Summary: In this study, a regional configuration of the Barents Sea modeling system composed of ROMS and CICE 5.1.2 is used to study the impact of the assimilation of swath AMSR2 sea ice concentration data versus daily means of SIRANO sea ice concentration data. Particular focus was given to sub-regions with the 2.5km domain for the Barents and Greenland Sea. Two sets of atmospheric forcing are used to introduce ensemble spread: 1) Integrated Forecast System developed at ECMWF which provided members 1-5, and MET-AA which provided member 6. The EnKF is used as the data assimilation system in this study. Three experiments are performed: 1) Control run without DA, 2) synchronous assimilation using SIRANO data, and 3) asynchronous assimilation using AMSR2 swath data. This study found that the assimilation of the swath AMSR2 sea ice concentration led to a 10% improvement in the MAD at the end of the assimilation period and 7% improvement at the end of the 7-day forecast period.

This is a very thorough and well written paper. I only have minor comments listed below. I recommend publication.

General Comments:

Line 304: Rephrase “As CICE does not…” to something like “While CICE 5.1.2 used in this study does not differentiate between stationary ice attached to land, CICE6 includes a landfast ice parameterization (https://zenodo.org/record/7419531).

Fig. 5: Why was this particular month chosen (April 2022), with the Easter Holiday occurring mid-April? You lose data for 5 days (April 14-18) versus the typical 2 days on weekends? Since you should have the SIRANO data, I suggest you add that information for the Barents, Greenland, and Entire region to the plot.

Figure 9 caption: “Mean Absolute Difference” is defined as Mean Absolute Deviation on line 166. Please make correction.

Lines 443-450: Do you have any graphics or tables to support your (29% lower), (14.3% improvement) statements?

Review on “Assimilation of satellite swaths versus daily means of sea ice concentration in a regional coupled ocean-sea ice model”, by Marina Durán Moro, Ann Kristin Sperrevik, Thomas Lavergne, Laurent Bertino, Yvonne Gusdal, Silje Christine Iversen, and Jozef Rusin, submitted for publication in The Cryosphere.

General comments :

The paper presents experiments where sea ice concentration observations (retrievals) are assimilated in a coupled ice-ocean model. The assimilation follows the method of the Ensemble Kalman Filter (EnKF). The goal of the study is to measure the impact of assimilating individual satellite observation swaths as opposed to assimilating daily average sea ice concentration derived from the same data set. The paper is very well written.

Three runs are performed. The first is a control run without data assimilation (or free run), the second run assimilates daily average sea ice concentration every 2 days (SYN), and the third run assimilates the sea ice concentration retrievals from satellite observation swath (ASYN).

One concern I have is the use of the mean error of the ensemble members instead of the error of the ensemble mean. I am wondering about this choice, since the ensemble spread should correspond to the error of the ensemble mean, and not to the mean error of the ensemble members. Please comment.

Specific comments:

Line 144: Please provide a reference for DFS, maybe Cardinali et al. (2004)

Line 223: What is the “K-factor” ? Is the observation-error covariance matrix diagonal ? If the observation-error variances are increased by a factor 20 (R-factor = 20), what is the based sea ice concentration observation-error variance ? Is the observation-error variance homogeneous ? Some of these questions are answered later in the text in detail, but it could be nice to have general info upfront for the readers.

Line 333: The fact that the ensemble spread does not reduce to zero does not mean that the system is well tuned. It is a necessary but not a sufficient condition. For a well-tuned system, we would expect some overlap of the shadows of the model spread and the observations uncertainty in figure 6.

Technical corrections

Line 279: “are are” typo, repetition

Line 367: “with a slight larger” should be “with a slightly larger”

Line 370: SSS should be replaced with salinity since that is the term used throughout.

Figure 8: Replace SSS with salinity or define SSS = Sea Surface Salinity