Reply on RC1

However, only the W. Indian sector agrees with this statement. When I look at Fig. 3, see that many cores do not extend back to MIS 4. There is no information in the Scotia Sea sector about MIS 4, In the Atlantic sector, the only core that extends to MIS 4 shows that the site was south of the WSI edge in MIS 6 and MIS 4. In the Pacific Sector, two sites show similar magnitude expansions of sea ice in MIS 4 and MIS 2 (SO136-111 and the SID in TAN1302-96), whereas the WSI estimate in TAN1302-96 shows that the site was south of the WSI edge in MIS 6 and MIS 2. While, as the paper rightly points out PS58/271-1 shows highly fluctuating sea ice and was probably close to the WSI margin. I ask the authors to re-evaluate and revise their quoted statement above.

guessing it is because it doesn't add much information beyond the fact that glacial/interglacial cyclicity is the dominant forcing? It might be useful to plot PC1 along with the stacked sea ice records in Fig. 4.
The reviewer is correct in the reasoning behind us only including the PCA figures in supplemental. However we agree that PC1 will make a good addition to Figure  4.
In section 3, there is a great deal of attention paid to the differences between MIS 5 through MIS 2. However, I had a really hard time seeing the MIS 5 substages referenced in lines 203-208. Perhaps some delineation of the plots in Figs. 1-3 would aid in this (i.e., colored bars indicating glacial/interglacial stages and substages). The changes during the MIS 5 substages are significantly more subtle than the difference between MIS 5 and MIS 4. The authors also refer to "the prominence" of the substages in Fig. 2 for cores PS75/072-4 and PS1768-8. However, for PS1768-8 there is only one small increase in sea ice around 100 ka. This does not depict the prominence of MIS 5b and 5d. PS1778-is much more prominent in Fig. 2. Is there a typo?
We agree that the MIS 5 substages in PS1768-8 are not overly prominent in Figure 2 and will rephrase to refer to Figure 1 where the MIS 5 substages are more obvious for this core. We will also remove the word prominent from the sentence.
It would be helpful if the authors briefly reiterated that positive values in Fig. 3 (note, Fig.  3 axes are unlabelled) indicate times when the cores were south of the winter sea ice extent in the past.
We will restate this point to make it clearer to the reader: "For the standardised records in Figure 3, positive values indicate intervals when the core site is located south of the mean WSI edge." Lines 211-212 state, "Standardised sea-ice records also show that, despite sea-ice expansion in most regions as early as MIS 4 (Figures 1, 2 and 4), the mean WSI edge was located south of the majority of core sites in this study until MIS 2 ( Figure 3)." However, only the W. Indian sector agrees with this statement. When I look at Fig. 3, I see that many cores do not extend back to MIS 4. There is no information in the Scotia Sea sector about MIS 4, In the Atlantic sector, the only core that extends to MIS 4 shows that the site was south of the WSI edge in MIS 6 and MIS 4. In the Pacific Sector, two sites show similar magnitude expansions of sea ice in MIS 4 and MIS 2 (SO136-111 and the SID in TAN1302-96), whereas the WSI estimate in TAN1302-96 shows that the site was south of the WSI edge in MIS 6 and MIS 2. While, as the paper rightly points out PS58/271-1 shows highly fluctuating sea ice and was probably close to the WSI margin. I ask the authors to re-evaluate and revise their quoted statement above.
We will amend the text to clarify that there are relatively few records that extend back to MIS 4. We will also specify the different patterns for the individual sectors, as mentioned by the reviewer:

"Most regions experience sea-ice expansion as early as MIS 4 (Figures 1, 2 and 4) but identifying the position of the mean WSI edge prior to MIS 4 is complicated by the scarcity of standardised records for this period (Figure 3). In the West Indian sector the mean WSI edge was located south of all the core sites in this study until MIS 2 (Figure 3). In the Pacific sector the SO136-111 and TAN1302-96 core sites were located north of the mean WSI edge during the 130-70 ka interval (MIS 5), with the standardised record for TAN1302-96 indicating it was north of the mean WSI edge during the 130-25 ka interval (MIS 5 through MIS 3, inclusive). In contrast, the standardised sea-ice record for core PS58/271-1…"
Line 225: This sentence is confusing because the authors say that cores from the western Pacific sector show little sea ice signal outside of MIS 2-4, but the western Pacific cores (PS75/072-4; SO136-111; TAN1302-96) are the ones that earlier were described as having variability in the MIS 5 substages. It's unclear whether this is a typo and the authors meant the Scotia Sea or if there's something about these western Pacific cores that is not obvious here.
The earlier reference will be amended to remove cores SO136-111 and TAN1302-96 from the discussion of MIS 5 substages.
In line 241, the authors estimate how much farther north the WSI edge was in the Scotia Sea during MIS 2. It took me a while to parse that this was probably estimated because there is a lot of sea ice in the cores during MIS 2 and they are currently up to 3.6 degrees north of WSI. It would be helpful if a line or two describing this logic could be included in the text. While on the topic of the Scotia Sea, I also wonder if the sediments below iceberg alley artificially show increased sea ice because sea ice diatoms are transported by the flux of ice bergs?
We will add additional text to clarify the logic behind our estimation: Figures 1 and 3

Short Line by Line Comments:
Lines 80-81: I was confused by this, and initially thought that the sites north of the WSIE also should be in abyssal depths. I looked up a bathymetric map of Antarctica and was surprised to learn that there is a network of ridges north of Antarctica, roughly in the same location as the WSIE boundary. I suggest mentioning the water depth of cores in lines 76-79 somewhere (and potentially in Table 1?) to underscore that the cores mostly come from shallower depths. It's not necessary for this paper, but if you're interested in a modern analysis of sea ice extent and water depth, Nghiem et al., 2016; doi:10.1016/j.rse.2016.04.005 is quite interesting. I found this paper while deciphering Lines 80-81.
We will add a comment on the average water depth for the core records presented.
Lines 82-83: The authors write, "dissolution of the more lightly silicified diatom species (generally sea-ice related species) increases, which biases the preserved diatom assemblage to reflect warmer and lower sea-ice conditions." This is a commonly written idea, but in my experience, finding actual data to back this up is challenging. I urge the authors to find a reference to support this idea of biased diatom assemblages (i.e. sea ice diatoms selectively dissolved) and/or increased dissolution of diatoms in general.

Leventer (1998) and Warnock et al. (2015) both discuss the dissolution of diatoms causing biased reconstructions and will be referenced in the manuscript in support of the sentence identified by the reviewer.
Line 147: Journal requirements differ, but as a reader, I would appreciate to be reminded of what the acronyms FCC, WSIC, SID, APF, and WSI stand for in this figure caption.
We will spell out the acronyms in the figure captions.
Line 217: It's unclear which cores are in the output region of the Weddell Sea Gyre. Is it the Atlantic Sector cores? Please include a notation.
We will add a note with which cores are in the output region of the Weddell Sea Gyre.
Lines 296-305: I suggest you spell out iron throughout this paragraph.
We will refer to it as iron throughout the paragraph.
Lines 314-320: This paragraph relies fully on non-peer-reviewed papers. If the Green et al., 2021 and Chadwick, et al submitted papers are In Press by the time this manuscript is submitted, this paragraph is fine (in fact it's great). But I just wanted to highlight it in case they're not.

Figures 1-3:
Please label major places referred to in the test including (but not limited to) Ross Sea, Weddell Sea, Scotia Sea…it would also be helpful to include very basic current patterns since you mention output regions repeatedly. It is very difficult to discern the difference between the grey and black lines (September sea ice vs Antarctic Polar Front). Perhaps the grey line could be lighter?
We will add labels of the Ross, Weddell and Scotia Seas and the general gyre patterns to Figure 1. We will also make the difference between the grey and black lines more distinct. Figs. 1-3, however the numbers change from figure to figure. I realize that this is because the number of records decreases, however, it makes it confusing to compare between figure. I suggest the authors number the records in Fig. 1 and leave the numbers consistent even though it means that Fig. 3 will not have record numbers 1, 3, 5, 6…

Each of the records is numbered in
We will keep the numbers consistent with Figure 1 throughout.
In the paper you refer to studies as reconstructing winter sea ice concentration and abbreviate it WSIC, but in the figures you call it WSI (%). Please make these consistent.
We will change the figure legends to WSIC (%).

Figures 2 and 3
do not have the x axis defined. Is it the same for every core? It should be. Is it +/-1? Even though it's normalized, it should still be labelled. Actually, in the legend, it says that the axes are still WSI (%), SID (months/year), etc. But, if this data is normalized, shouldn't it be unitless?
The x axis is consistent between all the records and hasn't been labelled because the values themselves are not relevant to the interpretation. The values should be unitless and we will make this change. We will add labels for Terminations I and II.

Supplementary Material
Figures S5 and S6 are not referred to in the text. Please either include a discussion of them or remove them from the supplemental material.