In their Measurement Report "Insights into the chemical composition of molecular clusters present in the free troposphere over the Southern Indian Ocean: observations from the Maïdo observatory (2150 m a.s.l., Reunion Island)", Romain Salignat et al. report on observations of a selection of trace gases in the marine free troposphere, enabled by the deployment of a sensitive mass spectrometer at an island mountain site, which is argued to be exposed to the free troposphere during night times in the selected time period. Thereby, the paper contributes to overall sparse observations of various acids and amines in the free troposphere (FT). These compounds play important roles in aerosol chemistry, including new particle formation, so the reported observations are certainly of interest to the atmospheric science community.

A downside of the study is that it is based on only a week's worth of usable data (primarily, I understand, due to unsuitable environmental conditions otherwise). As a consequence, the statistical analysis of the measurement results on the detected compounds, together with air mass history investigation, appears rather weakly supported. However, that weakness is acknowledged, and the analysis was sufficiently carefully conducted, resulting in potentially useful hypotheses in aid of our understanding of free tropospheric chemistry.

I do have some concerns and comments, but I am confident they can be addressed. Subject to that, I recommend publication as a measurement report in the Atmospheric Chemistry & Physics journal.

Major comments:

1)

How were instrument backgrounds ("zeros") determined, and what was their result?

This potentially important aspect of the trace gas measurements is missing (or, if I missed it, insufficiently discussed). I got interested in particular due to the detection of fluorinated acids, quite consistently throughout the week, and amines, with a continuous substantial presence, especially of C4 amines (Fig. 7). Indeed, the fluorinated species are suspected to be due to some instrumental contamination (P18, 1st paragraph), which zero measurements might be able to confirm, and implying the importance of zeros potentially also for other species.

1b)

The time series for the C4 amines additionally spiked my interest, as it exhibits numerous "spikes" followed by slow decays (Fig. 8). Are those instrumental or real? Either way, can the authors speculate on the causes for those spikes?

2)

Do the authors have idea about the identity of the other major peaks observed (prominent unidentified mass spectral peaks in Fig. 6a)?

They should at least discussed. Even if unidentifiable, a mass defect diagram could provide clues?

3)

"Molecular clusters" (title and text).

The term "molecular clusters" in the title (plus corresponding throughout the text) is misleading. Measurements were primarily made of certain acids plus two classes of amines. The only observed clusters (less the reagent ions/clusters) were made up by up to 2 sulfuric acid molecules. Often the reference is to the observed "ion clusters", which are hence primarily a product of the measurement method but not as such present in the atmosphere.

If that understanding is correct, I recommend adjusting the title (and text).

4)

Section 4.2.2 overall (and Figs. 10, S10):

a) I wonder how significant are correlations with "time spent over land other than Reunion" and "time since last passage over land other than Reunion" are, as they could only be quantified for one of the nights, if I understand correctly?

b) Can the authors speculate about the mechanisms of "time spent over Reunion" affecting FT composition, in particular amine concentrations? I have not discerned that from this rather detailed and complex discussion of that analysis.

c) Maybe a table of hypothesized source mechanisms for the observations of the various compound groups could help in summarizing this section.

d) I am not an expert with NO3-CIMS in particular, but could the anti-correlation of most signals with RH be connected to instrument sensitivity potentially varying with RH?

General comments:

5)

Up to Section 3.2, the impression is given that a new method is found of establishing FT vs BL conditions via a tracer (e.g., P3 L26 "paper presents a tracer", P4 L9 "identify a tracer", P5 L34 "determine a tracer"). Or I have found those formulations misleading. Only in 3.2, it becomes apparent (to me) that the same approach as in Rose et al. (2017) and other papers is used (including preceding ones cited in Rose et al.), except, I believe, for a weaker threshold value.

I would consider making that clearer already earlier.

6)

A figure illustrating the various models' horizontal grid sizes and locations would be useful (including a length scale). E.g., based on the map shown in Fig. S5 (which misses the scale by the way).

Minor comments:

Abstract:

I would consider restructuring the abstract so that the identification of FT conditions at the site (plus implications) is described BEFORE the findings regarding cluster compositions and sources. I.e., first describe the setting, then the results. (Like in the text actually.)

P1 L23: would remove the latter "the", as I am suspecting that not all clusters were detected. [Later edit: see also comment (3) above regarding "clusters".]

P1 L25-26: I believe this sentence has some grammar issue or is missing a part.

P2 L13: References to two technical papers are given, but missing somewhere here are references to studies that have reported on molecular cluster/precursor observations related to NPF.

P2 L14 (or earlier): It would be instructive to the reader to briefly explain the difference between BL and FT. "Remote region of the atmosphere" could also refer to near-surface air except in remote areas, such as the marine BL. The key is of course the vertical layering of the atmosphere and its relation to vertical mixing. I am missing here any reference to the vertical. (The FT is, in a way, always quite close: just a few kilometers (or less) up.)

P2 L25: Which "these events" is being referred to?

P3 L22: How far inland is the Maïdo station?

P4 L~20: Was there any humidity control part of the inlet/DMPS setup?

P5 L7: A lack of "blanks" for amines is noted. How? Wouldn't any blank always blank all compounds at the same time, as a TOF-MS was used?

P5 L10-15: I wonder if the detailed explanation of how amine signals were normalized in various figures wouldn't be better placed later or in the figure captions.

Table 1: Suggest to clarify somehow, which rows are for instruments and which for models.

P7, section 3.2., 1st paragraph:

The first paragraph talks about previous works studying BL development on Reunion but is very vague on what those studies have actually found ("diurnal cycle", "result of complex combination..."). I understand that this study's approach is conceptually simpler, but as the previous works are brought up, it would be more useful to the reader to briefly explain how the BL has been found to generally develop over the mountainous terrain around Maïdo.

Section 3.3:

In the beginning, it is pointed out that Meso-NH simulations were not performed for one period (OCTAVE), so that use of the CAT model was evaluated for the case of another period (BIO-MAIDO). I don't see the logic here right away? And then the 2nd paragraph continues by talking about Meso-NH and ECMWF simulations, causing more confusion. I assume ECMWF means CAT and onward discussion (incl. Fig. 3) is for the BIO-MAIDO period?

I suggest rewording and restructuring the story somewhat for a more consistent flow.

[Edit: I didn't remember Table 1 etc. when I read Section 3.3, explaining most of my confusion probably. Some more detailed or streamlines explanations here could be useful anyway.]

Fig. 5:

I would add some labels to the maps (can be abbreviated, with explanations in the caption) to accommodate readers with less geographic background knowledge, at least for Africa, Madagascar, and the Ocean(s).

P18 L31 - P19 L1: Please de-convolute this sentences. I cannot get my heard around it.

P19 L2-4: Are these increases at sunrise/sunset significant? If so, it appears that H2SO4 exhibits the same behavior, at least on April 14?

P19 L28-29: Would be nice to see that model fit in a supplemental figure.

Figure 8:

What were signals normalized to in panels 2-4? Would be good to explicitly explain in the caption. For example, malonic acid fractions apparently exceed a factor of 10 at times. Does that mean its signal was 10 times that of the sum of the reagent ions (cf. P5 L1)? If so, I will have more comments, but probably there's a misunderstanding.

P21 L17-21: Please de-convolute this sentence. I cannot follow.

P21 L24: "... than those measured" where and by whom?

The manuscript by Salignat et al. investigate the clusters and the air masses in the marine free troposphere of the Maido observatory (2150 m a.s.l.) from Reunion Island in the South Indian Ocean, by the deployment of advanced mass spectrometry techniques and modelling techniques. They first identify a tracer of FT at the site, then study the origin and positioning of air masses before their arrival at the site in FT conditions, and lastly investigate the molecular composition, diurnal variation, and potential origins of the clusters in MFT combining the two techniques. Studies on the sources, processes, and composition of clusters at high-altitude environment in the MFT are scarce but important. And this study provides new insights into a better understanding of MFT, particularly in this region. I would therefore recommend its publication after the authors address my comments and questions below.

Specific:

Page 1 Title. The reviewer felt the title was only reflecting/summarizing part of the results presented in this study. Maybe adding the origin of air mass or clusters as well?

Page 3 Line 6. It seems a little bit contradictory to the sentences in Page 2 Line 15 where the authors cited several NPF observation in low FT or the interface of BL and FT with some of them from stationary measurements.

Page 7 Section 3.1. The reviewer was wondering whether this section would fit better at the beginning of Section 2 as a separate subsection, considering it’s mostly description of the measurement site from literature instead of results from this study?

Page 8 Line 9. It would be nice to add the frequency/fraction of BL thicknesses below 6m as FT conditions. Are there any periods at night that are still under BL conditions, or the station is always at FT conditions at night? Based on Fig 4 from OCTAVE campaign, it doesn’t seem to be always at FT at night.

Page 8 Line 15. The reviewer was wondering about the vertical wind. Would it be a better tracer to reflect the influence of BL at the site? Is the usage of horizontal wind instead of vertical wind due to the commonly-unavailability of vertical wind dataset? Could the authors clarify this a bit?

Page 9 Line 15-16. It would be nice to provide a plot at least in SI to support the statement that the stricter threshold would not improve the results.

Page 12 Line 15-17. It seems bigger differences between the two models for Fig 3c-e compared to Fig 3b from northwest. Is there any explanation for this?

Page 13 Figure 4. There are FT periods with relatively high RH values (e.g. night of April 11-12, 13-14, 14-15). Would the water mixing ratios provide better comparison for FT vs BL differences?

Page 14 Line 13-15. Consistent with the relatively high RH values of the night of April 11-12, 13-14, 14-15 in the reviewer’s previous comment, it seems these nights have air masses from pristine marine air from Indian Ocean or southern Madagascar. The reviewer would probably category Fig S6a (night of April 13-14) to group 3 like Fig 5c (night of April 14-15) instead of group 2 considering the similarity (also shown in Fig S7). Is this because it’s the night with air masses passing through terrestrial boundary layer other than Reunion as mentioned in Line 27-28? Could the authors clarify this a bit?

Page 15 Line 17-21. Considering the much higher signals of reagent ions and isotopes have been excluded from the calculation of the total signal, what’s the unidentified species (>75% fractions) over the mass range 80-400Th? The fraction seems a bit high. Are they mostly organics? They are labeled as “Others” in Fig 7a with pretty high signals. More description and discussion on this are needed.

Page 18 Line 18-21.

-Could the coinciding drop of RH be related to the advection of air masses from the higher altitude of the station? Can the models help to explain this?

-Could the increase of SA be due to similar reason as the increase of MSA, i.e., the evaporation from condensed phase (Page 19 Line 26)?

Page 20 Line 15. The reviewer was wondering how could MSA fragmentation in the mass spectrometer form SO5-? Could the authors add a reference to support?

Page 25 Line 10-11. The reviewer was wondering why would lower time spent over the ocean lead to higher wind speed along the air mass path. Could the authors clarify a bit and add a reference to support?

Figure 8. Considering April 16 early morning till 10 a.m. local time was in FT conditions with not too low radiation (Fig 4), the reviewer was curious whether this day was a NPF event day since SA and MSA levels were not low? If yes, is it possible to study e.g. few hours of FT nucleation/clustering process?

Technical:

Page 7 Line 7-8. Change to “the intensity of which vary…”

Page 7 Line 9. Change to “to a large extent”.

Page 18 Line 2 and 14. Remove “the order of”.

Page 18 Line 16. It only seems to be up to ~50% in Fig 7 on April 12. Please double check.

Page 21 Line 15. Remove the duplicated “in”.

Fig S4. It would be nice to swap the x axis from -12h on the left to 0h on the right side.