Harlan et al. present a description of the measurement setup for the MBS CFA record and the associated uncertainties. Overall I found the  manuscript to be complete, well written and very readable!

I only have some minor remarks that the aurhors should address before publication:

The most important one is a potential issue with the relative delay estimation between the two parameters (Section 4.4): The authors determine the signal delays using the maximum of the derivative during the transition into the standard measuremens. This approach is certianly valid, but a small detail is missing from the discussion: The influence of the signal smoothing on this way of determining the ralative signal delays. For higher signal smoothing the maximum of the derivative will occure systmatically later as the derivative wil have its maximum at approximately halve the signal amplitude. This will introduce a slight systematic overestimation of the singal delay for paramters with larger smoothing relative to parameters with less signal smoothing. I would encurage the athors to add a few sentences in this regard.

The second  is the description of the differences between the 2018 and 2019 setup.s Judging from the flow diagrams in Figure 3, the systems were quite different with respect to the parameters presented here. I would suggest the authors expand this section significatnly to delineate all the differences also in terms of procedures.

Specific remarks:

L 123: “tempramental” is an odd choice of words. To avoid personificaiton mabye describe in terms of noise sensitivity

Section 3.0.2: See comment above.

Section 4.4: See comment above.

Section 5.0.1: Explain target meltrate especially in light of the density change downcore.

Section 6.1: What makes the pH measurements more sensitvie to the pressure fluctuations than the Na channel? From my understanding both are absorption spectroscopic methods and are thus more sensitvie to mixing ratio fluctuations than the fluorcense detection.

Section 6.5: Please explain the filtering aproach detailed enough so that others are able to repliclate the same thing, if needed. This will also help to judge the imprint of the filter on the final dataset in terms of its frequency content.

The Harlan et al. paper is the latest in a series of publications on the new MBS ice core, one of the rare high-resolution records of the last millennial. Harlan et al. present the CFA dataset, including meltwater electrolytic conductivity, sodium (Na⁺), ammonium (NH₄⁺), hydrogen peroxide (H₂O₂), and insoluble microparticle measurements.

Despite some language inconsistencies, the paper is well written and easy to read. Figures have been tested for several types of colorblindness and seem adequate in terms of colors used. The quality of the figures is good and the coherent use of colors in the different figures is appreciable.

However, I have a few concerns, especially about the lack of discussion about errors and data quality assessment, as well as a proper description of the interest and novelty of such a long and highly resolved record. The different major and minor comments are discussed below.

Major comments:

The dataset presented here is unique, with fine resolution and spanning a millennium. However, I feel that a key point is missing, namely the emphasis on what could be done with these new records, as well as highlighting the importance of such longer time records. Readers and future users would need information on how valuable the dataset is and on its potential future use.

This is particularly important as the actual minimum resolution of this dataset is considered to be 3 cm, the same resolution as the discrete (complementary) dataset published along the paper of Vance et al. (2024a). This also raises the question: what is the value, and the reliability, of a dataset with a resolution of 1 mm? A topic only partly addressed by the authors.

On line 31, it is confusing to read a mention of ice cores in the plural when the paragraph begins with the presentation of the new Mount Brown South ice core in the singular (line 29). I think it would be interesting to first present the site and the multiple cores (currently section 1.2) and then introduce Continuous Flow Analysis (currently section 1.1). This would also provide a more logical sequence for the rest of the content. In addition, the sub-section presenting these different cores could be expanded, typically by adding information on the length and position of Alpha, Bravo and Charlie cores in relation to the main core. Consideration should also be given to adding a paragraph (or a dedicated sub-section) on what has already been done on these different cores, and possibly summarizing the main findings of Crockart et al. (2021), Jackson et al. (2023) and Vance et al. (2024a). This would also provide more clarity on the addition of this paper compared to the other publications, particularly compared to the Vance et al. paper (2024a) which already mentions the CFA data (even if it's not clear whether Vance et al. use it or not). Finally, section 4.2 on MBS chronology could be added to this part of the text instead of being presented in the data processing section (while it has not been processed in this paper).

Are there any traces of melt layers and wind crusts in the cores? Have their (potential) influences been considered in the various recordings? Even if it has been said that one of the criteria for selecting a site is a minimum of snowmelt in summer, it seems essential to document their occurrence and explain the potential influence of these melt layers and wind crusts on the recordings, especially for the 1 mm resolution dataset.

My major concern is the absence of any discussion of the accuracy and precision of the measurements, which is an essential point in a data paper. Sub-section 6.3 only presents the limit of detection (LOD), i.e. the smallest concentration that can be measured by the instrument, but not how precise and accurate these measurements are. The relative standard deviation (RSD) - the standard deviation of concentrations of the standards divided by their known levels - of the lowest level standard can be used as a measure of precision, as in Griemann et al. (2022). With regard to accuracy, intermediate standards could be considered as samples (i.e. not used for calibration) and used to define the accuracy of measurements by evaluating the difference between their known concentration and the concentration measured during the run.

l.267: Have you compared the 3 cm resolution Na⁺ dataset with Vance et al.'s (2024a) discrete dataset? This could be used as (partial) data quality assessment.

Similarly, no error is mentioned for the third dataset presenting a decadal scale record. A global error calculation - including measurement and dating errors - should be associated with the decadal scale record.

Section 4.4 on signal delay time is not detailed enough. Are the numbers and errors the result of statistical analysis? How many estimates have led to these values? How regularly were the bulk and individual delay times measured? And, in practice, how is the “total” delay time applied to the different measurements?

I don't see how section 6.3 (l.256-261) can be used to verify data quality and depth scale accuracy. Firstly, it should be specified what data are involved here (conductivity) and the comparison between the conductivity measured by CFA and the discrete measurements of nssSO₄ by Vance et al (2024a) should be clearly mentioned. Secondly, approximating 90 cm as equivalent to 3 years is not correct: it is mentioned in the text that accumulation is highly variable on an interannual scale. Moreover, even if the accumulation were more or less constant, we assume that the authors are using a sliding average of 90 cm on the raw data (observed depth), but density changes with depth, so 90 cm at the surface does not represent the same number of years as at a depth of 200 m when the snow has been compacted into ice. Thirdly, I disagree with the statement “we are able to identify many of the volcanoes reported in Vance et al. (2024a)”. Looking at Table A2 in the Appendix, only 16 of the 32 volcanoes identified in Vance et al. (2024a) are also present in the conductivity record, i.e. 50%. This deserves an attempt at an explanation (for example, it is possible that the relative proximity of the coast induces a greater presence of impurities resulting in a higher background noise in the conductivity measurements).

Minor comments:

Some inconsistencies in the text:

• Two periods of CFA measurements are mentioned in the text but using different terms (CFA sampling/measurement/melting campaign) (l.54, l.218-219, l.238). The same wording should be used to facilitate understanding.
• The depths given for dry-drilled and wet-drilled sections are not the same throughout the text: for example, in l.95-96, dry-drilled and wet-drilled sections are respectively at ∼4-94 m and ∼95-295 m vs. 5-95 m and 96-295 m in l.218-219.

There is a problem with the table numbering. The first Table mentioned in the text is actually Table 2 (line 211). The numbering of the figures and the order in which they are presented should be changed in the text accordingly.

l.2: In the abstract, the authors mention a mean annual accumulation estimated at 20-30 cm ice equivalent. However, 20 cm is never mentioned in the main text. The abstract should only contain information that is present in the main text.

Figure 1 (near l.61): some values of lat-lon would be welcome.

Legend of Figure 2 (near l.80): "Shaded areas indicate cuts made with a bandsaw.” Do you mean dotted lines?

l.90: A few more details on the scraping of horizontal ends could already be given here: at the very least, mention how much thickness is removed and insist that this is taken into account in the depth logs, especially as one of the datasets is published at millimeter scale.

l.95: The authors should mention why the dry-drilled section only begins at a depth of 4 meters and how the chronology of these first 4 meters is carried out.

l.114: It would be interesting for the readers to mention which type of analysis will be performed in the future.

l.137: What are the resolution and accuracy of this cable-driven rotary encoder (in comparison to the previous system)?

l.145: The last sentence of the paragraph raises a question: why would we need more details about the gas extraction system if it was not used for data collection? The answer is only 3 pages later. I suggest adding a few words alluding to the effect of gas extraction on the CFA data presented here.

l.159: What is the impact of inaccuracies of one or two millimeters on the 1 mm dataset? This should be estimated and mentioned in the text.

l.165: It is mentioned that the procedure for correcting differences in measurements of the same stick is presented in Vance et al. (2024a). However, Vance et al. (2024a) state that “The scaling and shift factors will be described in detail as the CFA trace chemistry and water isotope datasets are developed and published”. So it seems that information is missing here.

l.184: For readers unfamiliar with CFA measurements, it might be interesting to explain the principle of calibration with standards run at the beginning and end of each measurement. Is it a single calibration with some standards run at the beginning and others at the end? Or two distinct calibrations to account for e.g. measurement drifts?

l.189: Shouldn't the MATLAB script mentioned be made available in a code availability section, as required by the ESSD guidelines?

l.205-206: Attention should be paid to the significant figures used here (either 79 ± 4 seconds or 79.x ± 3.6 seconds).

l.254: Is Figure 4 needed? It is not really discussed in the text and is quite redundant with Table 2.

l.279: When mentioning “baseline values”, do you mean “true” observed baseline or the limit of detection? It seems more accurate to define the LOD as a threshold.

Figure 5:

1) The last data point of each complete series reaches 0. If this is an artifact of the plotting (which it probably is), the last point of each series should be removed.

2) It looks like there are several points reaching 0 in the H₂O₂, Na⁺ and NH₄⁺ series. If this is the case, these points are probably below the LOD/baseline (especially for H₂O₂ and Na⁺) and should be removed according to text line 279. It may be interesting to try to plot the LOD for these series (even though they are low for H₂O₂ and very low for NH₄⁺).

l.286: Is there a reference to confirm that the layer thinning is negligible at this ice core site? This is surprising, given that layer thinning already has a significant effect at 80 m depth in records such as that from Philippe et al. (2016). I also recommend adding the cause of this “layer thinning” (due to strain rates) to the text, to leave no doubt for the reader.

Technical corrections

Pay attention to sentence syntax (same word used several times in the same sentence):

• l.2: the second word ‘accumulation’ can be deleted.
• l.184: twice the word ‘run’ in the same sentence.
• l.223-224: the first part of the second sentence (this gives a smoothing effect) repeats the first sentence (the dynamics lead to smoothing).
• l.226: ‘calculated’ used twice in the same sentence.
• l.227: similar as previous comment with ‘using’ and ‘use’.
• l.272: similar as previous comment with ‘applying’ and ‘applies’.

Some inconsistencies in the writing:

• l.99: lowercase after the ‘:’ (as in l.7 for example).
• There should always be a space between the number and the units (see ESSD guidelines): l.123, l.163, legend of Table 1, l.227, legends of Figures 4 and 5 (1 µm)
• (e.g. l.152) The word ‘meltrate’ is sometimes written ‘melt-rate’.
• l.178-179: standardize the expression “a # step calibration” or “a # step calibrations”.
• l.218-219: interval values (in this case, depth values given in brackets) must either be joined to the dash or separated by a space, but must be consistent throughout the text (including in Table 1 or in other intervals like in l.95-96, l.139-140, legend of Table 1, l.227, …).
• l.225: CFA system (with a lower-case s, for consistency).

Some bibliography citations need to be revised:

• General question: which order do you use to cite multiple references in text (for example, in l.112-113)?
• l.100: should be “(Bigler et al., 2011)”.
• l.122: should be “in Dallmayr et al. (2016).”

Purely technical corrections:

• l.18: delete the comma before (Legrand and Mayewski, 1997).
• l.122: a point is missing after the citation.
• l.148: “The position derived *from* melt-rate data”.
• l.199: “(15 seconds for both *CFA setup* systems)”.
• l.219: meltrate (instead of metlrate).
• l.220: the second parenthesis ) is missing at the end of the sentence.
• Table 1: min-1 should be in superscript.
• l.229: min-1 should be in superscript.
• l.254: a point is missing at the end of the sentence.
• Legends Figures 4 and 5: perhaps the expression should be “particles > 1 µm per ml”.
• l.272: I suggest adding a ‘by’ before ‘applying’.
• Table A2: a hyphen is missing at 213.58 m.

References:

Grieman MM, Hoffmann HM, Humby JD, et al. Continuous flow analysis methods for sodium, magnesium and calcium detection in the Skytrain ice core. Journal of Glaciology. 2022;68(267):90-100. doi:10.1017/jog.2021.75

Philippe, M., Tison, J.-L., Fjøsne, K., Hubbard, B., Kjær, H. A., Lenaerts, J.T.M., Drews, R., Sheldon, S.G., DeBondt, K., Claeys, P., and Pattyn, F.: Ice core evidence for a 20th century increase in surface mass balance in coastal Dronning Maud Land, East Antarctica, The Cryosphere, 10, 2501–2516, https://doi.org/10.5194/tc-10-2501-2016, 2016.