Review of Large synthesis of in 1 situ field measurements of the size distribution of mineral dust aerosols across their lifecycle by Paola Formenti and Claudia Di Biagio

This article surveys a vast amount of dust size distributions in the scientific literature (obtained by different groups using multiple techniques over different particle size ranges) and  synthesizes it into a dataset that allows comparisons amongst the different techniques. Significant details about the operating premiss of the various instrumentation are provided so that the reader can understand why this synthesis is necessary and the mathematical relationships (and assumptions) between the different techniques. I find this an excellent article that is clear and well organized, and I do not have any significant suggestions for improvement. I list a few minor grammar suggestions below.

Unfortunately, my comments only apply to the article itself, as I do not have access to the dataset. The links in the pdf resulted in a “doi not found” message from doi.org. The links on the ESSD website appear to be identical to the pdf links, but the ESSD links led me to www.easydata.earth, which NASA has blocked (apparently because the the site is not https protocol). I have requested access from NASA, but this could take weeks; if approved, the NASA approval might also expire at some point. Since there are many NASA employees who would be interested in this dataset, it would be wise for easydata.earth to upgrade to https protocol or for the authors to find another home for the data.

Line 151: Replace sizes with size.

Line 173: I struggled with this sentence because I found the first 8 words rather klunky in their arrangement.

Line 201: It is a little strange to reference the first author as ‘personal communication.’ I would just omit.

Line 207: Try a comma after ‘However’ and replacing ‘to note’ with ‘noting’.

Line 286: Replace ‘take’ with ‘taken.’

Line 334: I don't know that intensity is the best word, given its link to optical intensity for your likely audience. Perhaps try "Particle volume concentration above 10 um remains unchanged…"

Line 338: Replace ‘to evaluate’ with ‘evaluation of’.

Line 344: Remove “, but”, including the comma.

Line 345: Replace ‘while we identify’ with ‘and there is’, keeping the comma.

Line 353: “A large statistics of data… “  This is an important summary sentence that I believe could be made stronger with an active voice.  " We did this.... based on..." Basically, brag that you did this, as it is worth bragging about!

General Comment

Formenti and Di Biagio present a reference data set of mineral dust size distributions, compiled from a large set of harmonized datasets from a variety of field campaigns. The original data were corrected according to the applied set of instruments and mapped to a common log-scale size grid for easy use in atmospheric modelling. The procedure of correction and mapping is precisely described and statistical parameters like mean, median, standard deviation, and percentiles, are presented for three classes of mineral dust (source, mid-range transport, long-range transport) are provided. The presented dataset is of invaluable benefit since it provides for the first time a harmonized data set of mineral dust size distributions prepared for direct use in atmospheric modelling.

The manuscript is clearly structured and very well written. It is recommended for publication after a few minor revisions have been taken into account.

There is one general question to be asked before moving on to the minor comments: Is there a reason why the authors do not provide parameterized size distributions for the three dust categories? Of course, these parameterizations can be implemented by the individual researchers using the data. But the provision of parameterized size distributions may help making results from model studies on mineral dust more comparable.

Minor Comments

1. In the reference list, many journal names are not provided, the format of the reference list needs to be checked. Furthermore, check the use of names like Schütz or Müller, often they are given as Schutz and Muller.
2. The expression “providing with” is often used, but to my opinion not always in the correct manner. This should be checked.
3. The abstract may be shortened without losing its information content.
4. Line 36: Mineral dust particles do not emit solar radiation. The sentence should be rephrased.

Typos

1. Line 48: “orders of magnitude”
2. Line 51: “… and attributed THIS to clays …”
3. Line 110: “consist of “
4. Line 111: “one data point” should be rephrased as “one data set”.
5. Line 186: “i.e., only one data set …”
6. Line 247: Suggested title “Limitations of the chosen approach”
7. Line 334: The expression “The intensity of the particle volume …” should be rephrased, e.g., as “the magnitude of the particle volume …”.

This article compiles and unifies in situ dust size measurements from a large number of previous studies to form normalized distributions of particle volume versus diameter at three stages in the lifetime of a dust plume: near the source, a few days after uplift, and far downwind. The authors form averages across the studies for each stage, but to this reviewer (a modeler), what seems most valuable is the harmonization of the individual data sets to a physically consistent size variable: the geometric diameter. The authors additionally provide harmonized versions of the data that treat the particles either as spheres or make further assumptions to account for aspherical particles. What is particularly welcome, along with the accompanying files containing the harmonized data from each individual study, is the description of each study in the Supplement that lists the instruments and size descriptor that was measured, and most importantly, locates the measurements in space and time. Modelers can use these harmonized measurements from individual studies and information from the Supplement to compare their models consistently.

I think this paper is thorough, well-written and significant.  My comments are only to help improve clarity: 

line 10: are -> that are

line 11: proper -> properly

line 18: maybe use 'grid' for 'path'?

line 25: "decreasing to ~5 μm and ~2 μm for MRT and LRT conditions" This seems excessively precise, even with the approximation symbols, given the uncertainty and fluctuations of each of the three distribution curves in Figure 3.  Instead, I suggest describing only the shift of the mode from order 10 um to order 1 um as the dust moves downwind.

line 40: "started in the last decade" Dust models go back at least three decades to Tegen and Fung (1994) if not earlier (S. Joussaume, 1990; C. Genthon 1992).

line 55: being -> been

line 62: tenths -> tens?  line 67: "spurious": This variability is not necessarily spurious.  It may be instead just a reflection of local conditions that may not apply elsewhere.

line 79: by "ensembles" do you mean collections or compilations of measurements?

line 88: why exclude deposition samples?  Is this because of the additional assumptions that must be made to convert a flux into the concentration measurements compiled here?

line 106: How confident is this assignment of dust age? If there is a continuum of transport times, then some cases near the temporal boundaries of each category might be mischaracterized, a source of uncertainty. This uncertainty is fine if acknowledged, and I like the supplement because age is more clearly acknowledged as an estimate.

line 168: I understand the importance of harmonizing the different types of diameter present in the literature (e.g. geometric vs. aerodynamic).  What is unclear is what is meant by "differences in number concentrations" in this context.

line 197: CRI = 1.53 + 0.003i.  What is the citation for this?  The CRI depends upon wavelength, and should match the wavelength used by the instrument.  Is this range of instrumental wavelength small enough that a single CRI is representative?

line 224: Do these correction factors depend upon the observed distribution of shape parameters compiled by Huang et al. 2020: e.g. the aspect ratio and height-to-width ratio?  If so, a brief description of these assumed shapes descriptors would be useful.

line 244: "Additionally, the 25% and 75% percentiles are also calculated, despite keeping in mind their limited representativeness given the reduced number of samples in the datasets, especially for SOURCE and LRT classes." You should also note that each of the individual measurment studies have their own uncertainty that will vary with the instrument and duration of the measurement period that is not accounted for here. (I realize that this information is not always available or possible to estimate robustly.)

line 270: delete 'used"

line 286: take -> taken

line 287: "and when the first and/or the last bin of the corrected size showed unrealistic divergence" I do not understand what is meant by unrealistic divergence.

line 316: "A main mode located at ~10 μm" This is outside the range of normalization cited on lines 180-181. Maybe note that this mode near 10 um (and possibly beyond) is based upon only a few studies?

line 328: the normalized size distribution?

line 355: "Dust particles below 0.4 μm in diameter are seldom measured close to source regions, but are found in observations at mid– and long–range transport conditions." Could you speculate about why this is?  Is this emergence of a fine fraction the result of contamination by non-dust aerosols that is more apparent far from the source where dust concentrations are smaller? Alternatively, is it a consequence of the normalization, where larger particles are emphasized closer to the source and smaller particles are obscured?

Finally, I note that I had the same challenges to access the data as Greg Schuster.