The paper by Catherine Collins and co-authors presents a new rich dataset from the precious sediment samples collected at the base of the Camp Century ice core. The analyses include microCT, XRD, SEM-EDS and SEM imaging. The authors looked at the sedimentary facies in order to decipher the processes responsible for the deposition and preservation of sediments and infer some paleoenvironmental interpretation. The paper is well written and the figures have been made with great care, although several improvements can be done (see detailed comments below).

This paper refers a lot to another one submitted to the journal The Cryosphere by Bierman et al. (still a preprint at the time this review is written). Unfortunately, the reader needs to read the Bierman et al. paper to understand several things, such as the sample numbering and the stratigraphy. I think this essential information needs to be incorporated within this paper. For instance, the uppermost core section is the one on the right in the figure, which is not what one usually when presenting sediment cores. Incorporating panel B of figure 2 of the Bierman’s paper would be very helpful.

A disturbing feeling while reading the discussion section of this paper is that many arguments for the interpretations of the sediment facies are coming from previous papers. A clear distinction between the interpretations based on this new dataset, and the data presented elsewhere is needed.

My main concern is about the interpretation of the sedimentary facies. I’m not a specialist of glacial/proglacial/subglacial microfacies, but in such settings, it is critical to have an idea of position of the ice, the direction of the flows, the topography, the slopes, etc., to make a solid interpretation, which is impossible to obtain from a single sediment core. Moreover, it is very likely that the sediments recovered at the bottom of 1350 m of ice experienced some glaciotectonism. This is not discussed in the manuscript, and it should, because it is not sure that the ice cap was cold based all the time, especially if one of the interpretations suggested by the authors is that a river flowed at this site. In short, I think that the observations remain interesting but that the facies interpretation is pushed too far. It would be useful to have the input of a specialist of microsedimentological analyses of glaciogenic sediments.

For the statistical interpretation, the choice of the variables considered is not well justified; for instance, why including the depth of the sample as a variable (it could be an explanatory data).

Finally, there are many small mistakes and inaccuracies in the µCT dataset description making the paper arduous to read. There are details about the µCT data acquisition in the online public repository (congratulations for that effort), but it would be nice adding some information in the paper and/or in the repository in order to be able to reproduce the measurements (see my detailed comments below). Unfortunately, I was unable to open all the movies made on processed images. While the movies are a very nice addition, it is quite impossible to analyse these images. The addition of images that can be handled by an image analysis software would have been better in my point of view, but this is already very nice.

Detailed comments:

Line 65 : « sequence stratigraphy » has a specific meaning in sedimentology that is not related to the reality you want to describe. I suggest changing this by “the lithological succession”.

Line 166: “shrinking as sediment freezes”. I’m not a specialist of these structures, but it does not seem logical to have shrunk when ice has a 10% higher volume. Can you expand on that?

Lines 1972-173: this statement is strong and should be supported by evidence or a reference to other studies.

Lines 1978-1982: these sentences belong to the introduction, not the method section.

Method section: there is little information about the quality of the storage and the transport conditions during the four transfers of the samples that occurred through time. If there is no information, this should be stated. Also, the reader is referred to another paper for the sample handling; I suggest adding a more specific reference, for instance, what figure one needs to look at? Or to add that information or figure in the supplements.

Lines 186-187: please add some technical information about the energy used to scan the samples. The sample size, the geometry of the acquisition and if any filter was used. Was the reconstruction made using the NRecon software correcting for artefacts such as beam hardening?

Line 292: perimeter is a quantity that is not robust when small grains are measured. What is the size of the grains analysed by Fiji, in number of pixels? One needs to have at least 300 pixels to have a robust measurement. See for instance, Francus and Pirard (2004). Testing for sources of errors in quantitative image analysis in P. Francus (ed.) 2004. Image Analysis, Sediments and Paleoenvironments. Kluwer Academic Publishers, Dordrecht, The Netherlands.

The area measured in 2D slices underestimate the size of particles (same ref and many others). This should be stated somewhere in the text.

Lines 295-298: I believe this might be problematic to compare these shape parameters taken on images with different resolutions (see again the same paper motioned above). I’m not sure what the Tukey-Kramer honestly significant difference statistical test does, and maybe it is good enough, but you should discuss a possible bias.

Lines 301-313: I’m not sure to understand in what order these statistical analyses have been performed: first, first PCA on the image analysis data set on quartz grains only, and then K-means clustering on all data? Correct? Please try to clarify your text.

Lines 316-322: are these new observations, or corroborating the results of the paper cited? The wording is not clear about this.

Figure 2: this figure is good-looking, and I’m sure you spent a lot of time on it. However, it needs to be improved. First, historical images are not visible in the panel a). The figure would gain in readability if the top and the bottom of the section were indicated. A scale is missing. Adding a depth scale is also needed to help read the following lines. Panel b): What is exactly the full view? Do you mean a topogram, i.e., the equivalent of a radiograph? On what is based on the colour code? And what does it mean? These are false colour, right? So what do you gain here to transform the greyscale original images to these colourful images? If you gain something, that is correct, but if not, you should consider greyscale for the CT images. Also, I think that Climate of the Past has a policy regarding these figures to make sure that colour-blind people can read them. Finally, what exactly is showing the Particle view? Have you segmented all the denser particles from the volume and made a sum of them in one direction? How have you made this segmentation? What is the smallest size of the threshold particles?

The labels “units 1” and “unit 2”, in the 3^rd column of panel a) seems to have been inverted, making the explanations below very difficult to understand.

The vertical clear ice inclusions are not obvious. Are those the very narrow vertical lines, one on the left of the image, the other in the centre?

Line 333: please add the name of the samples, i.e.1063-6 to 1062-1, in a similar way than you did below.

Line 335: what size are the clasts?

Line 336: sample 1062-4 seems to be a better example of these ice lenses with a braided lenticular pattern.

Lines 337-339: it is not clear what the reader has to look at in figure 2a that is representing ice. Where exactly are these 2 ice-rich layers?

Lines 340-341: it seems that the topmost sample of unit 1 is 1061-D5, right? Unless you call a “section” something that corresponds to a core tube in Bierman’s paper. This is why you need to incorporate the information from the Bierman’s paper about the sample names in this paper as well.

Line 342: I count 7 samples.

Line 343: how have you obtained this density and the % ice content? Is it with µCT? If yes, this is not trivial to obtain, and you should explain how you acquired these numbers.

Lines 344-345: not all the samples display 45˚ bedding: samples 1061-D1 and 1061-D2 display horizontal contacts.

Line 345: where is sample 1060-D3 in panel a)? Maybe you mean 1061-D3? If this is the case, then panel b) labelling needs to be corrected. Actually, sample 1060-D3 does not seem to exist elsewhere.

Line 348: this sentence “The samples in this unit are 1060-C3, 1060-C2, and the lower portion of 1060-C1.” should start the paragraph. I do not see the bedding, the grading and the cryostructures in the µCT-Scan images.

Line 349: the text says here there is no bedding but the line above, there was bedding. Please review your text.

Line 351: can you better show on the picture the reticulate structure that you mention here?

Line 352: bedding is visible in 1060-C1, but cryptic in 1060-C2.

Line 353: How the ice content has been measured? (Same question as above)

Lines 354-355: how can the reader know what is the a) sample and the b) sample in Figure 2? One can guess that the b) is on the right, but please add something about this in the figure.

Lines 355-356: Authors write “Directly below the contact bedding curves from sub-horizontal, downward to nearly 90° which continues into 1060-C”, but the bedding is not visible in the picture of 1060-C2.

Section 4.1: are the results presented here from the observation of the µCT scan image only?

Figure 3: Please change the label “14A” (standing for 14 Angstrom I suppose) from the legend into something like “clay minerals”. Also, it seems that there are more amorphous minerals in samples 1060-A2 and 1059-7. Could you comment on that?

Line 383: “selected” instead of “select”, right?

Figure 4a: what the vertical axis in the area plots means? Is it depth in the section of the sample? This is not easily readable. I would only keep the histograms. Also, in general, one plots those particle size distribution using a logarithmic scale.

Figure 4b: this plot is not readable. I suggest making scatter plots with the size, this will be helpful to check if size influences the two other parameters.

Lines 424-425: it would be a good idea to repeat here the formula of the parameters, so the figure is self-standing.

Lines 434-435: I suggest adding here what group the samples belong to (for ex. A-samples,…)

Line 436: “maximum mean value” is more appropriate, same for the minimum.

Line 437: is roundness distribution unimodal? I really doubt it is, the spread is very wide when looking at the figure.

Figure 5b: is the sum of the two plots making 100%. If yes, why do you need two plots?

Lines 448-449: see my earlier comment on potential bias of the size of the particles on the shape indices.

Line 454: sorry, but Fig.5a does not allow to see the coatings. Fig5 is too small for that.

Line 470: “grains within the core”: which one?

Lines 462-471: it is a pity that there is no detailed account of all these features, to have at least a semi-quantitative view of their occurrences.

Section 4.3: how the analysed grains have been selected? Random selection is quite important to avoid biases, or maybe all visible grains were selected.

Figure 6: the figure is very nice, but the element code in panel g is not clear and the scale shows the ∞ character, I suppose instead of the µ one.

Figure 7c: the variable labels are too small.

Line 495: why have you included depth as a variable? I suppose it is the depth in the core. If it is, then I think this is biasing your statistical analysis, forcing the samples with the same depth to be similar (spatial autocorrelation). I think you should remove this variable, and redo your statistical analysis without that variable. This brings the question how the variables included in the analysis have been selected? Can you expand on that?

Line 506: in the introduction, you write that the sediment core is made of several units, but here you assess this unit assignment. From this, the reader understands that the units have been previously defined, and I suppose this was made in previous papers. The authors should clearly distinguish the interpretation derived from the dataset presented here from the other proxy not presented here. Also, could you suggest a change in the unit assignments using your dataset only?

Line 530: I suggest adding “subsequent” here : (…) before subsequent cooling (…)

Line 535: which cryostructure are you talking about?

Line 536: this is the first time you mention that slumping is occurring. You should first demonstrate that the sediment is indeed a slump.

Lines 538-539: slumps can also produce debris flows, which are sediment facies that are not well sorted. If there is a normal grading, it is more likely that a turbiditic current occurred, implying that the environment is not compatible with sediments flowing downslope. Many questions come to mind here: for instance, was the site in an aerial or aqueous environment?

Lines 535-544: this interpretation for unit 3 is very hypothetical. I don’t think that these inferences can be made out of a single core. One needs to have observations about what is happening laterally.

Line 556: sediment content: what do you refer to here? Grain size? The horizontal alignment of grains is present only in a few samples.

Line 562: fluvial sediments are usually better sorted that these.

Line 565: “Multiple lines of evidence”: please specify which ones. The ones from this dataset or from other papers? Your dataset is not very convincing that this unit has been created by a river.

Line 569: is there information about the topography under the ice cap?

Line 591: this is counter-intuitive: glacial tills are usually coarser than fluvial sediments.

Figure 8: the font size in the boxes are too small.

Collins et al. (2024) present a comprehensive multiscale analysis of the sub-glacial core collected from Camp Century, Greenland, analyzing physical, geochemical, and cryostratigraphic characteristics. The authors use X-ray diffraction (XRD), micro-computed tomography (µCT), and scanning electron microscopy (SEM), to identify and characterize five distinct sedimentary units, each with characteristic structures, mineralogical compositions, and cryostructures, and then use these data to infer a sequence of depositional environments, spanning glacial to interglacial to glacial. Their findings add to a growing body of work on this subglacial material and provide further context into Greenland’s paleoclimate history and ice sheet dynamics, though some of their interpretations remain speculative without further supporting evidence.

Overall, this paper is well-written and polished, and the data and generally analyses are robust. However, my primary critique of this work is that the authors' interpretations occasionally overreach given the data presented and some aspects of the presentation lack sufficient detail to clearly follow their logic. While the authors’ interpretations are plausible, it appears they could represent one possible scenario among several. This is particularly uncertain when inferring the timing of cryostructure formation or depositional environments from qualitative observations. I believe the paper provides a solid empirical foundation and presents a valuable dataset but lacks specific evidence for some interglacial interpretations.

I describe my main concerns with the interpretation of the sediment facies below and then follow up with minor concerns.

“The higher ice content and the difference in cryostructures between Units1 and 3 supports the hypothesis that Unit 3 was deposited in or by liquid water.”

The specific differences in cryostructures are not clearly presented in this manuscript. The paper lacks plots or tables that detail ice content by depth, as well as quantitative descriptions of ice lens characteristics such as shape, spacing, orientation, and thickness. These details could provide valuable clues to the origin of the ice lenses and would allow for comparisons with theoretical predictions. Including a figure (in the supplemental material if necessary) that details and compares the various cryostructures within each unit in high resolution would greatly enhance clarity, as these structures are frequently referenced as evidence for interpreting the origins of each unit. Currently, it is challenging to contextualize the manuscript’s claims alongside the physical evidence. The supplemental movies provide a helpful visual, but lack sufficient detail to quickly orient the reader. 

(As a note, I was unable to view the CT scan videos using an M1 Mac on either Safari or Firefox and ultimately had to download Chrome, which played them successfully.)

Regarding the Interpretation for Unit 2

The authors suggest that Unit 2 is a ~1-m-thick remnant of basal ice (or even firn) preserved from a previous glaciation that survived interglacial conditions near the surface (e.g., Figure 8). This seems challenging to support without corroborating evidence, such as ice crystallography or age constraints. The resemblance to “basal ice” is unsurprising, given that this layer is located in subglacial frozen sediments near the ice-bed interface, akin to a frozen fringe. Theory predicts that a frozen fringe could grow meters thick and form ice lenses in basal till under freezing conditions (e.g., Meyer et al., 2023), following similar physics as frost heave in permafrost—an occurrence well-documented in the field (e.g., Christofferson and Tulacyzk, 2003; Fitzsimmons et al., 2024). A recent paper posits that clasts can migrate into ice through thermal regelation (Pierce et al., 2024), which could provide a mechanism to progressively incorporate till from the underlying layer. Ice crystal structure analysis could clarify whether this is highly sheared basal ice left over from a previous glaciation or ice grown through cryosuction, which would exhibit distinctive characteristics. Without such data, the timing of its emplacement remains speculative. One might imagine Unit 3 being deposited over Unit 1, with later ice lens growth. Based on the information provided, I don’t see how it could be said with certainty.

Additionally, while the presence of vertical cracks is intriguing and could potentially represent freeze-thaw cycles, I do not believe they are a smoking gun. This layer could have experienced any number of glaciotectonic events that induced brittle strain, and the history is difficult to deduce from a single core, which is essentially a point source.

Regarding the Interpretation for Unit 3

The authors describe Unit 3 as a slump deposit, partly based on the presence of normal grading. This interpretation is not readily apparent to me from the evidence provided in Figure 2. The core sections labeled as Unit 3 are divided between 1060-C3 and the lower portion of 1060-C1, with a substantial gap between them, which complicates a continuous interpretation. In the lower section (1060-C3), the sample is clast-rich, while the upper portion (1060-C1) contains mostly fines. This contrast presumably forms the basis for describing Unit 3 as normally graded. However, 1060-C3 does not visually appear graded; it lacks the clear stratification expected in a slump deposit and could simply represent basal till, with clasts suspended in a finer matrix.

The missing sediment between 1060-C3 and 1060-C1 leaves some ambiguity in interpreting the contact between these layers. This discontinuity between a poorly sorted till and fines could indicate separate depositional events/mechanisms, weakening the argument for a continuous slump feature. Additional data—such as grain size distributions with depth or clast fabric/orientation relative to underlying till sections—would lend stronger support to the slump hypothesis (though it would still be difficult to ascertain from a single core without lateral context). Clast orientation specifically could help distinguish a traction till, characterized by horizontal shear, from a flow till, deposited downslope. In lieu of these constraints, alternative interpretations are plausible.

“The paucity of grain coatings in Unit 3 compared to Unit 1 (Fig. 5) could indicate that the slumping process disrupted grain coatings or mixed sediment from Unit 1 with sediment from the upper units, 4 and 5.”

If Unit 3 was deposited as a basal till and subsequently covered by fluvial deposits, then the lack of grain coatings might be expected. My impression is that the main evidence for interpreting Unit 3’s deposition as subaerial or subaqueous, rather than glacial, is the presence of weathered coatings in Unit 1 below, which suggests significant exposure at Earth’s surface. This would imply that ice retreated, exposing Unit 1, before Unit 3 was subsequently deposited. An alternative hypothesis could be an ice readvance that deposited Unit 3. While the paleo-climatic sequence for this specific region isn’t my specialty, are there regional constraints that would support this specific locale being ice-free for the full duration of the interval given by the age constraints of Units 1 and 5, or is this largely unknown?

“The sub-horizontal, braided, lenticular cryostructures are consistent with syngenetic permafrost formation, suggesting based on cryostratigraphy, little influence of liquid water in Unit 1 after permafrost formation (French & Shur, 2010).”

Based on the evidence provided in this manuscript, the timing of these lenticular cryostructures remains ambiguous. While the authors suggest they formed as syngenetic permafrost in a subaerial environment, it is equally plausible that they developed subglacially over a range of timescales. Without additional dating or contextual evidence, attributing them definitively to subaerial permafrost formation is speculative.

“Overall, the similar sedimentary structures and mineralogy suggest that Unit 3 was originally a part of the subglacial till formed below the ice (Unit 1) that was subject to slumping due to saturation of liquid water during interglacial conditions.”

While I concur that this likely originated as subglacial till, the assertion that it is a slump deposit in an interglacial seems rather uncertain. As mentioned above, the timing here could have other possibilities. Even as a slump, perhaps it could be subglacial as well, perhaps representing the collapse of a canal’s sidewalls.

“Water and vegetation: During interglacial conditions, the permafrost landscape was subject to freeze-thaw cycles as shown by the vertical lenses of clear ice in Unit 2. Till, saturated by water, flowed downslope and buried the ice forming Unit 3. Interglacial conditions supported plant growth and the development of a headwaters fluvial system that eroded the upper portion of the flow-till deposit, stripping grain coatings but not changing grain shape. The fluvial system then deposited bedded sand, initially fine-grained and then coarser-grained material.”

The evidence presented here doesn’t necessarily indicate a unique interglacial, subaerial depositional environment:

1. Ice lenses could have formed in these layers subglacially over an ambiguous range of timescales.
2. Unit 2 may have grown after the deposition of Unit 3. If Unit 3 represents a slump (though this is not convincingly demonstrated), it’s plausible that it collapsed as a subglacial feature, potentially forming canal walls.
3. The introduction (Lines 60-62) states the presence of plant and invertebrate fossils in the core necessitates the site was ice-free during MIS11, but the authors do not explicitly state what units described in this paper contained signs of biological life. Fossils in Units 2-5 would significantly bolster the interglacial deposition argument, yet this isn’t explicitly discussed in this manuscript. In particular, are there biological remains in Unit 3?

Line 172: What evidence is there that this has undergone minimal deformation? Line 348 mentions deformed bedding, for instance, with no further explanation. Has previous work assessed the microstructure in detail? A citation would be warranted here.

Line 316: "Our multiscale analysis supports, refines, updates, and provides more justification and detail regarding unit delineations proposed earlier (Bierman et al., 2024; Christ et al., 2021, 2023; Fountain et al., 1981)." It would be helpful to identify in the discussion how observations made in this paper agree or disagree with past interpretations of the units.

Line 322: Technically, it is all subglacial material. Does this refer to Units 1 and 2?

Minor punctuation errors

• Line 47: "1969)), but the sub-glacial material" → missing comma.
• Line 172: "minimally, if at all, deformed."

Works cited

Meyer, C.R., Schoof, C. and Rempel, A.W., 2023. A thermomechanical model for frost heave and subglacial frozen fringe. Journal of Fluid Mechanics, 964, p.A42.

Pierce, E., Overeem, I. and Jouvet, G., 2024. Modeling sediment fluxes from debris‐rich basal ice layers. Journal of Geophysical Research: Earth Surface, 129(10), p.e2024JF007665.

Christoffersen, P. and Tulaczyk, S., 2003. Response of subglacial sediments to basal freeze‐on 1. Theory and comparison to observations from beneath the West Antarctic Ice Sheet. Journal of Geophysical Research: Solid Earth, 108(B4).

Fitzsimons, S., Samyn, D. and Lorrain, R., 2024. Deformation, strength and tectonic evolution of basal ice in Taylor Glacier, Antarctica. Journal of Geophysical Research: Earth Surface, 129(4), p.e2023JF007456.

This is a solid paper from Paul Bierman’s group on the surprising discovery of the sub-ice sediment from the Camp Century ice core recovered several decades ago and presumed lost.  This follows earlier papers by Christ et al (2023) and Bierman et al (2023, 2024). Below are some specific and general comments on the ms.

Abstract:  Because more people read abstract than ms, put into the Abstract your key findings: in situ MIS 12 till overlain by MIS 12  basal debris, then a sequence of MIS 11 fluvial sediment. You want to be sure that the casual abstract reader has caught your main conclusions, which may lead them to actually read the paper.

Introduction:  Line 1 of Intro is misleading “Understanding past ice-free times allows us to predict the response of the Greenland Ice Sheet 35 (GrIS) to current and future climate warming (Gemery & López-Quirós, 2024)”  This is not a claim in the cited article, and certainly not true as stated.   “current and future climate warming”  has no analog in the past in terms of rates of GHG change. This complaint in no way is meant to diminish the importance of paleo work, which provides the only means of testing how well climate models are accurate for times of different planetary energy balances and different Earth-surface characteristics.

Line 60 “the site was not covered by ice during MIS 11 and indicates the maximum limit for ice extent at that time”  I think it is safer to say “the site was not covered by an ice sheet during a significant fraction of MIS 11” as there is no duration in the authors’ datings from the core.

Figure 6.  This is a very important figure. From the lithostrat column it “looks like” three primary units. I take from this that Unit 1, the thickest unit, is pretty uniform in the characteristics you are using.  But the other units are much more distributed into different clusters, which suggests that there may be different depositional and re-working processes at work in the younger units, which complicates genetic interpretations.

1. Discussion

Line 519  “The rich record of ice sheet history”  I suggest deleting “rich”.  I see this as an important depositional record that reflects a complex deglacial history during MIS 11 that is difficult to constrain in terms age beyond the MIS 11 window, and also of processes that produced the sediment in the core segments.

It would be very helpful to go through this section unit-by-unit starting with Unit1, with the authors briefly summarizing  their interpretation of the depositional environment of their 5 “stratigraphic units” in the subglacial sediment and the reasons for that interpretation.  Details for their interpretation can follow in the sections under this heading.

5.1 Sub-glacial Core Stratigraphy: Synthesis of Physical, Chemical, and Mineralogic Observations

You should impose an order on this section, which summarizes your interpretations of the genesis of each unit based on all the data you generated starting with the basal Unit 1 and work your interpretation upward. As it is the text goes back and forth between the units, and it is too much work for the casual reader to follow.

Line 535: I was not convinced the unit 3 was a slump.  Could it not have been a “basal dirty ice unit from the MIS 12 glacial cycle? The underlying Unit 2 is dirty basal ice.  Why couldn’t Unit 3 be the upper part of Unit 2 that liquified when Unit 2 was exposed at the surface during an interglacial, with the fluvial processes imposing a different order on your primary measured variables?

Figure 8 This is a very helpful figure. I suggest adding to the right hand box the overall environmental condition as is done in Unit 3.  Unit 1: Glaciation, Unit 2 Deglaciation, Unit 3 early Interglaciation, Unit 4 Interglaciation, Unit 5 transition into the next glaciation?

Key points as I saw them, which is very close to section 5.2.  I think it makes more sense to list the items under 5.2 in terms of the 5 units identified

The sub-ice sediment consists of sediment deposited both when ice was over the site and when it was ice free.  Under the rule of interpretation following the “lines of least astonishment”, it seemed to me the sequence below is the most parsimonious interpretation of depositional events.

1. The 5 units start with a basal till (Unit 1), overlain by sediment entrained in the basal ice of a deglaciating MIS 12 glacial advance across the site (Unit 2)
2. Unit 3. Represents MIS 12 deglacial sediment that is likely derived from sediment carried with in the deglaciating MIS 12 ice, as well as reworking of dirty basal ice as the ice margin retreats inland, resulting in some sediment characteristic similar or very similar to the basal till of unit 1.
3. Once the landscape has stabilized in MIS 11, fluvial sediment dominates the environment, and this process evolves with increasingly aggressive flow in your upper units that may suggest the onset of an advancing MIS 10 Greenland Ice Sheet?