Review

The authors present in-situ observations with a high altitude aircraft of size-resolved ultrafine aerosol particle concentrations, ice water content, ice particle size, carbon monoxide mixing ratio, and meteorological quantities in the tropical tropopause layer over Nepal, India, and Bangladesh, during the Asian Monsoon season in 2017.

During the eight flights, recent nucleation of ultrafine aerosol from the gas phase (new particle formation, NPF) was identified from the size-resolved ultrafine aerosol particle concentrations. NPF events that produced large amounts of new particles in the altitude range 11-16 km were identified both in clear air and cloudy air, with cloud ice reaching number concentrations as high as 3 cm^-3. While earlier in-situ observations have identified NPF inside cirrus clouds, and numerical modeling showed that NPF in cirrus clouds is possible based on known mechanisms, such systematic occurrence of NPF in cloudy air in the tropical tropopause layer, producing large numbers of new particles, is remarkable. The observations presented by the authors, a result of their significant scientific, technical, and organizational ability, are a major contribution to a better understanding of the upper troposphere. In this respect, the manuscript is of high quality. The analysis and discussion of the observed ultrafine aerosol concentrations is very good with respect to the structure of the upper troposphere and and carbon monoxide. In particular, the discussion of the role of surface air and pollutants for the new particle formation is very interesting: The moderately elevated carbon monoxide values in the majority of observed NPF case show that lofting of surface air contributes the observed NPF events, but the lofted air is almost always diluted when NPF has been observed. The findings that in-cloud NPF is strongly suppressed in the presence of predominantly liquid-origin ice particles (as opposed to in-situ cirrus ice particles), and that the observed NPF is largely independent of the carbon monoxide content (which indicates time since surface contact and NPF precursor load) are important. The statistical analysis of NPF events in the presence of ice particles is a very useful quantitative analysis of the results.

Where the manuscript falls short is the valiant, but lengthy, complex, and mostly inconclusive attempt to produce quantitative relationships between ice cloud properties and ultrafine aerosol concentrations / NPF strength (Sec. 5). Relationship that are identified are weak and only applicable to subsets of the collected data, producing much hypothesizing and conditional statements that prevent strong conclusions. The main insight is that below the integrated radius (IR) threshold of ~ 1 μm cm^-3, the observed ultrafine aerosol concentrations are independent of IR, while above it, the maximum concentration of ultrafine aerosol falls linearly with increasing IR. This insight may hold in general, but it also could be limited to this data set. The reason why it is so difficult to construct quantitative relationships between the observed quantities is, in this reviewer's opinion, the complexity and nonlinearity of the processes (transport, precursors, scavenging, mixing, chemical conversion, time since NPF, etc.) that shape the observations, rather than any failure on the part of the authors.

This reviewer's recommendation is to focus on the very significant results and insights in this work, of which there are plenty, and drop the complex, hypothetical, conditional, and inconclusive elements of the analysis that weigh the manuscript down and compromise the overall quality of the work. The manuscript would also benefit from proofreading for English; some expressions are used in an unusual way ("constrained" is a better expression than "confined" in many places, etc.), and the language could be simplified for clarity and ease of reading. A major revision is recommended to provide sufficient time for any necessary changes.

Specifics

Line 58: "From the CLOUD experiments, which were performed under a variety of controlled conditions, it can be deduced that the intensity of NPF (the formation rate of new particles per air volume and per time unit) depends on the concentration of the NPF precursors."

Hasn't this been known long before the CLOUD experiments - e.g., https://doi.org/10.1029/2003JD004460, and others?

Line 65: ... indeterminable, because ... (expand a little bit on this for the reader's sake)

Line 76: "however" may not be necessary here - could be dropped for simplicity.

Line 77: ". Under real conditions in the atmosphere, however, the concentration of precursor material is spatially and temporally highly variable."

Pleas provide one or two references.

Line 93: "Investigations concerning the occurrence of NPF within clouds, or in their immediate vicinity, are sparse ..."

This statement would appear incompatible with the work of Clarke and Kapustin (2006), who report decade of data on particle production, transport, evolution, and mixing in the troposphere, much, if not most of it, near clouds.

Line 95: "... possible reasons for this are discussed by Wehner et al. (2015)." 

Please briefly give some of these reasons - this will be illuminating to the reader.

Line 229 "... principally based on the difference of both quantities (cf. Weigel et al. (2011)). "

This could be removed.

Line 247: "coincidently"

replace with "coincide"

Line 280: "computations"

replace with "computational"

Line 375: "The encountered in-cloud NPF events at altitudes between approximately 11 km and 16.5 km (~ 355 ? 385 K) had a mean event duration of 14.5 seconds (ranging from one second to a maximum of about 300 seconds)."

"event duration" means "flight time spent in air with in-cloud NPF", is that correct? If yes, please make sure that this is clear, because the reader might otherwise assume that this refers to the time period during which NPF took place.

Line 648: " It is not likely that a high number of interstitial, non-activated aerosol is accountable for the abundance of submicrometre-sized particles."

Please substantiate that "it is not likely", or if substantiation is not possible, remove the passage.

Line 666: "... likely suffice ..."

Please substantiate this, or if substantiation is not possible, remove the passage.

Line 912: "The IR turned out as appropriate cloud ice related parameter to juxtapose with NPF data."

This is a very confident statement given the very limited explanatory power of the IR.

... the maximum concentration of ultrafine aerosol falls exponentially with increasing IR ...

... the maximum concentration of ultrafine aerosol falls exponentially with increasing IR ...

New particle formation inside ice clouds: In-situ observations in the tropical tropopause layer of the 2017 Asian Monsoon Anticyclone

This paper describes CPC measurements of sub-15nm particles over Kathmandu at up to 20km altitude. Elevated concentrations and new particle formation were frequently coincident with cloud ice at 11-16km altitude. The authors determine that the NPF occurrence can be limited by the integral radius of the ice particle size distribution, which, as they point out quantitatively, makes sense if precursor vapors condense onto the ice instead of forming particles. This and the other findings they document are very interesting and based on a valuable dataset and sound reasoning, although sometimes only weakly supported by the data (see major comment below). The paper has a bit too much text devoted to qualitative and/or speculative details, and a long-winded writing style which sometimes detracts from its key messages. However, if my comments below can be addressed, this generally very good paper will be well worthy of publication in ACP and it should be highly cited.

Major comments:

Based on the scatter in Figure 8, the integral radius doesn’t work as well for controlling NPF as one might hope. Of course, given the difficulties of measuring NPF on an aircraft, not knowing accurately precursor concentrations or air mass history, one should not expect too much. Perhaps because of this, the authors don't currently present quantitative metrics for whether or not the IR is any use. So maybe the authors can disentangle the data more brutally to extract some numerical information on the usefulness of the IR? If they excluded data with IWC below a threshold, say 0 on their log scale in Figure 8, then is there some correlation coefficient between IR and N_uf? Even if no meaningful correlation can be presented, perhaps the message can be firmed up with further stratification of the data? Otherwise, it is hard to justify the lengthy text and detailed discussion associated with the IR in the paper, and the authors could instead substantially streamline these sections and perhaps focus instead on drawing out more quantitative conclusions about the role of in-situ vs liquid-origin cirrus.

Weigel et al 2020a is only referred to for specific details, why not take advantage of it for the broader context to avoid repetition (e.g of lines 173-193 in Weigel et al 2020a), and maybe even call this Part II of a 2-paper series (something to discuss with the editors)? The current situation is quite confusing. As another example, the first paragraph of the summary is background information that describes findings that are very similar to those in Weigel et al 2020a and there was not sufficient referencing provided.

Minor comments

The diversity of units used in the field is confusing (not the authors’ fault). Will be helpful to present aerosol number concentrations in cm^-3 alongside mg^-1 (as concentration in cm^-3 is relevant for the molecular collision frequency leading to NPF) and altitudes in km alongside K and hPa throughout the text and especially in section 3.2.

Figure 2 might be more helpful as a frequency distribution of number concentrations exceeding a threshold, or if the existing figure is accompanied by something like that.

L389: “Furthermore, there is no obvious indication that the number of ice particles present had a direct influence on the NPF strength” seems inconsistent with later L516 “Although an ultimate observational evidence is currently lacking, however, these findings suggest that NPF is entirely prevented in cases when Nice substantially exceeds 2-3cm^-3” – maybe add a qualification to the earlier statement to make this later statement seem less at odds with it.

L409 Could cite and discuss Bianchi et al 2020 (an understandable omission given the date this article was posted). https://www.nature.com/articles/s41561-020-00661-5

L429 what about ammonia?

Personally, I find it much easier to read and review papers, especially long ones, if the figure captions are on the same page as the figures (and preferably presented when they are first mentioned in the text rather than at the end, though this is less important). By the time I have found the figure that relates to a point in the text, opened the paper in two more instances of my browser, found its caption, looked back to the figure, understood the caption, gone back to the figure, understood the figure, I have forgotten why I was interested. Maybe this is a matter of opinion, but the ACP guidelines here:

https://www.atmospheric-chemistry-and-physics.net/submission.html#reviewfiles

say “Figures and tables as well as their captions must be inserted in the main text near the location of the first mention (not appended to the end of the manuscript)” so it seems I am not alone.

While generally nicely written, the paper is long, and the writing could often be more economical. I encourage the authors to go through each paragraph sentence by sentence as if there was a page limit, and use more efficient phrasing and omit unnecessary details. The paper would be easier to read and the authors would save on page charges.

A few sentences are written confusingly: “In particular, the abundance of in-cloud NPF concentrates between ratios of 1:30000 and 1:500000, which may not further surprise, as the large aerosol number concentrations are indicative to result from NPF.” What is an “NPF concentrate”?

Also, there are some typographical errors; I pick out only examples. Many commas (e.g. before “that”) reminiscent of German should be removed. On line 234 “principle” is confused with “principal”. Finally, while not strictly incorrect, manuscripts are usually “drafted”,  not “draughted”. See https://www.merriam-webster.com/words-at-play/using-draft-and-draught