Dear authors,

Thanks for submitting your responses to three comments. Your responses show that the line of your planned revisions will improve the paper, so I would like to invite you to submit the revised manuscript for another review round. It is hard to assess the quality of the forthcoming revised paper accurately at this stage. Therefore, I would like to re-consider this paper after I receive review reports on the revised manuscript.

Kenny Matsuoka
TC/TCD Editor

Dear authors

The revised manuscript was reviewed by a refree. The refree judged that your rebuttals to retain the title and to introduce new terminology “braided river” are not valid, and suggested to change them. I echo this view and requst a full consideration on these points. I think it benefitial to have speculative statements in the papers, even if there is a set of evidence that partly do not support speculations. However, such set of eviddence in prior work should be adequately presented so that the readers can have the full knowledge base and develop a balanced view.

I would like to ask another revision to respond to the refree comments. Please provide a response letter as well as the marked manuscirpt, when you submit the revised manuscript. I will review the next manuscript by myself.

Kenny Matsuoka
TC/TCD Editor

Dear authors,

Thanks for submitting the response letter together with two versions of the manuscript for further consideration. The main difference between these two versions is whether the feature is referred as “active subglacial channel” or “subglacial river”. The authors made a point that “subglacial river” is an accepted glaciological term in the response letter so I evaluated the latter version for the further consideration.
I made a careful review for the final judgement and found a large number of points that need to be considered before the acceptance of this work in its final form. Please consider following suggestions and address questions in the next revision, and provide a marked manuscript and response letter as well. Page/line numbers refer those in the marked “200615_final_V2” manuscript.

General comments:
- Title: as I said above, the authors convinced me to keep using subglacial river. However, I don’t think that the current title captures the main point of this study. The possibility of subglacial river largely depends on a continuous topographic valley, which is hypothesized in this study. My suggestion is “Possible impacts of a 1000-km-long hypothetical subglacial valley towards Petermann Glacier in northern Greenland”. I am open for other alternatives, but please be careful to clarify that this is a sort of thought experiments and assessments of a hypothetical subglacial valley which does not appear in the recent bed topography dataset but cannot be denied because of the inadequate data coverage. Also, geographical scope/interest of this study is somewhat unclear at the beginning but gets clearer in Results. Please clarify the geographical scope of this study somewhere in Introduction so that it is clear for the readers from the beginning. Also I found both Petermann Glacier and Petermann Ice Stream in this manuscript. I am not a Greenland specialist, but I think Petermann Glacier is more common than Petermann Ice Stream. In any case, use only one of them consistently throughout the manuscript.
- Another general suggestion is to re-organize Introduction, Section 2.3 and Appendix A so that the motivation of this work and bed modifications can be better presented as a single information package. The current manuscript presents this point at three different locations, and each of them is neither comprehensive nor standalone. A possible solution is to make Introduction shorter, and change Section 2.3 to a new Section 2 “Limited knowledge of bed topography” or “Bed topography: current knowledge and hypothesis” (I’m open for other section titles) before new Section 3 “Model and methods”. In this way, the bed topography is explained before the current Section 2.2 “basal sliding and hydrology”.
- I cannot accept the majority of Discussion for the publication, because they are relevant to the main results only to a small degree, or they are speculative based on thought experiments. Please revise Discussion largely; shortening this section does not reduce the value of this paper. Rather it helps readers to get the main points of your paper and stay focused.
Line-by-line suggestions and questions:
P1L1: Re-organize the beginning of the abstract so that it does not start with a question. For general readers, it is hard to immediately understand thoughts behind this question. It is better to explain the current status of our understanding of bed topography (including its limit), and potential impact of the hypothetical valley on ice flow and mass balance of the ice sheet. After stating these points, this question can be better understood.
P1L2: This is only one place you say Petermann Glacier, and you say Petermann Ice Stream at other locations. Please sort it out.
P1L13: Is “present day active long subglacial water river” observed? I cannot catch what you mean exactly.
P1L20: “mirror like” radar reflectors are expected only when there is a subglacial lake. Subglacial water channels with a smaller spatial extent do not necessarily make “mirror like” reflectors.
P2L6-: Mention the fractional length of the mass-conservation-derived BedMachine topography to the entire hypothetical valley. E.g. “The mass conservation method was applied to the lowest one sixth of main flowline towards the Petermann Glacier, so the majority of the bed topography in this region is derived by Kriging so that its error largely depends on the location of radar data.”
P2L31: change to “a possible continuous subglacial valley”¨
P3L3: at the beginning of this paragraph, add the spatial resolution of the model (5 km). You said “as detailed above” at P5L7 but you didn’t show it actually.
P3L4: geographical model domain is uncertain. Is it better to say “….the state of the entire Greenland Ice Sheet”? (or is the simulation made only for the northern Greenland? Anyway, please clarify).
P3L7: change “final” to “last”
P3L10: Do not cite a manuscript in preparation in a way like (Greve et al. 2020). As a rule of thumb, manuscripts in preparation cannot be cited.
P3L11: change to “computed ice topography”
P3L34: local depressions are not necessarily related to subglacial lakes. If the authors refer only to subglacial lakes, then this filling can be made only if the flat part is larger than a certain extent (i.e. lake extent). I assume that the authors rather refer to any type of small, local depressions in the bed topography.
P4L3: spell out RMSD and I think modeled and observed surface flow speeds are compared both in the logarithmic scale (rather than only observed speed in the logarithmic scale).
P4L4: Remove a citation to a manuscript in preparation.
P4L9: typo? “Petermann in at an average…”
P4L10: It is said that the modification is made to make the average elevation around -400 m. However, it is unclear for me why the presented procedure gives this average depth. Probably, the authors aimed to say that the derived hypothetical bed topography gives the valley floor at about -400 m. However, as far as I can see in Fig. 2a, BedMachine gives slightly shallower valley floor (it is represented as yellow/brown, instead of distinct green for your hypothesized valley). So, I doubt that this procedure gives the bed elevation of -400 m. Did you make any other adjustments and if so why is this floor depth of -400 m chosen? Overall, I would like to see Section 2.3 improved significantly to present the bed topography modifications more clearly. Some suggestions follow.
P4L10: Does “standard topography” refer BedMachine topography? If so, please say so explicitly rather than make a new term.
P4L12: Change to “consequences of the modified bed topography”.
P4L14-16: Show the size of the valley polygon (how wide is the polygon from the valley center line? Might be 5 km, the resolution of the model??). Later at line 17, this polygon is converted to the center line. In principle, polygon (with a certain areal extent) cannot be converted to a flowline. Development of the modified bed topography is subjective by nature, which I can accept, but please be careful to explain the process as clear as possible.
P4L16: What is “a new BedMachine”? Do you refer a modified bed topography, i.e. “Valley”? Please clarify.
P4L16: If I understand correctly, the authors also used kriging (i.e. same procedure and parameters of BedMachine), which the authors criticized in Introduction. Is it correct? Then except for the saddle removing step, what do the procedures presented in the second paragraph bring in this study? Why do you need the steps, instead of simply removing rises from BedMachine?
P4L28: Re-consider the title of Section 2.4. I cannot see why it is “idealized”. This section can be probably named as “Sensitivity tests on geothermal flux” or “Possible melt water supply from NEGIS”.
P4L29: probably “whether” instead of “where” (you compare only two experiments), and add “… is routed to this valley.” Do both Control and Valley bed topography route the subglacial water to this valley? Or do you enforce such water flow in the model?
P4L33: Fig. 7b cited here, though only Figs. 1 and 2 are cited to this point. All figures need to be named in the order of citation in the paper.
P5L1: Does Fahnestock et al. infer 1.5 W/m2 at the hot spot? If so, please say so and if not please explain why it is set to 1.5W/m2.
P5L5: this section shows (1) Control, (2) ControlIS, (3) Valley, (4) ValleyIS sensitivity tests. However, at P5L32, sensitivity tests are mentioned again and cites Appendix B. Appendix B presents four more sensitivity tests, not 1-4 above, which is very confusing. Please re-organize this section and mention additional four sensitivity tests in this section. And more fundamentally, do you really need additional sensitivity tests presented in Appendix B?
P5L7: you didn’t say the resolution of the model (5 km), though you say “as detailed above” here.
P5L9: I think it clearer if you say “basal water thickness”
P5L10: Is it better to say “the simulated basal water depth is affected by the Valley bed topography not only in the Petermann catchment but also the North Greenland in general.” I think most readers understand that this paper primarily focuses on the Petermann catchment, and it is suddenly said North Greenland here.
P5L11: better to say “separated” instead of “independent”, and change “deeper basal water” to “thicker basal water”. I already pointed out above, but also change “basal water depth” to “basal water thickness”, and “deeper and uninterrupted” to “thicker and uninterrupted”. I don’t point out this “deeper vs. thicker” issues below but the same suggestion is applied for all locations.
P5L14: I don’t call a water body of 10 cm thick as subglacial lakes. It is very unclear to show the thickness as “larger than” (even 100 m is larger than > 0.01 m!). It is much clearer if you say something like “thicker water bodies of about 10-15 cm were found at 6 locations along the valley”. (Obviously I made up 15 cm and 6 locations!).
P5L15: do you mean no difference between Control and Valley experiments by “no change”? The word “change” often refers temporal changes, so please describe it in a more explicit way.
P5L20ff: It seems to me that the location of thicker water shifted eastward at 74-75N as the bed topography is modified. Is it directly associated with the shift of valley position? I think it is better to mention this in the text. What features are directly related to the hypothetical topography and what features are more “wow, it’s unexpected”?
P5L26-27: I cannot read “stuck” features from Fig. 4.
P5L29ff: It is very hard to follow these statements. I think that the authors are trying to mention the feature of a “blue-red” pair blowing from the valley to northwest in Fig. 3d. Do I understand correctly? Then why don’t you explain it when you discuss Fig. 3 a few lines above?
P6L2: Is 1.6 m3/s a simple conversion of 10,000 m2/a (check the unit)? Also why do you convert it? Are you trying to tie this water flux to some observations at NEEM?
P6L5ff: What is the point of Fig. 5? Fig. 4 shows water flow directions, which are calculated based on the subglacial hydraulic potential I believe. Then why do you want to show ice and bed topography (i.e. more indirect way to show the water flow direction) here? If you argue that Fig. 5 is necessary in addition to Fig. 4 and others, consider presenting it in a smaller size. Also, is the bed topography presented here Control or Valley? Clarify it in caption.
P6L9-L14: I cannot follow the argument here. Why can’t you conclude that the bed is sloping or flat simply from the bed topography? In what context is post glacial adjustment discussed here?
P6L15ff: I am puzzled. Fig. 2 shows no topographic modification in the Steenby Glacier area (forgive me if I call it wrongly, but I want to point a glacier east of Petermann). No topo modification and no increase of water thickness in Fig. 3c are consistent, but the significant increase of surface flow speed (Fig. 6) is inconsistent with the other two. Please explain. It seems that this point is discussed and concluded negligible (P6LL23-24), which is incomplete and not convincing. Also, as the bed topography is modified, ice thickness is modified. Then the difference in surface flow speed is a combined effect of increased (thicker) ice deformation and basal sliding. Is it more straightforward to examine ice velocity of the deepest modeled ice layer?
P6L27ff: I am afraid that Figure 7 needs to be revised to present these points clearer (or the authors read Fig. 7 too much). The most obvious thing is that basal melt rate along the valley increases only if the Valley topography is used. And regardless of the topography, the additional NEGIS hotspot increases basal melt rates towards the NEGIS and towards the west. I cannot immediately see the differential amount of basal melt towards NEGIS and west, when Control and Valley topography results are compared. Maybe the authors found this west-east contrast by examining the modeled results, but Fig. 7 is in any case inadequate to make this point. Also, this and following paragraphs in this section include discussion in addition to results. Please distinguish discussion and results more clearly and move statements to discussion as appropriately.
P6L35: Please carefully check the statements in this paragraph, and its evidence presented in the figures. I cannot agree with this statement about the region north of 80N.
P7L1-3: You discuss the water flux here, while you have discussed basal melt rate in this paragraph. Reorganize.
P7L4ff: The majority of the discussion here is irrelevant or at least not significantly connected to the results presented here. In particular the first few paragraphs explain general knowledge, which is not necessary for this paper in my opinion. I suggest to delete the first 1-3 paragraphs and the 5th and 6th paragraphs because of inadequate relevance to the results. The 4th paragraph is somewhat relevant, but it is not directly linked to specific results in the current manuscript. If you retain this paragraph in the new version, please deepen the discussion and clearly connect the discussion to the results you obtained.
P8L16ff: I am not yet convinced that this discussion and Fig. 8 are necessary in the final form of this paper. I see the importance to examine boundaries of surface/ice and subglacial/water drainage basins. The current Fig. 8 shows the subglacial/water boundary for the Valley topography (I assume so, but it is not clear in the caption). If the authors have important discussion points on this issue, please present both boundaries of surface/ice and subglacial/water drainage basins for (1) Control topography and (2) Valley topography in two panels next to each other, and discuss how (2) is different from (1). Keep it in your mind that topography modification is hypothetical and the main aim of your work is to draw an attention to improve the bed topography of northern Greenland. You don’t need to examine all small features.
P9L4ff: I think that the discussion about the link between the NEGIS hotspot and Petermann Ice Stream is valid. However, this paragraph starts with the discsussion about R-channels and so on. Please re-consider this paragraph and make sure that the discussion is directly pointed to the most important issues.
P9L17ff: I don’t really think that discussion in this and following paragraphs are directly relevant to the hypothetical subglacial valley. Nonetheless, if you really want to point out these issues, limit the topics which are directly relevant and highly important for the ice-sheet dynamics. I really think that the 4-page-long discussion in the current form can be dramatically shortened to 1 page or so. It actually helps the authors to address the most important issues directly related to the results.
P11L6ff: Please update this section according to your revisions in Discussion. The link to the NEGIS hotspot can be mentioned here, but the statements to this point (P11L16) can be shortened. I didn’t point out at other locations, but the authors occasionally call the feature as “active” subglacial river, though the word “active” was removed from the title. I don’t clearly see meaning added with this word, so suggest the authors to carefully consider what it really means. Also, the statements about wet sediment flow system and paleo-fluvial rivers are too weak to be concluded/presented here. My take-home message from this paper is that more radar surveys are necessary to examine the absence of these “removed” rises. If this is the case, please clarify this point both in Abstract and conclusions, and possibly in Introduction.
P11L9: change “are not real” to “may not be real”. The lack of evidence does not means that there are no rises.
P12L3ff: Overall revisions are necessary to the two paragraphs for Appendix B. For example, it is unclear what “linear” and “idealized” mean (L5), why 26 valleys are used for 4 additional experiments (L5), what “grid point values” of “along-valley rectangle” mean (L6), what “short SICOPOLIS simulation” (L7) and “standard SOCOPOLIS topography” (L9) mean, and what “4th test” refers (L9). There are just examples to show why I require overall revisions to Appendix B. I expect concise, but adequate descriptions in Appendix. Also, above all, do you really need Appendix B?
P.17 Figure 1: the legend says Petermann surface catchment, but I think it is a buffer of 200 km used to modify the bed topography, rather than defined from the surface topography. Also, the label shows Petermann Glacier but the text mostly refers as Petermann Ice Stream.
P.19: Figure 3: change the unit for Fig. 3a/b colorbar from meters to millimeters. Also the caption does not need to say “from SICOPOLIS simulations for the year 1990”. Also explain the label “NEEM zone” in the text. If you meant the NEEN ice core site, show the core site using a small symbol and explain it in the caption.
P.19 Figure 3c and 3d: consider adjusting the colorbar. The current figure does not support a statement in the main text about larger amount of water in the NEEM zone. The entire valley looks similarly red.
P.20 Figure 4: The labels of NEEM Zone for Figs. 3 and 4 point to different locations at a glance; adjust them so that both labels point to the right regions. Maybe adding arrows to Fig. 3 helps. Add the descriptions to the lower two panels (please clarify that these are zoom up of panels a and b). I understand that the authors chose the logarithmic scale because of the largely variable range of water flux shown here. But the river locations are shown in Fig. 3 already. Then is it actually better to show the water flux in the linear scale or at least adjust the colorbar so that the water flux increasing along the river can be better seen? I any case, clarify in the caption that the water flux colorbar is in the logarithmic scale if you stick with the logarithmic scale.
P.21 Figure 5: see the comment at P6L5ff.
P.22 Figure 6: Is it possible to show the extent of Petermann Ice Stream? The text says that the basal sliding increases significantly in the valley (> 10 m/a) but only little in the ice stream (< 0.5 m/a). It is hard to see the latter point in this figure. Maybe you can add the grounding line, or use non-white background color so that we can see near-zero difference (white) more clearly.
P.23 Figure 7: see my comments above. And revise the unit for the basal water thickness colorbar from meters to millimeters.
P.24 Figure 8: Re-consider about this figure as suggested above (see my comments to p.8). If you retain Fig. 8 nearly as is, add colorbar, and the legend for blue/orange polygons. It is not easy to distinguish this orange polygon as the lower bed is shown with similar color. In any case, I expect that you modify Fig. 8 largely or delete it from the paper.

Dear authors,

Thank you so much to submit the revised manuscript with great care. Your responses are satisfactory, and I am happy to accept this article as is. Great work!

Kenny Matsuoka
TC/TCD Editor