Ocean and Land Colour Instrument (OLCI) visible radiances and SIC estimates. These comparisons confirm the quality of the AMSR2 SIC retrievals but expose limitations in the modelled sea ice. The data assimilation also suffers spurious sea ice loss of up to 10 % in some Arctic winter locations, though effects on the atmosphere are locally contained, and sea ice losses are rapidly corrected by the forecast model. More generally, the atmospheric forecasts are improved in polar regions in the early forecast range, particularly in the winter seasons, confirming overall benefit from coupled SIC assimilation, likely due to improvements in the MIZ. Atmospheric forecast improvements propagate to midlatitudes by day 4, where wind and temperature errors are locally reduced by up to 1 %, indicating that sea ice assimilation can improve weather forecasts also outside polar regions.
This study describes the assimilation of sea ice concentration estimated from microwave radiances (AMSR2) into the ECMWF coupled forecasting system. The study analyses the effects of this assimilation in both the Arctic and Antarctic, finding indirect improvements in the sea ice concentration near the ice edge by evaluating the experiments against wind, temperature, radar wave height in the Arctic and by microwave radiances and and SIC estimates in the Antarctic. There are limitations in the sea ice model and the data assimilation causing sea ice concentration loss in pack ice but they are corrected quickly. The most interesting result shows that the sea ice data assimilation is improving atmospheric forecasts in the midlatitudes, outside the polar regions. The topic is timely and relevant to the sea ice data assimilation community, and the methods are thoroughly described, the figures are well presented and the paper is well written. The reduction of error in atmospheric forecasts outside the polar regions caused by the sea ice data assimilation is novel and interesting. However I believe parts of the methods and results need to be improved before publication.
Methods
The study uses a background error of 0.05, a horizontal correlation length of 200 km, and an observation error for SIC of 0.2. However, these choices are not sufficiently justified within the paper. It should be clarified whether these values are inherited from the sea ice data assimilation system described in Browne et al. (2026), and, if so, why these values are considered appropriate for the present application. Given that these parameters directly control the magnitude and spatial extent of the SIC increments, some justification or sensitivity analysis would strengthen the study.
The description of the observations used for evaluation could also be expanded. Figures 13 and 14 provide an evaluation using independent observations, but there is relatively little context given for these observations. I suggest adding a brief section to the Methods describing the independent observations in more detail. The data period, spatial resolution and limitations of these observations should be discussed here. Â The observations section in the methods could be split into "observations assimilated" and "observations for independent evaluation"
Results
The results of the assimilation of AMSR2-derived SIC are only qualitatively evaluated against OLCI. A quantitative independent validation against OLCI, including metrics such as bias, RMSE and correlation, or scatterplots, would strengthen the manuscript. Comparison with an additional independent SIC product would also provide a more robust assessment of the quality of the AMSR2-derived SIC and the resulting analyses.
The evaluation period is quite limited, and although the results are promising, this is a relatively short period to assess how robust the improvements in atmospheric forecast skill really are. The authors should make this limitation clearer, and the authors should clarify the extent to which the results can be generalised beyond the assessment period.
Specific comments
Line 15: The statement "likely due to MIZ improvements" is somewhat vague. It would be useful for the authors to provide a clearer explanation of the mechanism linking the paper study (sea ice assimilation in the MIZ) to the reported result, or use more cautious wording.
Line 99: How are melt ponds and melt pond lids represented in the sea ice model? There are several different approaches to representing melt ponds in sea ice models, and this could be relevant given the discussion of the AMSR2-derived SIC. The description of the sea ice model could therefore be expanded to clarify how melt ponds are represented, as well as any other relevant parameterisations.
L112: typo: "forecasting forecasting system"
L143: Some more details of the IAU procedure would be useful here, alongside a reference. I assume it is the same procedure described in Browne et al. (2026a), but it should be provided if so.
L225-226: "These are the aforementioned addition of sea ice scenes" - I don't understand what the authors mean here.
L308: closer the Canadian archipelago -> to the
Figure 8 - 14 - The figures and accompanying text mention normalised change in RMSE forecast error, normalised change in standard deviation and statistical significance at the 95% level, however it is not completely clear how these metrics are calculated in the study. It would be clearly state how these are calculated within the methods section.
L565: The two sentences " Satellite names such as Metop-C, Sentinel, Jason and Cryosat are documented at https://space.oscar.wmo.int/satellites. The coverage of in-situ platforms such as buoys and ground stations is quite limited in polar regions and hence it is no surprise that the fits to these observations do not significantly change. " seem out of place and probably belongs in the methods.Â
 L649: a the -> at the