General:

This manuscript presents a newly developed polar nephelometer for aerosol phase function characterisation, along with comprehensive calibration discussion and quantitative error analysis. While other groups have demonstrated polar nephelometry with similar configurations, the novelty of this instrument is its reduced size. The authors provide an extremely detailed and rigorous discussion of the data analysis, including the main sources of uncertainty in the data inversion procedures. The performance of the nephelometer and predicted errors are demonstrated via laboratory experiments using monodisperse and unimodally distributed non-absorbing aerosol with known optical properties. While it would have been interesting to see an exploration of more novel materials or more significant instrumentation developments, the strength of this manuscript is in the thorough analysis. I especially appreciate all the detail discussed in the appendix and shown in Supplementary figures. As such, it is a valuable and excellent piece of research that is highly recommended for publication in AMT after a few minor revisions.

Specific comments:

L95: One of the novel features of the uNeph, compared to other polar nephelometers, is its relatively small size. It would be worth reiterating this advantage elsewhere, e.g. the conclusion section.

L97-109: Details on the optical system are scant. It would be useful to provide more details on:

• Type of laser and operating mode (i.e. pulsed or cw)
• Whether beam was collimated and estimated beam size within the measurement chamber
• Range of optical densities for ND filters used
• Type of photodetector

L111: Based on Figure 1, it would appear that the camera lens forms part of the chamber seal, and the rooftop reflector is outside a window – is that correct?

L117-121: Can you clarify the dimensions of the region of interest within the camera pixel array that was used, and whether pixels were binned at all? Despite the large pixel array indicated in Section 2.1, Figure S3 indicates a much smaller array.

L120: More details about the objective (lens) would be helpful, e.g. camera position relative to the focal length, the field-of-view…

L162-164: Depending on the CPC mode used, this would imply that the aerosol flow to the uNeph was around 3.5 or 4.7 L min^-1 – can you clarify what you estimate the volume flow in the uNeph to be (and hence the estimated residence time).

L485; L721-22: I believe the N limit in the summations over i indices refers to the number of angles – is that correct? It’s unclear whether this is incremented by pixel angle or by absolute angle. How does the angular resolution of measurements vary across the full range of measurement?

L592: It would appear that the V_tot predicted by the retrieval systematically overestimate the volume when compared to the SMPS measurement. Could this be because the measured size distribution (Figure 8) appears by eye to deviate from an ideal lognormal distribution?

L618-619: It would be interesting and useful to test your calibration and data processing procedures for absorbing spherical particles in the future.  

L797: This seems like a very conservative (i.e. erring on the side of large uncertainty) way of quantifying this error. No corrections needed, just an observation!

Table 1: Please clarify whether “q25” and “q75” refer to the 25^th and 75^th percentile quantile points, or other values.

Figure 2: No image appears in the manuscript

Figure 5: It is unclear the purpose for showing the line depicting “3% of uNeph air measurement” in the lefthand panels

Figure 6: Check units on stated number concentrations in caption

Figure 8: The information presented in (b,c) plots might be more clearly depicted in a table

General comments on Figures 6, 7, S2, S4, S8, S9, S11, S12, S13, S14, S15, S16, and S17 : The font size in the axes labels, legends, and/or schematics is far too small.

The manuscript provides a detailed explanation of a new imaging nephelometer called μNeph. Data processing methods are described, followed by validation of the measurements with monodisperse and monomodal polydisperse test aerosols and then finally an example of the potential to perform retrievals of particle size and concentration using μNeph data. The demonstrated accuracy of the μNeph is impressive, especially considering its small size, and I expect it to contribute significantly to our ability to measure aerosol phase matrix elements. Additionally, some aspects of the described data reduction procedure represent advancements over prior approaches, like the use of a monatomic gas (e.g., argon) to characterize the stokes vector of the incoming laser light source. The manuscript is clear and well written making it a very useful resource for future users of μNeph data and other imaging nephelometer efforts. I can thus confidently recommend publication, after the following minor points have been addressed.

SPECIFIC COMMENTS 

LN 30: I think this should read "...aerosols makes them difficult to..."

LN 36: It should just be "radiance" here, not "irradiance".

LN 71: It may be good to also cite the very well known Amsterdam–Granada polar nephelometer that is capable of measuring all phase matrix elements (e.g.,  Muñoz et al. 2012).

LN 95: Would the authors be able to elaborate on potential uses that they envision for the instrument outside of the laboratory (e.g.,  perhaps onboard a UAV)? The small size of the instrument seems like it would be advantageous for many such applications.

LN 116: Perhaps this would be clearer as greater/less than symbols (e.g., "0≤θ≤90°") to avoid confusion with the minus operator.

LN 117: Presuming the laser beams fill most of the length of the image, this CCD resolution would imply a raw measurement resolution of closer to 0.1°. Are multiple beam cross section integrals averaged to obtain the coarser 1° resolution? Also, how was 1° selected as the final resolution?

LN 118: I assume "A/D" stands for "analog-to-digital" but it still may be good to define the abbreviation explicitly. 

LN 121: Please specify the diameter of the pinhole used to image the lasers.

LN 127: It would be helpful to state exactly where the Thorlabs PAX polarimeter was placed in the optical path. Is it possible to place it in the measurement chamber? If so, why are q1 and q2 not known exactly, but instead later solved for using argon measurements? Perhaps this is due to the lack of exact knowledge of the camera's pinhole in the Thorlabs polarimeter's coordinate system? Please elaborate.

LN 180: Apologies if I missed it, but how frequently are dark images obtained? Are these coefficients updated during a given measurement?

LN 194: It would be helpful to replace "several" with the exact number of pinhead positions.

LN 234: While some referencing of supplement figures in the main text in passing is certainly fine, here and in several other places (e.g. LN 258, LN 329, LN 390, LN 499, etc.), a figure from the supplement is described in very significant detail. In these cases where whole sentences or paragraphs are devoted to supplement figures I would suggest either moving that text to the supplement, or the figure into the main text.

LN 240: Are these Ξ limits fixed or a function of scattering angle?

LN 240: Will the value of Ξ at which saturation occurs be dependent on aerosol loading and dark current? The effect is probably small, but it seems that short exposures (low dark current) and high loading (high counts in center of last beam) would be more likely to saturate pixels at the center of the beam than a longer exposure with lower aerosol loading, even if both cases produced the same value of Ξ.

LN 282: Have the authors considered using a beam expander? This was considered for the PI-Neph of Dolgos and Martins (2014) in order to reduce the signal at each individual pixel, while also enlarging the sampled volume providing statistics for a larger number of particles in each image acquisition.

LN 287: Is the precision of the laser reference photodetector known? If so, it would be helpful to state it.

LN 328: Do the authors have any theories as to the primary drivers of the instrument calibration drift?

Eq 3: I'm wondering if something like F∥ and F⊥ would be clearer than F1 and F2, which are very easily confused with phase matrix elements. Although, perhaps the close connection (and same units) with phase matrix elements make the variable names appropriate. 

LN 371: Is there a physical basis for the assumption |q1|=|q2|? (Or, more precisely, I think q1=-q2, right?) Are polarization artifacts expected to impact q1 and q2 identically in some way?

LN 384: Please clarify if 1.7% is the width of the PSL size distribution or uncertainty on the mean PSL diameter. If the former, was an uncertainty on mean diameter provided by the PSL manufacturer? If so, how much variation does that uncertainty translate to in terms of Mie calculated F11 and F12?

Sec 4.3: Would it be possible to further refine q, and characterize q1 and q2 separately, using the maxima and minima of the PSL measurements of F1 and F2? 

Eq 4: Is instrument stray-light background implicitly included here or is it not considered in the error model? 

LN 411: For consistency, maybe drop the squaring "2" on "σ^2_e,l" or add it to the other three variables.

Figure 5: I'm not see a section 3.5.3, as referenced in the caption. Please update to the correct section number.

LN 463: What was the reason the values 1.04 and 1.08 were assumed?

LN 493: This is the second time the blue lines are defined as best fits.

LN 534: Please explain how GRASP was tailored to produce the μNeph-GRASP Inversion. What changes were made relative to the standard build of GRASP? Also, for future work, it may be helpful to note that there was a high size parameter resolution version of the GRASP kernels developed to retrieve PSL properties in Espinosa et al. (2017). Those files are likely available by request from the GRASP team.

LN 549: I think "F11/F12" should be "F12/F11".

REFERENCES

Muñoz, O., Moreno, F., Guirado, D., Dabrowska, D. D., Volten, H., & Hovenier, J. W. (2012). The Amsterdam–Granada light scattering database. Journal of Quantitative Spectroscopy and Radiative Transfer, 113(7), 565-574.