Review by Susan Conway
The manuscript contains novel results pertaining to alcoves in the glaciated mid-latitudes of Mars that the authors argue are carved by glacial erosion akin to cirques on Earth. The contribution is worthy of publication, but the paper (and most notably, the discussion) needs to be shortened and the methods (and some of the results) expanded to better represent their work. Only the conclusions that are robustly supported by the author’s collected data should be presented and the paper shortened via that mechanism. I have included detailed comments on the attached PDF and reproduced below those that require a response and either a change to the manuscript or a reason why the manuscript has not been changed (I note from the previous review that the authors made many replies to the review and yet did not necessarily change the manuscript – the reviewer’s question about map projection of Fig 3 is an example, where a response was provided as a reply, but the information should also have been added to the text, so I have had to raise a similar comment in this review).
In general, the paper reads as if the authors were arguing strongly that these cirque-like alcoves were formed by the erosion of wet-based glaciers in the first version and after a round of reviews they were forced to acknowledge that their data do not allow them to conclude this (I wrote this text before looking at the previous reviews). So even though the paper does acknowledge that these cirque-like alcoves could be formed by the erosion of cold-based glaciers, it is somewhat begrudging. I encourage them to take a fresh look at the paper and try to “clean it up” by focussing on the conclusions that are best supported by the data collection effort they have undertaken. Briefly these would be in my opinion: that the alcoves they investigate are likely caused by significant glacial erosion, they are bigger than cirques on Earth, their orientation/size/distribution shows climate signal also seen in gullies and GLF globally, the timescales for formation under wet and cold based glaciation are realistic, but different (with different implications). An effort should be made to shorten the text substantially. Hence, I feel that the main concern in the previous round of reviews “My main concern is that from the very start of the manuscript, the wet-based model is accepted as an assumption” has not been fully addressed.
In summary, the main points that need addressing before publication are:
1. The discussion is very speculative and not focussed so the reader gets lost as to how the authors’ work even pertains to what is being discussed. I think the authors should focus their paper to present 4-5 solid conclusions which have a direct link to the data they have collected and remove the other conclusions with associated discussion to make the paper easier to understand and read.
2. That cirques on Mars indicate wet-based glaciation is going too far. Even that these alcoves are probably cirques is already enough of a finding without the authors needing to go further. The wet-based glaciation can appear where the authors consider rates of erosion and timescales, but should not form part of the abstract, the introduction and conclusions. The other parts of the discussion dealing with this are speculative and unfounded, so should be reduced/removed. See my detailed comments. This was the main concern of one of the previous reviewers.
3. The comparison in the discussion to rock glaciers (and its recurrence in the abstract and conclusions) is speculative and no comparison data from Earth are directly presented to support this point. Personally, I do not see the resemblance, yet I should not need to make a personal judgement if this is a conclusion of the paper, I should have the data presented to me and be convinced by the authors’ arguments. This is not the case – please see my detailed comments – so I suggest removing this comparison as it distracts from the more robust conclusions in the paper.
4. The methods need some clarification so the reader fully understands the data and methods used. Namely:
a) Demonstrate the difference to other alcove forming processes on Earth up front by integrating Table 6 on page 15 and removing section 5.6. This provides additional justification for the down-selection using the alcove morphometrics and frees up space in the discussion.
b) Clarify the data used to make the initial alcove classification (e.g., simple, joined, staircase, etc). Longitudinal profiles are included in Figure 4 which presents the classification, yet it is not explicitly stated that they have been used to inform the classification and what attributes of them were used. If they were indeed used then section 3.2.3 which speaks to the effect on the long profiles in the uncertainty of the elevation data used, makes more sense. The authors should make sure to update this section to incorporate the effect of the elevation uncertainty on each attribute they list in the methods as being critical for the classification, as well as comment c below. A similar point was raised in the previous review, but the manuscript not changed in response.
c) when considering the uncertainties in the elevation data in section 3.2.3 please address how these may also affect the ACME data collection, specifically consider the noise in the HRSC product (clearly visible as step-artefacts on Figs 4 and 6), and how well the CTX and HRSC data were co-registered. Noise is accentuated in topographic derivatives such as slope, which is amongst the parameters extracted. Presumably the position of the long profile was determined based on the CTX image data (if this or is not the case then it should be described in the methods as mentioned in point b), hence co-registration is critical to have reliable and representative elevation data. Please state what projection system was used for the morphometric analyses and consider whether this introduced any uncertainties/distortion (including the slope calculation from the HRSC DTM).
d) How the different ice-related-morphologies were recognised in the HiRISE images should be explained in the methods, with references to support their ice-related-origin. The results of this work should be in the results and then this frees up space in the discussion.
5. Further, the inclusion of all the alcoves types in the methods and in the first part of the results makes the paper bulky and are unnecessary as these results are not used to support the main conclusions. I strongly suggest omitting them. In the methods it can simply be stated that “alcoves that show any of the following morphologies [bulleted list of properties of joined, staircase, channel, etc], were not included in our database”. I understand that work was done by the authors to map these landforms, but this is not a masters’ thesis where one has to demonstrate how much work was done, and so I do not think this is adequate justification to include them in the paper.
6. The objective distinction between what was previously mapped as GFL and the alcoves mapped by the authors is not clear to me. It seems that many of the alcoves mapped by the authors contain GLF missed in this previous survey and as demonstrated in the HiRISE survey many of the alcoves contain deposits that have one or more ice-related morphologies (these need to be tabulated somewhere, as noted in my detailed comments) which could be the extension of the VFF up into the alcove (I am talking about the visible extension of VFF as can be seen on images and not mapped outlines, which are never totally reliable as they are usually made using low resolution image bases suitable for global studies – not meant to be looked at “in detail”). Global surveys are often incomplete so this statement is not a criticism of the previous work. I think the back and forth discussion when comparing the GLF and cirque-like alcove distribution would become clearer if these landforms were treated as a continuum. This is a similar comment to that raised in previous review point 3 and was not addressed by the authors by a change to the manuscript.
Detailed comments (please refer to PDF for placement as no line numbers were included)
Page2:
***please include references for each of these types of ice, especially because subsurface ice captures debris covered ice, so the distinction that is trying to be made is not clear
***it is not clear for a general reader why this is "in addition" to the previous sentences that point to evidence of wet-based glaciation, so make it clearer this is also being used to make that case or remove
***suggest being more specific and saying "where ground penetrating radar data are exploitable"
***this is vague, be more specific
Page3
***not a good citation for the LDM and not in reference list
***this is vague, be more specific
***not in reference list
***not good refs for the age, better: Schon et al PSS 10.1016/j.pss.2012.03.015 Willmes et al PSS 10.1016/j.pss.2011.08.006
Page 4
***this text should appear after the first sentence as it applies to both panels a and c.
***place with text describing legend items, above
***missing space
***This section is not convincing. There is no good evidence presented that cold based glaciation cannot create cirques. We do not observe them on Earth because currently cold based glaciers hide them and any currently exposed cirques have experienced warm-based conditions at some point confusing the signal. On Mars there is a lot of time to do geomorphic work because of the lack of plate tectonics and active hydrosphere/biosphere, so "it takes too long" is not a good reason to throw out cold based glaciation. I think simply ignoring the uncertainty is dishonest to the reader. It also sets a precedent for future works to use cirques as "proof" of wet-based glaciation on Mars or other planets.
I don't disagree that terrestrial cirques are generally associated with wet-based glaciation, but this does not prove that cold-based glaciers cannot make them. In order to use these as evidence of wet-based glaciation on Mars or even suggestive of wet-based glaciation on Mars there needs to be solid proof that water is needed to form cirques on Earth, which to my knowledge does not currently exist in the literature.
I think all this paragraph should be in the discussion and not the introduction
Page 5
***it would be useful to mention briefly what processes contribute to cirque growth/formation, see the nice summary in the intro of this paper: https://journals.openedition.org/geomorphologie/13057
NB: this paper also highlights that not everyone thinks that cirques are principally glacial, I am not saying the weight of evidence is on their side (e.g. Evans ESurf 2020), I am just saying it is better to acknowledge that cirque origin is not a completely "solved problem"
***move to discussion
Page 6
***the colour-keyed DEM is not very useful to show the form, I suggest using the air image (as this approximates the CTX most closely) with contour lines. Or use the hillshaded+colourised DTM and another panel to show the actual image of the cirque
Page 8
***all panels have a north arrow, just delete this text
***Russia
Page 9
***please add longitude labels on bottom axis because the longitude lines do not run straight up-down so the top labels cannot easily be linked to the bottom
***can be deleted, redundant with legend
***white rectangles are extremely hard to see, make more visible
***please state the version of the CTX mosaic that was used
***please state how this mosaicing was done and how the alignment between HRSC and CTX was managed (or not) and if it was not please say how much the mismatch was and therefore the inaccuracy on the placement of your profiles. to assess the mismatch it is easiest to use the orthorectified ND4 image and the CTX mosaic. Note the CTX mosaic is not controlled, so is unlikely to align properly with the HRSC which is controlled at level 4.
Page 10
***please list the images used in your data availability statement
***these classes were determined using the image data only? if so explicitly say so. If not say what other data was used and how.
***In the figure the longitudinal profiles are included, but in the table the longitudinal profile characteristics are not cited.
Page 12
***please state what the i and ii panels mean before getting into the descriptions
***there is no b-ii in my version
***ridges not visible in elevation data
Page 13
***colourised elevation
***elevation values
***make clear if this is in in addition to the number above it, or subsampled from it (perhaps in the caption as *)
Page 15
***this is the first time craters are mentioned as possible origins for these features. However, craters on the lip of the slope are highly unlikely to have the same morphology as craters on the plains, so this comparison is invalid and does not rule out that these are craters. This perhaps should be something in the discussion, at least should be mentioned whenever this comparison is brought up
***specify you downsampled from (1266 - 81) which were considered similar to cirques based on image analysis only
Page 16
***Fig 5 really doesn't show this
***simply state "by using standard circular statistics calculation methods"?
***Fig 5 the "H" line is misleading in planview, use a different line colour or use points.
include the contour lines and labels so it can be seen that "H" is the difference between the max and min height, and the min is not necessarily at the point where L starts, i.e. the midpoint
Page 17
***Figure 6 seems to show a lot more uncertainties than discussed here. Please add discussion on the noise of the HRSC DTM and the resolving power as demonstrated with the comparison with CTX (I note here that you must make sure these two datasets are spatially aligned to make this comparison valid)
***the longitudinal profile is not used to decide on which alcoves are included in the study, as far as I understand from the text (which if I am mistaken please correct the text), but the elevation data are, so please instead discuss the effects of the elevation data uncertainty on the measurements made by ACME2, which are used to downselect the data further
***this should be recorded as an attribute and presented as a result.
Page 18
***save interpretation for after the results are presented
***given the steps in this profile that clearly do not exist in the images, it casts doubt on the "overdeepening" being a real signal or an artefact like the steps.
Page 19
***briefly say what methods were used and how it was controlled to the HRSC data in sect 3.2.3
***did you do circular statistics to calculate this? in essence this means the alcoves are in all orientations? Aspect would be better shown as a rose diagram or histogram
***15°
***this would have to be done using circular statistics
Page 21
***explain the symbology, i.e. what is the blue box, the organe line, what are the circles, what are the whiskers...
***a first result should really be a map of where they are , i.e. fig 11
***are located at
***throughout the first paragraph of section 4.2 (and in the remaining sections) make sure that the term “alcoves” is not used to mean cirque-like alcoves because otherwise the reader becomes confused as to which group of data is being discussed.
***please mention how this was calculated (in methods if it takes more than one sentence to explain, then refer to methods from here)
Page 24
***Fig 11 is not super-easy to understand without the topography/image mosaic for context. It also needs to be a lot bigger (fill page width)
Page 25
***stacked bar charts are really hard to interpret, put bars side-by-side?
***this section contains a mix of methods, observations, results and interpretation, please add a methods section to describe the morphologies, then results to summarise them, and then this section can be limited to interpretation
***In general, this section is long and hard to follow. There is a lot of speculation and discussion about each landform and it is hard to follow what the authors are arguing for. Each landform has many possible interpretations and discussing every one in turn in detail makes it really hard to understand why the authors are even focussing on these “details”. I strongly suggest cutting back this text and organising it differently. If the only point that is trying to be made in this section is that the 9% of cirque-like alcoves with HiRISE coverage a large percentage (how many total is not clear) of the alcoves are filled with some kind of icy materials then this can be a lot shorter (which I interpret is the main message).
A section should be added to methods to describe how the following features were recognised and citing previous literature to say BRIEFLY how they are known to be icy materials:
1. crevasses/washboard, 2. Lineations 3. polygons etc
Then a section in the results saying the percent of the alcoves that have one or more of these and then the separate percentages (with table containing the alcove ID, HiRISE Id, lat/long and features identified).
Then perhaps simply a line or two in the discussion about how these observations can be extended to the other filled alcoves in the CTX survey (these filled alcoves need to be identified and reported clearly in the results, which is not currently the case). The details on “possible maybe perhaps” till deposits, glacial dynamics, rock glacier deflation etc etc should be left out and focus maintained on providing evidence pointing to ice as a major component of the fill material.
*** so the definitions of these should be in the methods and then you can report this part as results
*** coverage of what we interpet as remnant or active ice
*** replace with "some"
*** unfounded speculation, when you restructure this section such unfounded statements should be deleted and replaced with augmented statements.
*** name them or don't speculate on them here
Page 26
*** if this is all the only point that you want to make with this observation, then delete all the following parts, these are unnecessary details
*** so instead of doing this, just use ONE of the terms and stick to it please
*** said like the reader already knows what you mean, describe these before you discuss their relation to other features
*** so the interpretation is that these features are related to ground ice? If so say so explicitly
Page 27
*** what is shown in 13e looks like 13ca and doesn't seem to match the description made in this paragraph
*** this is going a little too far I think without showing examples of this morphology. in order to not overload this manuscript I suggest just removing this comparison
*** 13e looks nothing like the arcuate ridges described in this paper, so either you are referring to the wrong image or something else is wrong here
*** this seems to more accurately reflect fig 13e
*** this is too much of a conceptual jump
*** unfounded speculation
Page 29
*** missing space between words
Page 30
*** it is impossible to know if what you interpret as "less developed" landforms is how the "more developed ones" looked without a time machine or at least a good knowledge of the process(es) and their rates. I suggest complete removal of this section, it is complete speculation
*** missing word?
Page 33
*** this section seems like it should be first in the discussion
*** I did not read any "strong" evidence for glacial erosion in section 5.1, perhaps you meant another section (but I cannot find it)? section 5.1 demonstrates the ice fill of some of the alcoves and speculates based on selected images on the possible development of the alcoves from gullies but doesn't specifically assess the likelihood of glacial erosion processes (simply assumes it must be a given). Rephrase
*** I think the comparison between alcoves and GLF orientation is somewhat misleading. You show in your HiRISE study that many alcoves are filled with icy materials that are not mapped as GLF, this suggests they are perhaps occupied by glaciers that for some reason did not get classified as a GLF (no good images there when Colin did his survey? too small? don't have a clear divide between the alcove-fill and the VFF below?)... so I suggest rewriting this section with this caveat in mind
*** See Kreslavsky PSS 2008 for a nice diagram of this effect
https://doi.org/10.1016/j.pss.2006.02.010
*** are
*** pretty sure this should be 2019 - did you mean this one - doi.org/10.1144/SP467.3
Page 35
*** 2023b
*** these two papers do not talk about meltwater
*** corresponds
*** I can't make this make sense, I have read it three times
*** delete? totally unconnected with the rest of the section
Page 37
*** section 5.2.2 would be a lot shorter if this is the explanation the authors prefer
*** I don't think it was demonstrated that these icy units are more or less remnant than the GLF or indeed the VFF in the area, so either make sure this demonstration takes place or rephrase this statement.
*** why? they are morphologically indistinct from the GLF and VFF in the area which are thought to be debris covered glaciers? They do no resemble rock glaciers in my opinion and you do not present any comparisons to defend this interpretation, so I strongly advise not including it, it just distracts from your data
*** such hummocky structures occur when there is melting, which is probably not the case on Mars, also why would the upper part of VFF where the alcoves are have dead ice? This seems a nonsensical argument.
*** the timescale of formation cannot discriminate between cold and wet based because there is no constraint on the age of the alcoves nor on the time taken to form them. delete this sentence
Page 39
*** their initiation would predate, but they would develop during occupation by the LDA? Rephrase?
*** I don't see how this argues against the alcoves forming in concert with the LDA - debris from them should superpose the LDA - in cirques on Earth it is rare that the headwall is always completely covered, cirques form in alpine glaciations and not plateau glaciation
*** or it is the process contributing to the cirque development - hence a short description of the processes leading to cirque development on Earth in the introduction would be useful
*** not sure what you mean by this, but the VFF certainly extend into the alcoves as shown by your HiRISE observations
*** this doesn't seem unreasonable
In general, the part above reads as if the authors want these alcoves to have formed alongside GLF in a wet-based regime, but every argument they make has an equally convincing counter-argument that they were forced to add in previous reviews, but they are still trying to argue for the wet-based.
Page 40
*** this section should be removed, it just goes in circles and does not rely on the data in this paper. mostly it is speculation with no firm conclusion at the end of it all
*** this seems to mildly contradict what is said in the previous section, which concludes both wet or cold based glaciation are potential candidates over different, yet realistic timescales
*** this seems to lead nowhere, so I suggest remove it
Page 41
*** they don't need to keep pace with talus production on Earth, but on Mars, where the rate is likely (a lot) slower as there is not rapid frost-shattering
*** strong wording given the noise in the DTM, rephrase or remove
*** hard to say require when it totally unknown what processes are going on
Page 42
*** I don't think you need a massive discussion about processes that you can reject based on morphology/morphometry without much effort. I suggest reducing the text to an absolute minimum and simply refer to Table 6 to say these features are unlike those formed by other alcove-forming processes on Earth and most like cirques - this compilation should be presented when you present your selection criteria, page 15 and does not belong here
*** these are conditioned on melt and highly unlikely to occur on Mars, also their scale is not big enough and they only affect the active layer and cannot erode into bedrock. I would discourage this kind of hasty comparison based on "looks-like-must-be"
*** if you consider these as "similar" then you have a lot of other features to describe, including rock avalanches, first order fluvial catchments in badlands, debris flow headscarps...
*** already this point is debated where there is no ice on Mars, so why open a can of worms? your data don't have anything to contribute here
Page 45
*** this cannot be observed as you do not have a time machine, rephrase. I also do not think this conclusion is justified, and should be removed, along with the associated section.
*** also unjustified, suggest remove
*** remove, not justified
Page 46
*** merge with previous paragraph and shorten both to make a shorter snappier conclusion point
***remove
*** these are extremely tenuous arguments and this paper does not show that this is the case, remove |
Please see attached document.