Overview Comments

The authors set out to explain the cold bias in ERA5 temperatures over Antarctica prior to 1979 reported by Bromwich et al. (2024) and Dalaiden et al. (2025). The uncertainty in the SST and sea ice conditions (SSC) over the Southern Ocean during this period is explored to determine the SSC contribution to the cold bias in ERA5. About 30% of the ERA5 cold bias is attributed to Southern Ocean SSC uncertainty. The desire for a reliable SSC prior to 1979 is well justified. This is a valuable study that just needs some more contextual content.

Specific Comments:

1. Line 30: It is not strictly accurate to call the period prior to 1979 pre-satellite. There were a few primitive satellite sensors in orbit back to the 1960s, as described by Hersbach et al. (2020). It is true that satellite observations became much more extensive starting in late 1978.
2. The reconstruction by Fogt et al. (2022) https://doi.org/10.1038/s41558-021-01254-9 produced a much smaller sea-ice decrease across 1979 than shown by ANT-REC. The seasonal sea ice extent reconstructions were primarily based on monthly mean pressure and temperature records across the Southern Hemisphere extratropics and midlatitudes from 1905 to 2020, similar to ANT-REC. ANT-REC does provide a much more comprehensive analysis, Fig. 5 for example.
3. Some more discussion of the results of Bromwich et al. (2024) is desirable. Their estimate of the cold bias in ERA5 was ~1C for all of Antarctica, compared to 0.7C here based on station observations. Also those authors discussed the widespread Southern Hemisphere issues with assimilation of satellite (atmospheric) observations prior to 1979 and the jump in performance of the Southern Hemisphere forecasts across 1979 (Fig. 8(e) in Soci et al. (2024) https://doi.org/10.1002/qj.4803).

Please see uploaded file.