This article describes the implementation of dust mineralogy in a regional model, COSMO5.05-MUSCAT, and presents an overall evaluation of the results. This implementation constitutes a first approach towards a more integrated representation of the dust mineralogy and its impacts, e.g., through the refinement of the definition of the optical properties.  As such, the article addresses a relevant topic for the atmospheric and climate modeling communities and merits publication. However, in my view, some methodological aspects deserve a more detailed explanation and part of the highlights in the abstract and conclusions could be further clarified. 

General comments:

• Airborne minerals particle size distribution

The authors assume the size distribution of the minerals reported in the soil mineralogy map of Nickovic et al. (2012) as equivalent to that in the airborne particles. Observational evidence (e.g., Kandler et al., 2009) show that phyllosilicates are often found in airborne dust in coarser sizes than those reported in the soil maps. The implications of that assumption for the size-resolved airborne mineralogy have been discussed in previous works (e.g., Perlwitz et al., 2015a,b, Pérez García-Pando et al., 2016, Gonçalves Ageitos et al., 2023) and a number of modeling studies include some form of adjustment between the soil mineral fractions and those in the aerosol (e.g., Scanza et al., 2015; Perlwitz et al., 2015a,b; Ito and Shi, 2016; Li et al., 2021; Gonçalves Ageitos et al., 2023). In my view, the authors should justify their choice to define the size distribution of the airborne minerals and further discuss its impact on their results throughout the article. 

• Evaluation of the dust mineralogy

The evaluation section presents a comparison of the modeled dust optical properties against different products and retrievals. This evaluation is relevant to prove the model’s ability to represent the dust cycle (a necessary step to reproduce the dust mineralogy), but it could be shortened and/or included as supplementary information. 

The focus should be put on the evaluation of the mineralogy, either through direct measurements or the use of mineralogy-sensitive optical properties (e.g., single scattering albedo).  Furthermore, the mineral observations from in-situ data reported in Appendix A include information on aerosol samples with various size ranges, while the caption on figure 5 suggests that the scatterplots only consider bulk aerosol measurements. A size-collocated evaluation of the mineral fractions would increase the number of data available and provide relevant information on the ability of the model to reproduce the mineral content in different size ranges. Something that, on the other hand,  could be linked to the assumed size distribution at emission (see the comment above). 

Finally, the authors present a comparison of the dust vertical profile with LIDAR products, where they add the vertical profile of the modeled mineralogy. While they hypothesize that considering explicitly hematite would lead to better agreement with observations, this is not proved in their case study. I would suggest to clearly acknowledge this is the abstract and conclusions where they highlight this hypothesis. 

• Mixing state of the iron oxides and mass density

COMO-MUSCAT uses a representation of the different minerals as external mixtures. In the case of iron oxides, previous works suggest that these minerals are often found as accretions in other mineral particles (e.g., Kandler et al., 2009). Iron oxides have greater mass densities than other dust mineral components, which would make their lifetime in the atmosphere shorter. The methods section should clarify which is the assumed mass density for the different minerals, and particularly for the iron oxides. There are a couple of remarks in the conclusions related to this aspect, but it is unclear to the reader how this was dealt with in COSMO-MUSCAT.  

Gonçalves Ageitos, M., Obiso, V., Miller, R. L., Jorba, O., Klose, M., Dawson, M., Balkanski, Y., Perlwitz, J., Basart, S., Di Tomaso, E., Escribano, J., Macchia, F., Montané, G., Mahowald, N. M., Green, R. O., Thompson, D. R., & Pérez García-Pando, C. (2023). Atmos. Chem. Phys. 23(15), 8623–8657. https://doi.org/10.5194/acp-23-8623-2023

Ito, A., & Shi, Z. (2016). Delivery of anthropogenic bioavailable iron from mineral dust and combustion aerosols to the ocean. Atmos. Chem. Phys. 16(1),  85–99. https://doi.org/10.5194/acp-16-85-2016

Kandler, K., Schütz, L., Deutscher, C., Ebert, M., Hofmann, H., Jäckel, S., Jaenicke, R., Knippertz, P., Lieke, K., Massling, A., Petzold, A., Schladitz, A., Weinzierl, B., Wiedensohler, A., Zorn, S., & Weinbruch, S. (2009). Size distribution, mass concentration, chemical and mineralogical composition and derived optical parameters of the boundary layer aerosol at Tinfou, Morocco, during SAMUM 2006. Tellus B: Chem. Phys. Met. 61(1), 32-50. https://doi.org/10.1111/j.1600-0889.2008.00385.x

Li, L., Mahowald, N., Miller, R., Pérez García-Pando, C., Klose, M., Hamilton, D., Gonçalves Ageitos, M., Ginoux, P., Balkanski, Y., Green, R., Kalashnikova, O., Kok, J., Obiso, V., Paynter, D., & Thompson, D. (2021). Quantifying the range of the dust direct radiative effect due to source mineralogy uncertainty. Atmos. Chem. Phys. 21,  3973–4005. https://doi.org/10.5194/acp-21-3973-2021

Nickovic, S., Vukovic, A., Vujadinovic, M., Djurdjevic, V., & Pejanovic, G. (2012). Technical Note: High-resolution mineralogical database of dust-productive soils for atmospheric dust modeling. 12(2), 845–855. https://doi.org/10.5194/acp-12-845-2012

Pérez García-Pando, C., Miller, R. L., Perlwitz, J. P., Rodríguez, S., & Prospero, J. M. (2016). Predicting the mineral composition of dust aerosols: Insights from elemental composition measured at the Izaña Observatory. Geo. Res. Lett. 43 (19), 10,520-10,529. https://doi.org/10.1002/2016GL069873

Perlwitz, J. P., Pérez García-Pando, C., & Miller, R. L. (2015a). Predicting the mineral composition of dust aerosols - Part 1: Representing key processes. Atmos. Chem. Phys. 15(20). 11593–11627. https://doi.org/10.5194/acp-15-11593-2015

Perlwitz, J. P., Pérez García-Pando, C., & Miller, R. L. (2015b). Predicting the mineral composition of dust aerosols - Part 2: Model evaluation and identification of key processes with observations. Atmos. Chem. Phys. 15(20), 11629–11652. https://doi.org/10.5194/acp-15-11629-2015

Scanza, R. A., Mahowald, N., Ghan, S., Zender, C. S., Kok, J. F., Liu, X., Zhang, Y., & Albani, S. (2015). Modeling dust as component minerals in the Community Atmosphere Model: Development of framework and impact on radiative forcing. Atmos. Chem. Phys. 15(1), 537-561, https://doi.org/10.5194/acp-15-537-2015

Specific comments

L8,9 - How is the improvement of model performance related to the introduction of the mineralogy? 

L13,15 - Please, clarify how your results back up the link between hematite and the improved interpretation of the LIDAR product. 

L16 - Why does the comparison with in-situ measurements show how important they are?

L27 - Jickells et al. (2005) focuses on dust as a source of iron for marine ecosystems, rather than its direct radiative effect.

L29 - Chatziparaschos et al. (2023) is a modeling study. It focuses on k-feldspar and quartz as INPs. One could cite Harrison et al., 2019 as an observationally based study: 

Harrison, A. D., Lever, K., Sanchez-Marroquin, A., Holden, M. A., Whale, T. F., Tarn, M. D., Mcquaid, J. B., & Murray, B. J. (2019). The ice-nucleating ability of quartz immersed in water and its atmospheric importance compared to K-feldspar. Atmos. Chem. Phys, 19, 11343–11361. https://doi.org/10.5194/acp-19-11343-2019.

L44,46 - Specify the source of the soil types information as FAO classification. Nickovic et al. (2012) did not include additional mineralogy measurements, but added the phosphorus content of the soils (an element present in different minerals). Please, rephrase.   

L55 - There are previous works implementing mineralogy in regional models: 

Menut, L., Siour, G., Bessagnet, B., Couvidat, F., Journet, E., Balkanski, Y., & Desboeufs, K. (2020). Modelling the mineralogical composition and solubility of mineral dust in the Mediterranean area with CHIMERE 2017r4. Geosci. Model Dev, 13(4), 2051–2071. https://doi.org/10.5194/gmd-13-2051-2020

L56 - There’s something missing in the sentence. 

L66,77 - There are multiple modeling exercises that consider a different mineral particle size distribution in the aerosol than in the soil. Some examples are: Scanza et al. (2015), Li et al. (2021) for the CAM model, Gonçalves Ageitos et al. (2023) for MONARCH, Chatziparaschos et al. (2023) for TM4, Myriokefalitakis et al. (2021) for EC-Earth3, Ito et al. (2016) for IMPACT, etc.  

L78 - I would recommend stating the main objectives of the paper before describing the content of each section. 

L176 - “Effective fractions of minerals in soils are determined by combining soil texture classes and applying modifications derived from modelling approaches.” This is not clear. Please, clarify. 

L179 - Phosphorous is not a mineral, and as far as I understand it has not been included in COSMO-MUSCAT. Please, clarify. 

L186,187 - See my general comment above. This assumption must be justified in view of previous evidence and or its implications discussed more thoroughly in the article. 

L191,193 - How does the GMINER dataset consider the mineralogical composition changes during the emission process?  “By only taking into the account the soil mineralogical composition of the particle classes that would be emitted, that being, silt and clay sizes.” This sentence is incomplete. Please, if it refers to the previous sentence, rephrase and clarify. The fact that the soil mineralogy encompasses the clay and silt size ranges does not mean that the size distribution within these particle sizes measured in wet sieved soils would correspond to that of the emitted aerosol. 

L199 - “That causes a larger allocation of mineral fractions to clay sized populations than could exist in undisturbed soil.” Please, clarify if this is the case for all minerals. Also, refer to the potential implications of this in your modeling study. 

L205,206 - “Results were later supported by Perlwitz et al. (2015b) who found that the particle composition of the clay sized emitted minerals is identical to that of the fully dispersed soil as given by Claquin et al. (1999).” Please, review this sentence. Perlwitz et al. (2015) found that reaggregation before emission improves the modelled aerosol mineral fractions below 2 um (e.g., for feldspars). 

L213,214 - The mentioned works do not present an evaluation of the size distributed mineralogy of airborne dust particles against mineralogy measurements. Atkinson et al. (2013) and Journet et al. (2014) do not show any evaluation of the modeled mineralogy. Hoose et al. (2008) shows an evaluation of phyllosilicate mass fractions (i.e. kaolinite, illite/smectite) without considering the particle sizes and the modeled results are not particularly correlated with the observations.

L231 - It has been shown that iron oxides (e.g., hematite) are usually internally mixed with other minerals (in the form of accretions in the surface of the mineral). This mixing has implications in terms of the transport (i.e., the hematite particles are much denser than the average dust particles, therefore if transported as externally mixed, they would be removed from the atmosphere efficiently close to sources). How are these aspects treated in COSMO-MUSCAT? Discuss how the external mixing assumption could affect the results shown.

L240 - The evaluation of the modeled AOT against AERONET retrievals provides relevant information on the COSMO-MUSCAT ability to reproduce the dust cycle (see my general comment above). The authors could consider complementing this evaluation with single scattering albedo (SSA) from AERONET, which is sensitive to the mineralogical composition. Besides the selection of stations close to dust sources, other AERONET parameters could be used to identify retrievals dominated by dust.

L279,280 - Please, specify the link between the LIDAR data and the mineral resolved emissions of dust.

L408 - Please, see my general comment above. I would recommend adding to this comparison the size-dependent mineralogy evaluation. Also, if possible, I would include the observations in the x-axis and the model in the y-axis in Figure 5. The interpretation of the figure would be then clearer with values above the 1:1 line representing an overestimation, and below, an underestimation. 

L412 - I would recommend adding some quantitative metric to back up the assessment, e.g. “good agreement with measurements”. The number of points used for the evaluation is also relevant when interpreting the results. 

L513 - See my general comment above.

L569,574 - Balkanski et al. (2007) focuses on the dust absorption, rather than the evaluation of the AOT.   

L576 - Please, clarify why the measured quartz and feldspars are less reliable than other minerals’ content. 

L596,597 - See my comment above (L55).

L612 - See my comments above (L240 and L569-574).

L622,625 - See my comments above. These aspects must be clarified and justified in the methods section.

Technical corrections

There are some acronyms in the text that are not defined the first time they appear. I would also recommend the authors to review that the references are appropriately included (and review the format in the bibliography). 

The authors present a study describing the implementation of dust mineralogy in COSMO5.05-MUSCAT regional model. The results are compared with lidar, satellite, AERONET measurements and dust composition data from literature. This study is particularly relevant because there aren't many models that consider the dust mineralogy. I have a concern with respect to methodology and a few other points that should be addressed before publication:

1. In the abstract, introduction and conclusion, the authors wrote that this study is the first implementation of explicit representation of dust mineralogy in regional model. However, there is at least one older reference (Menut et al., 2020) dealing with this topic in regional model. Please add the reference.
2. Even if the simulations are compared with measurements and show relatively good results, how to know if the total amount of dust is conserved with mineralogy and with no mineralogy description? Is it possible to add a reference simulation with no mineralogy to estimate the potential benefit of this development?
3. I do not understand why you could consider the same relative part of silt and clay from the soil to the aerosol. There is a lot of literature available to take into account the relative part of clay and silt from soil to aerosol (eg: Scanza et al., 2015 simplified in Menut et al., 2020, Gonçalves Ageitos et al., 2023…)
4. In the Methodology you discuss model configuration and emissions, but you don't mention deposition processes? How are minerals deposited?
5. Finally, you used the same density for each mineral and you used a fixed composition to calculate Q_{ext, 500nm} for each size class. How far is this fixed composition from the simulated one? Can you explain why do you choose to do that? It would seem relatively easy to run sensitivity tests taking into account the simulated composition range.
6. For the validation, you used only 5 AERONET stations. In fact there are more in your simulation area. Why don’t you use all available data? How do you choose your stations? Can you plot the Ångström coefficient to be sure they're mostly dust?
7. The validation is almost complete (lidar, satellite, AERONET measurements and dust composition from literature). Only mass concentration at the ground is not use. It could be interesting to use INDAAF data (https://indaaf.obs-mip.fr/catalogue) to compare your simulation with the PM₁₀
8. What is the cost in computation time of implementing this level of detail?
9. Can the mineralogy representation be used with activated chemistry? Are there any impacts of this representation on heterogeneous reaction?

References

Gonçalves Ageitos, M., Obiso, V., Miller, R. L., Jorba, O., Klose, M., Dawson, M., Balkanski, Y., Perlwitz, J., Basart, S., Di Tomaso, E., Escribano, J., Macchia, F., Montané, G., Mahowald, N. M., Green, R. O., Thompson, D. R., & Pérez García-Pando, C. (2023). Atmos. Chem. Phys. 23(15), 8623–8657. https://doi.org/10.5194/acp-23-8623-2023

Menut, L., Siour, G., Bessagnet, B., Couvidat, F., Journet, E., Balkanski, Y., & Desboeufs, K. (2020). Modelling the mineralogical composition and solubility of mineral dust in the Mediterranean area with CHIMERE 2017r4. Geosci. Model Dev, 13(4), 2051–2071. https://doi.org/10.5194/gmd-13-2051-2020

Scanza, R. A., Mahowald, N., Ghan, S., Zender, C. S., Kok, J. F., Liu, X., Zhang, Y., and Albani, S.: Modeling dust as component minerals in the Community Atmosphere Model: development of framework and impact on radiative forcing, Atmos. Chem. Phys., 15, 537–561, https://doi.org/10.5194/acp-15-537-2015