The manuscript presents ambient measurements of Saharan dust intrusions at ground level in Southern Spain. The polarized imaging nephelometer enables the measurement of the phase function and polarized phase function and at several wavelengths at scattering angles between 5° and 175°. Especially the polarized phase function is an important quantity in passive remote sensing. They could show with their spectral measurements, that the polarized phase function behaves similar in the observed wavelength range (405 till 660 nm) for intense dust events. Whereas dust which was mixed with local pollution shows a different spectral behavior, especially at the shortest wavelength of 405 nm. Additionally, simulated (polarized) phase functions are shown. However, I see that for a publication in ACP, the research findings should be set into a broader context and the novelty of the study should be worked out stronger. After addressing my comments below, and please take my major comments serious, a publication in ACP can be recommended.

Major comments:

1. The paper would benefit a lot, if you include lidar observations to your dust cases. I know that the University of Granada operates a polarization lidar (otherwise I would not ask for it) which could clearly show the Saharan dust layers above the station (Sect. 3.1).
   
   Additionally, it may provide you with directly measured lidar ratios. Then, there is just the gap between the PI Nephelometer observations at the ground and the lidar measurements in the lofted Saharan dust layers. So please include your colleagues from the lidar group with their lidar observations at UGR.
2. Do you have additional size distributions for your cases? The spectral slope of your scattering properties might depend on particle size. Having an additional size distribution would add certainly a lot of value to your discussion. Could you estimate the amount of pollution in your polarization measurements? It would be an important information in quantifying your observations. By requesting additional size distributions and lidar (lidar ratio) observations underlines my request to broaden your field and take all available information into account to deliver a comprehensive picture and to place your results in a broader context.
3. You have not written a lot concerning your uncertainty estimates. In Tab. 1 you provide uncertainties without mentioning what you are reporting. I guess, it is the standard deviation of the hourly mean, but it is stated nowhere. If it is the case, I am wondering about the systematic uncertainties of your measurements. Probably, it is reported elsewhere. But we need an assessment of the systematic uncertainties. Otherwise the results are not comparable to other measurements.
   
   In Tab. 2 you don’t provide any uncertainties at all. Please add them.
   
   Your phase matrix elements (Fig. 6,7 & 9) do not contain any uncertainties. Putting an error bar to every point would certainly overload the figure, but having at least 3 points with error bars in each plot would help to assess the range of uncertainty.
4. After finalizing my review, I am now reading the comments of Jean-Baptiste Renard and I want to enforce his point, that the (spectral) scattering properties depend on particle size (see my major comment #2). And there is much more literature on the size effect of the polarized scattering properties than mentioned in his comment. The size effect is almost not discussed in the manuscript. Please try seriously to get more size information than just a PM10 concentration.

Minor comments:

• Title: “polarized polar imaging nephelometry” – in the manuscript, you always state Polarized Imaging Nephelometer (PI-Neph). Why do you use the term “polar” only in the title?
• L20 Please mention already in the introduction the angular range of the instrument.
• L56 a strange unit.
• L67 complex refractive index
• Your introduction (L70-85) is focused on passive remote sensing. I am missing the active remote sensing with e.g., CALIPSO or EarthCARE and the linked laboratory studies which focus on the backscattered light close to 180° (e.g., Sakai et al., 2010, Järvinen et al., 2016 or Miffre et al., 2023).
• Eq 5 + 6: Do you use the decadic logarithm (log) or the natural logarithm (ln) for your definition?
• According to eq. (7), the units of the LR should be sr and not sr-1. Please change it throughout the manuscript (including figures and tables).
• How do you extrapolate to the scattering angle of 180°. Please add some description and assessment of the related uncertainties.
• From the reader’s perspective, I would suggest a different section numbering to avoid the 4^th level of subsections (e.g., 3.2.1.1). My suggestions are:
  • Give the meteorological conditions (currently Sect. 3.1) an own section (Sect. 3) before you go to your results of the PI-Neph in Sect. 4 (Results or Aerosol phase matrix from different dust scenarios)
  • Section 3.2.3 and 3.2.4 can be combined to new section: Sect. 5 – Discussion with 5.1 and 5.2 for the two respective subsections.
• The meteorological conditions (Sect. 3.1) are given in great level of detail. I am wondering, if the 4 CAMS model outputs for each case are necessary, because this information is not used in the next sections. To my opinion, it can be combined to one CAMS output per dust case. In this section some lidar observations of the Granada station would be helpful to demonstrate the vertical layering of the dust above your station. It must not be a difference in the strength of the dust outbreak between the two cases in March, but on 15 March, the dust was mixed to a larger extend towards the ground and therefore to your PI-Nephelometer. Lidar observations would reveal the vertical layering of the dust.
• L200 Libya
• L250 Please provide the values for usual dust outbreaks and not just a reference. Similarly, in line 373
• Fig. 5 Please add more detailed steps to the time axis. The scale for SAE is not optimal.
• L329-334 This text might be moved to the figure caption of Fig. 6. At least Fig. 6 needs some more explanations in the caption. The same holds for L 374-377 and Fig. 7.
• L340 “notable spectral separation” – I am not sure if it just a manner of scaling the y-axis. Fig. 6a1 has a maximum value of 10^3 whereas the other subfigures extend to 10^4. Therefore the spectral separation is better visible in Fig. 6a1.
• Figs 6,7,9 lower row: the y-scaling is quite coarse, please add some ticks in between.
  
  You may add the value of the PM10 concentration for the different time steps because you are discussing it a lot and it is not always visible from Fig. 5+8.
  
  The term “moment” seems not the best choice, maybe better “instances” or “time steps” or something else. Please consider changing it here and in the text.
• L381-383 Sentence unclear.
• L387-389 Be more precise. The whole paragraph on page 11 needs some rephrasing to be more precise.
• Tab 1+2 “The angular range of the PI-Neph is used as the integration range of sigma_sca.” Please provide the angular range otherwise this information is not very helpful.
  
  Why are some values at 405 nm are missing on 25 March?
• L450: A newer and more comprehensive overview of lidar ratios was reported in Floutsi et al., AMT 2023.
• Fig 8. Sorry, but it is really challenging to get anything out of Fig. 8. Please use more space to show your results, half of a page minimum. Or remove the figure from the manuscript. I see only dots and can hardly infer any value from the y-axis. You probably want to show that there are more dust outbreaks during summer.
• Tab 2 and Fig. 9: Do you show the measurements for the whole day? Or for one hour? Or for the periods marked in Fig. 8?
• Fig 10 and surrounding text: BC and BrC are not defined.
• L539-541: Polarization measurements are very valuable for separating dust and non-dust contributions. This potential is used in the active remote sensing community for two decades now, starting with Shimizu et al., 2004, continuing to Tesche et al., 2009 and Mamouri and Ansmann 2017. You now adding PI-Nephelometer measurements to this separation.
• Fig 11 Do you expect that the seasonal average of -F12/F11 should go back to zero for 180° or could it stay negative?
  
  Please indicate the wavelengths for the Granada Amsterdam Light Scattering data base in the caption or even in the figure. By the way, 488 nm are much closer to 515 nm than to 405 nm. Why did you choose to show it together with the 405 nm?
  
  Yellow is probably not the best choice – could you choose a different color? And overall, the final publication should get the figure in a higher resolution. It is hard for me to distinguish the open circles from the stars.
  
  Caption: “top” instead of “up”
• Do you have a pure pollution case for comparison? It would certainly enhance the message to have a contrasting pollution case presented or repeated from Bazo et al., 2024 in this paper as well.
• Paragraph (L598-612): You already mentioned Teri et al., 2024. Here some more discussion to the polluted dust cases in the Eastern Mediterranean would be helpful. Overall, I would recommend to place your work in a broader context besides of previous measurements at Granada station.
• L654-657: Sentence unclear. Please rephrase to be more precise.
• Fig. 12 The negative values for the -F12/F11 close to 180° for the pure dust case are not seen in the observations of the extreme dust events. Is it a modelling artefact or is it missing in the observations? What is your explanation? Please discuss in your paper.
• L668: “To our knowledge, these are the first measurements of this type for ambient mineral dust transported to southern Europe” – only in Southern Europe. Where else ambient mineral dust measurements have been performed? What is the difference to your observations?
• L691 versus --> to
• L699 SSA and LR are intensive properties
• The author’s contributions section is missing.

References:

Järvinen, E.; Kemppinen, O.; Nousiainen, T.; Kociok, T.; Möhler, O.; Leisner, T. & Schnaiter, M.: Laboratory investigations of mineral dust near-backscattering depolarization ratios, Journal of Quantitative Spectroscopy and Radiative Transfer, 2016, 178, 192 – 208.

Floutsi, A. A.; Baars, H.; Engelmann, R.; Althausen, D.; Ansmann, A.; Bohlmann, S.; Heese, B.; Hofer, J.; Kanitz, T.; Haarig, M.; Ohneiser, K.; Radenz, M.; Seifert, P.; Skupin, A.; Yin, Z.; Abdullaev, S. F.; Komppula, M.; Filioglou, M.; Giannakaki, E.; Stachlewska, I. S.; Janicka, L.; Bortoli, D.; Marinou, E.; Amiridis, V.; Gialitaki, A.; Mamouri, R.-E.; Barja, B. & Wandinger, U.: DeLiAn -- a growing collection of depolarization ratio, lidar ratio and Ångström exponent for different aerosol types and mixtures from ground-based lidar observations, Atmospheric Measurement Techniques, 2023, 16, 2353-2379.

Mamouri, R.-E. & Ansmann, A.: Potential of polarization/Raman lidar to separate fine dust, coarse dust, maritime, and anthropogenic aerosol profiles, Atmospheric Measurement Techniques, 2017, 10, 3403-3427.

Miffre, A.; Cholleton, D.; Noël, C. & Rairoux, P.: Investigating the dependence of mineral dust depolarization on complex refractive index and size with a laboratory polarimeter at 180.0degree lidar backscattering angle, Atmospheric Measurement Techniques, 2023, 16, 403-417.

Sakai, T.; Nagai, T.; Zaizen, Y. & Mano, Y.: Backscattering linear depolarization ratio measurements of mineral, sea-salt, and ammonium sulfate particles simulated in a laboratory chamber, Appl. Opt., OSA, 2010, 49, 4441-4449.

Shimizu, A.; Sugimoto, N.; Matsui, I.; Arao, K.; Uno, I.; Murayama, T.; Kagawa, N.; Aoki, K.; Uchiyama, A. & Yamazaki, A.: Continuous observations of Asian dust and other aerosols by polarization lidars in China and Japan during ACE-Asia, Journal of Geophysical Research: Atmospheres, 2004, 109, D19S17.

Tesche, M.; Ansmann, A.; Müller, D.; Althausen, D.; Engelmann, R.; Freudenthaler, V. & Gross, S.: Vertically resolved separation of dust and smoke over Cape Verde using multiwavelength Raman and polarization lidars during Saharan Mineral Dust Experiment 2008, Journal of Geophysical Research: Atmospheres, 2009, 114, D13202.

The manuscript provides a summary of ambient measurements of F11 (phase function) and F12 made in Granada, Spain by the Airphoton/GRASP-Earth PIN100 as well as measurements of PM10 and a suite of other optical quantities. Two cases with very high concentrations of advected Saharan dust are explored in detailed followed by a summary of several cases with more moderate dust loading. Results are contextualized by single scattering calculations for spheres and spheroids showing theoretical F11 and F12 for hypothetical urban, mixed, and pure dust aerosols.

The complex shape of desert dust particles limits our ability to accurately predict the optical properties required in aerosol modeling and remote sensing but direct, ambient measurements of these quantities, especially phase function and other scattering matrix elements has been very limited. This paper presents some of the only ambient dust scattering matrix measurements to date and the results have great potential to help improve on these past limitations. The identification of the strong dependence in F12 to the fine mode, particular at shorter visible wavelengths is also a valuable contribution. I do however feel the manuscript fails to properly contextualize results, particularly in its accounting for potential measurement errors. At times it also seems to jump to very strong conclusions based on what seems to be limited or inconclusive evidence. To be suitable for publication in ACP, I feel that these issues, outlined in detail below, must be addressed. 

GENERAL COMMENTS

<> The exact method for obtaining several key optical quantities is not specified. Four aspects of the data were particularly unclear: 

(1) The PI-NEPH does not measure F11 at exact backscattering, but LR is provided throughout. How was F11(180°) estimated given the available data?

(2) The TSI integrating nephelometer is described and it is stated that "σsca can be also obtained with the integrating nephelometer" but it’s not clear if this data is ever used, or all σsca values are obtained via the PI-Neph. If the latter is the case, then there is no need to introduce the TSI instrument. 

(3) Equations 5 and 6 show general definitions for SAE and AAE but I don't think the exact wavelengths used are ever stated. Please specify the wavelengths used in these calculations.

(4) How were the particles sampled and did any aspect of the process have the potential to produced size dependent biases relative to the ambient aerosol. This is particularly import for the dust dominated scenes since a significant portion of dust particle volume concentration occurs in particles several microns in diameter that are easily lost to inlets and tubing.

<> At times measurements seem give nonphysical results. For example, -F12/F11 is unrealistically low in Figure 9b.2 and trends in AAE do not always match spectral trends in SSA (see specific comment – LN 467). Measurement errors are unavoidable, but I think these issues further emphasize the need to include well characterized uncertainties on the measurements, as the other reviewers have noted. 

<> In my view, some conclusions are not well supported, especially regarding relationships between optical features and the underlying microphysical properties of the aerosol (see specific comments for examples). The inclusion of additional measurements like those of size distribution or chemical composition, if available, could go a long way here towards proving or disproving some of these claims. Additionally, GRASP could be a very powerful tool for interrogating some of the relationships between optics and microphysics hypothesized, at least in the context of spheroidal particles. This seems to have been started in Section 3.2.4 but it was not clear to me why this analysis was not extended to actual inversions of the phase matrix (and possibly absorption) measurements.

<> I found the structure of the text to be a bit difficult to follow. Perhaps moving some sections towards to beginning may help provide a clearer "landscape" for the reader as they progress through the text. Moving section 3.2.3 higher up, or at least providing a clear and consistent definitions of the aerosol types observed/hypothesized earlier on would help a lot.

SPECIFIC COMMENTS 

Eq 1: I see the reason 2π and 4π were kept separate, but I wonder if reducing it to 1/2 would make this equation a little easier to read, especially given the improved consistency with Eq 4.

LN 140: This sentence is confusing. It is not clear that the single beam novelty they are referring to is the PIN-100. Also, it should be "University of Maryland, Baltimore County".

Eq 7: How is F11(180°) obtained? F11 has a lot of structure in the backscattering direction and linear extrapolation is probably not very accurate, especially for nonspherical particle like dust. The method for extrapolating should be some discussion of the potential resulting uncertainties should be provided.

LN 197: "pression" -> "pressure"

Figure 2/3: If all four subplots needed the color bars should be made consistent for all subplots and some discussion should be added to the text explaining what features differentiating the four subplots should be provided in the text. As it is, there is no significant difference that stands out to me between the four model times shown.

Figure 4: It would be good to include the name of the specific satellite/sensor that obtained these images.

LN 273: I assume this is a regulatory limit? This should be made clearer in the text.

LN 284: It is well established that nonspherical particles can have weaker scattering near scattering angles of 180° but that only contributes a very small amount to the total Bs which integrates from 90° to 180°. If there is a specific reference stating Bs decreases with particle nonsphericity that should be included here, otherwise I would suggest this sentence be removed or rephrased.

Figure 5: The time series for Event 1 should be extended back to at least 7:00 on March 15th to be consistent with F11 and F12 shown in Figure 6 and to give better context to the background conditions prior to the dust intrusion. 

Section 3.2.1.2: Because data at each angle in a PI-Neph is based on the sum of many individual pixels, saturation tends to happen gradually and result in a steadily increasing bias rather than a sharp, obvious upper limit in F11. The challenge of avoiding such saturation given an aerosol with a combination of a strong forward peak and very high concentration is not discussed in the text. Was each image checked to ensure that the counts of all pixels were below the upper limit of the detector's dynamic range? The text should include at least a brief explanation of the steps taken to ensure saturation did not bias the measurements.

LN 341: Again, it is not clear to me how the values of g and Bs on their own–which are generally accepted to have the strongest dependence to particle size–are able to clearly suggest the predominance of spherical vs nonspherical particles.

LN 343: It's worth noting that holding intensive properties constant but increasing the aerosol concentration will increase F11 at all angles. Given that the authors seem to be using the PI-Neph to interrogate changes in microphysics here, I wonder if P11 would be a better variable to directly analyze? Total scattering and spectral dependencies between the three wavelengths can easily be captured in σscat.

LN 348: "constant tendency" needs to be made more precise. It looks to me like F11(~180°) is an order of magnitude higher in panel c.1 relative to a.1 in Figure 6.

LN 349: While the spectral dependence is less, the red channel, and maybe the green too, appears lower than blue to me in Figure 6a.1. The language here should also be made more precise.

LN 381: Should this say, "increase in F11" rather than "decrease"?

LN 383: Many things can increase the forward peak of F11 and it is known to be particularly sensitive to aerosol size. However, it seems to be implied here that change in shape is the most probable cause. This should be explained. 

Figure 7: I think the caption should say "four different moments".

LN 425: I found this a bit hard to follow. Is the statement that some studies suggest that SAE is a useful discriminator between of the pure mineral dust and dusty mixtures with urban background pollution?

Table 1: Aside from the ± variability windows, is there any information in this table that is not already provided in Figure 5? If not, I wonder if those could be represented by bars in Figure 5 and the table could be moved to the supplement.

LN 443: "absorbing particles" should be made more precise. Is this referring specifically to black carbon? 

LN 463: AAE=1.71 is significantly higher than pure black carbon so it may be too strong to say that "absorption is mostly explained by black carbon".

LN 465: Again, it is not clear to me that g can be used as a proxy for non-sphericity.

LN 467: It should be noted that while the SSA measurements on July 5 may show stronger spectral dependence, the AAE measurement is actually lower on July 5 than Apr 16.

LN 469: It should probably be emphasized here that eBC is only equivalent BC for a given amount of absorption. Even at the longer wavelengths at which eBC is derived, very high levels of dust could produce significant absorption.

Figure 9: Blue -F12/F11 is much lower than any prior theoretical calculations (including those presented here in Figure 12) or measurements have shown. There are also some unnatural looking "kinks" in F11 around the same angular region. This emphasizes need for good uncertainty estimates and careful data screening.

LN 504: I'm not following this sentence. Why would high levels of noise mean that the measurement is "very sensitivity to aerosol mixtures"?

LN 506: This also left me confused on first read. Is the idea here that the blue channel of -F12/F11 is depressed in anthropogenic (or absorbing?) aerosol and the feature shows up at low dust concentrations when the dust is not optically dominate? 

LN 517: I think this sentence is redundant with the first sentence of this paragraph.

LN 540: I believe it is important to emphasize here that utility of 405 nm P12 as a marker for smaller, likely anthropogenic particles may be unique to the region/circumstances studied here. For example, because of the size invariance rule for single scattering properties, an aerosol that is identical but with diameters only ~27% (1 - 515nm/405nm) smaller would produce scattering patterns in the blue that are identical to what was measured in the 515 nm channel, which did not show this unique anthropogenic "fingerprint". This is not meant to diminish the finding, but instead to recommend that it be clearly caveated by the statement that this fingerprint may not be generalizable to all regions/seasons with dust intrusions. 

LN 603: Should this say, "blue wavelength"? I generally consider 400 nm to be the upper limit of the UV. Either way, the 473 nm measurements of Espinosa et al. (2017) were certainly not in the UV.

LN 655: I'm not sure what is meant by "optical properties" here. Either way, GRASP is certainly capable of simulating separate fine and coarse single scattering properties and refractive indices, as is evident in the forward calculation results show in Figure 12. I strongly suggest the authors include such inversions here.

LN 707: I still am not following how instrument noise can be linked with particle properties.