Thank you for your author response comments. I invite you to submit your revised manuscript.

Dear Tiantian Feng, Tong Hao, and co-authors,

Thank you for your revised manuscript. I have read it in detail alongside the reviews and your earlier responses to the reviews. Although the manuscript has improved slightly from its first version, there is still much more progress needed before it could be considered for final publication. I would need to see more substantial changes made to the manuscript before I would be able to send it back to the reviewers to evaluate again. Below, I detail the improvements that would need to be made. All line numbers refer to the differenced ("Authors' Tracked Changes" or ATC) document.


Major areas where improvement is needed:

1. Formal separation of Methods, Results, and Discussion into sections.
The manuscript as written does not adequately separate its ideas into the appropriate scientific categories of Methods, Results, and Discussion. A substantial amount of text in the Results section actually describes methods and should be moved to there; similarly, there needs to be a Discussion section of its own that contains contextualizing, hypothesizing, and/or speculation about the meaning of the results within what is already known about Antarctic subglacial lakes (in the literature).

L237-242: This text should be moved to the Methods section

L263-274: This text should be moved to the Discussion section

L285-291: This text should be moved to the Methods section

L491-495: This text should be moved to the Methods section (where the simplified volume change calculation is described) or the Discussion (where the implications of this simplicity should be discussed). In the Discussion, the bias should be addressed - do you think the method overestimates or underestimates the volume of the lake, and does the magnitude of the bias depend on the size of the lake? If so, how does that affect the volume change calculations presented? See Point #4 below.


2. Better sophistication of the discussion.

The authors have added text in response to both reviewers' assessments that a discussion / interpretation of the results was missing. However, the text added is inadequate. For instance, one major point that is made at least twice in this revision is that "under the influence of basal topographic features, [the lakes] form a complex hydrological connectivity network" (L390-391). This is already pretty well assumed and understood by the community, and as posed here is too vague to be of any contribution. The authors need to do some thinking about what their research has revealed that was not previously known, and go forward from there to present these results toward possible implications. For instance, what is the significance that Lake Rec1 is actually two lakes that fill and drain on independent schedules? Is there any significance that both these lakes are located exactly upstream/downstream of one another on the same higher-order stream shown on Figure 9? These are just some examples; I do not mean to suggest that these are the only two questions that need to be probed, but rather I pose them as ways you might put your specific results into a broader context about our understanding of subglacial water flow. This is crucial if the manuscript is to be accepted for publication.

I found the response to Reviewer 2's first suggestion ("Why did the draining-refilling pattern change significantly after January 2022 (the end of Period III according to Figure 9c)? Why did the water stop migrating from the upstream lakes to the downstream lakes?") to be unconvincing. If, as the authors say, "active subglacial lakes reach drainage thresholds at lower volumes over time, possibly stemming from channel erosion and expansion," then what allowed the onset of Period III (May 2021 in Figure 9c)? At this time, the lakes appear to enter a period of larger (not smaller) capacity, implying higher thresholds for drainage.

I did not find a response to another of Reviewer 2's suggestions for discussion, "What is the core reason that some lake outlines need to be updated? Is it because the existing records are inaccurate, or the lake extent naturally changed (due to different water volume, perhaps) over time?" This is a natural and important question to address in the Discussion.

L522-523: "the drainage regime in the RIS is characterized by long-term, sustained water transport rather than episodic outburst floods." Is this consistent with previous findings of the Recovery Lakes, and/or with previous findings of other lakes (e.g. in the Siple Coast Ice Streams)? This is another opportunity to put your results in the context of the literature.


3. Addition of implications to the Abstract and Conclusions.
As mentioned above, both reviewers requested additional text of a "discussion" or "broader implications" nature. When these ideas are better developed, they need to be added to the Abstract and the Conclusions, as well as to the main text.


4. Uncertainty analysis.
A central component of this research is the volume change over time of each individual lake and of the RIS system as a whole. Figure 9 shows that the periodic recharge and discharge of each lake are similarly sized; such "differences of large number"-type problems in science require uncertainty analysis. Specifically, the uncertainties on the discharge / recharge volume rates are needed, as this would let us know if the net volume change is significantly different from zero, within measurement errors. For instance, on L467 and L481, volume change numbers are presented, but without uncertainty. Similarly, on L474, the finding is reported to be "significantly higher than discharge of the previous period"; however, without quantified uncertainties, it is impossible to say whether the change is significant or not. The reviewers, and now the modified text, suggest that the coarse "constant area" approximation for water volume used here is not accurate compared to the viscous changes described by Stubblefield et al. (2021). What remains is to quantify the magnitude and (if possible) the sign of the bias, and incorporate this uncertainty into the volume change numbers reported.


5. Explanation of the gap between lakes.
L248-251: This sentence was added to address a question by Reviewer 2 on why the gap between the lakes, as defined by the authors, seems to vary (e.g. Figure 4). This sentence does not address that concern. As pointed out by Reviewer 2, one can see in Figure 4 that the gap between gray zones (Rect1-1 and Rect-2) varies across the three panels. I suspect that this stems from the slightly different placements of the three ICESat-2 tracks analyzed in the figure - with close study of Fig. 4a, one can see that the red track (Fig. 4b) crosses the lakes at their closest approach, while the blue track (Fig. 4d) is at a point of wider separation. This should be explained, probably in the caption of Figure 4.



Minor changes that would improve the manuscript:

L141-142: This summary of why steep areas are "prone to measurement errors" in ICESat and ICESat-2 is too simplistic and too critical of the ICESat-2 products. The original sentence mentioned the root cause, footprint positioning accuracy; this is important to keep.

Use of "left" and "right" to refer to directions within the lakes: This should be corrected to compass directions, like "north" and "south" or "northeast" and "southwest", depending on the lake.

The downstream lakes are important but make up a very small part of Figure 1. On line 280, they are referred to and it would be natural to cite a figure here (Fig. 1) although in its current form, Figure 1 does not show these lakes very well. I suggest adding a panel to Figure 1 that shows these better, perhaps without the IceBridge lines, similar to what you have done for the middle group of lakes.

Figure 5: It is difficult to distinguish the data (dots) from the visual guides (lines). I recommend making the dots larger, making the dashed lines finer, or possibly both.

Figure 5: It is not evident what components of the ICESat-2 time series are "data" versus "visual guides", as this panel has only lines (no dots). I understand that ICESat-2 samples much more densely in time than the other eras, but this needs to be explained in the caption or perhaps by pointing to Figure 8, which does represent individual observations as dots.

L366 - The decline is referred to as "continuous", but this is not accurate as there are brief periods of rise interspersed. This should be corrected to "nearly continuous". Same on L267, L318, and perhaps other places where this or similar wording is used.

Figure 9: The caption incorrectly says the histograms show elevation difference, while the axis labels say instead volume difference (km3).



If these shortcomings are rectified, I will be able to reconsider the manuscript for review and publication.

Cheers,
Kristin Poinar

Dear Tiantian Feng, Tong Hao, and co-authors,

Thank you for your revised manuscript. I have reviewed it and find that you have improved the paper. However, the paper requires further revisions before it can be sent back to review. I detail these requirements as follows.


- The abstract is too long (350 words). It needs to be shortened to 150-250 words, per the journal guidelines at https://www.the-cryosphere.net/about/manuscript_types.html

- I appreciate that you have added the values of the measurement uncertainties of the lake volumes. However, the uncertainties are too precise, reported to 0.001 km3 compared to lake volumes reported to 0.01 km3. Please revisit the calculation and report both the measurement and uncertainty to the level of significance the results hold (I suspect 0.1 to 0.01 km3).

- In the long paragraph added to Section 4.4, there is much causal speculation interspersed with the reporting of results. This should not be in Results, but Discussion.

- Section 5.1: This section is much longer than it needs to be. The first paragraph states that filling–draining cycles affect the outlines over time, in agreement with previous findings, with some lakes dynamic and others dormant over the observation period. The current manuscript takes 9 sentences to report this - which is much too long. This is true of the other four paragraphs in Section 5.1 as well.

- L 505: "Subglacial lake outlines naturally evolve over time in response to filling-draining cycles" - This statement needs citation.

- Section 5.2: Most of the first half of Section 5.2 should be moved to the Methods. In the second half, the idea that the reported volume changes systematically underestimate the true changes should be made much more clear here - I suggest use of the words "bias" or "underestimate". Finally, I can't see how the assertion that the satellite data captures the signal accurately is supported here. If errors vary from 0.53 times the "true" signal to 0.93 times, this is quite substantial and would be hard to call "accurate". Unfortunately, these large errors cast doubt on the next two sentences, which assert that there is large natural variability. I am afraid that this argument is very difficult to buy. The authors should rethink this section and rewrite it so that it more honestly presents the reliability of the findings.

- L 564: "... indicating a relatively short filling-draining cycle for lakes in the primary drainage networks." This is ungrounded - short compared to what expectation? This needs citation. Alternately, you could state the approximate interval (2-5 years) and remove the commentary that this is long or short.

- Finally, I encourage you to deposit your code into a open-access repository to increase the probability that others can build on your work. A pointer to this repository should then replace the sentence "Other code used in this study is available from the corresponding author upon reasonable request."

I hope you will take the effort to address these changes. When I receive the newly revised manuscript, I will be able to send it out for review again.

Cheers,
Kristin Poinar

Dear Tiantian Feng and co-authors,

The reviewers have now re-evaluated your manuscript. I am pleased to accept your manuscript conditional on fixing four minor issues discovered by the reviewers. I have pasted them below for your convenience. I do not think these edits will be difficult for the authors, nor do they change the science discovered and reported in any way.

Early congratulations on the acceptance of your manuscript. Once you have implemented the technical corrections, your manuscript will proceed to the typesetting team. Please work these diligent folks to finish the publication process.

Cheers,
Kristin Poinar





Comment 1: I noticed that the manuscript alternates between phrases such as “draining-filling” and “filling-draining,” as well as “hydrologic” and “hydrological.” If these terms are intended to convey different meanings, please clarify. Otherwise, I suggest using the terminology consistently throughout the manuscript.

Comment 2: Please carry out one more careful language edit to correct small grammatical and formatting issues. Below are a few examples I noticed:
Title: Ensure capitalization is consistent and aligned with the journal guidelines.
Line 93: “are” should be “is.”
Line 98: “require” should be “requires.”
Line 246: “four remains active” should be “four remain active.”
Line 254: “contrast” should be “contrasting.”
Line 262: Remove the comma before “contains transitional elevation signals that create uncertainty.”
Line 273: “See” should be “see.”
Line 669: “run of the code” should be “run the code.”

Comment 3: In the code and data availability statement, the final Zenodo link should be updated to: https://doi.org/10.5281/zenodo.18354195

Comment 4: For Figure 11, the authors state, "The shaded areas indicate the uncertainty range for each lake," but I do not see any shaded areas in the figure.