The mixing state of LAPs in snow has large impacts on LAP-induced snow albedo reduction and surface radiative forcing (RF). However, most land surface models assume that snow grain shape is spherical and LAPs are externally mixed with the snow grains. On this background, the authors improved the snow radiative transfer model in the land model (ELM v2.0) of the Energy Exascale Earth System Model version 2.0 (E3SM v2.0) by considering non-spherical snow grain shapes (i.e., spheroid, hexagonal plate and Koch snowflake) and internal mixing of dust-snow and systematically evaluates the impacts on surface energy and water balances over the Tibetan Plateau (TP). In general, this study is well written and can advance understanding of the role of snow grain shape and mixing state of LAP-snow in land surface processes and offer guidance for improving snow simulations and RF estimates in Earth system models under climate change. I think this manuscript can be accepted if address the following questions.

1. The authors spent much of this manuscript on comparing the modeling results with remote sensing snow data under the control simulation using the previous settings, which was not the main concern of this manuscript. This analysis only demonstrated the capability of the original land model in snow cover fraction or SAR simulation, but can not reflect the improvement of your work. So, I suggest the results based on the original land model and your improved version can be simultaneously compared with remote sensing data to highlight the improvement of your simulation.

2. In the Abstract, the authors said “Compared with two remote sensing ……, the control ELM simulation with the default settings of spherical snow grain shape, internal mixing of BC-snow, external mixing of dust-snow and without TOP can capture the overall snow distribution reasonably.” So my question is what is your contribution? From your abstract, I can not find better simulation results after improving these snow microphysics properties. I think the authors should highlight the advances in reproducing the satellite observations after the improvements.

3. The two used remote sensing data of ASR have larger differences themselves, so I wonder whether it is appropriate to use these two data in comparing with your simulations. Maybe only using MODSCAG/MODDRFS data is better from your results.

4. From the results, I found the snow grain shape and TOP have larger impacts on snow cover fraction or SAR simulation, while the impact of mixing sate is relatively lower. So my question is why you spent most of your content on snow grain shape and mixing sate, but the impact of TOP was only simply discussed in the last Section 4.4.

5. When discussing the impacts on water cycles, the authors only spent little content while this part was very important. In addition, this is no figure in the manuscript in this part. I think more discussions are needed because the impact of snow grain shape and mixing state of LAP-snow on SAR or RF have been fully discussed in previous studies while the impact on water cycles is little mentioned.

6. In the Discussion, the authors spent much time on discussing why the snow cover was underestimated in spring. As mentioned in the Discussion, the underestimation may be caused by the use of a constant snow accumulation ratio, empirical snowmelt shape factor, and the complex vegetation-snow interaction processes of snow interception and dynamical removal from canopy. So I think this was not closely related to the contributions of this manuscript. While the discussions from Lines 468-517 are the main limitations in your work and the directions the future work moves on. I suggest the authors can discuss the limitations in the first place, then simply discuss the underestimation.

This study presents an overview of the impact of several grain shapes, the mixing state of light absorbing impurities and sub-grid topographic effects on the surface energy and water cycles in the Tibetan Plateau using ELM. The authors illustrate that non-spherical shapes and sub-grid topographic effects impact the surface energy and water balance considerably during spring, while the impact of the mixing state of light absorbing impurities is smaller. The manuscript is well written and the topic is relevant. The impact of the mixing state of light absorbing impurities and topographic effects, and to a lesser degree also grain shape, is often overlooked in ESMs and climate models. This manuscript thus provides a necessary overview of the importance of these processes for such models.

The following comments should help with solving the remaining issues, with, for example, P1 L1 meaning page 1, line 1.  

General comments

• The comparison with satellite observations is limited to the control run only and could be expanded to some of the other experiments as well. It would be especially interesting to see how well ELM performs for the ‘combined effect’ experiment. Also, if the authors could make a figure showing the difference between model and observations could help visualizing the results.
• It is not clear to me what grain size is chosen in this study and why. Furthermore, it is also good to notice that a homogeneous snow grain size that does not vary in time, as is employed in ELM if I am not mistaken, could have a significant impact on the results. Wet and dry snow metamorphism, refreezing and the presence of slush or ice can all impact the grain size and albedo. This may explain for a large part the differences observed with remote sensing observations. Similarly, it is not clear to me what concentrations of LAPs are chosen.
• This study investigates several grain shapes. It is not clear to me, however, why these grain shapes are chosen in particular. Are these grain shapes often occurring on the Tibetan Plateau? If the authors could show that these shapes are relevant, it would strengthen the results of this study. Also, what grain shape will you assume in the final ELM version?
• In the manuscript, the authors show that the impact of TOP on the results is quite large and larger than the impact of the mixing state. This is in my opinion an important result of this work and should get more attention. It should get mentioned in the abstract and the conclusions as one of the main results. The impact of the mixing state of LAPs should not be overexaggerated in the abstract and conclusions as well.

Specific comments

P1 L18: “The mixing state of LAPs in snow also has large impacts…”. In the manuscript you show that the impact of the mixing state is considerably smaller than compared to grain shape and TOP. So please rephrase.

P1 L27: Please format units like W/m² with negative exponents, i.e., W m^-2, throughout the manuscript.

P3 L42: Please provide the definition of albedo here as well.

P3 L44: ‘Sky conditions’ is unclear

P3 L67: Define the asymmetry factor.

P3 L69: Define the single-scattering albedo.

P4 L75: In my opinion, more should be said about the physical processes that make the albedo drop because of dry and wet deposition. So why is there a difference between the impact of dry deposition and wet deposition. Also, do you expect dry or wet deposition to be most relevant for the Tibetan Plateau?

P4 L102: haven’t --> have not

P6 L140-141: “ 1) SZA dependence of surface irradiance”. This is a bit confusing, it suggests that you have implemented a new routine for SZA dependency, but if I understand it correctly, it is part of the SNICAR-AD model. Please clarify.

P6 L144: There are six types of atmospheric profiles now included, but are clouds considered as well or is it only for clear-sky conditions? As clouds alter the spectral distribution of irradiance and limit direct radiation, it often has quite a large impact on the broadband albedo.

P6 L147: Assuming that you will introduce the single-scattering albedo and asymmetry factor earlier, please also define the extinction cross section.

P6 L160: Why do you choose to investigate these three non-spherical grain shapes?

P7 L171: “… and Rs is the specific-projected-area-equivalent radius.” This has already been said and can be removed here.

P7 L181: “…where ω_dust and ω_p are single-scattering albedo of pure snow and dirty snow…”. I suppose it is vice versa? i.e., ω_p is for pure snow?

P9 L228: A dot is missing after the citation.

Table 1: What impurity concentration do you assume?

P10 L237: um --> μm

P10 L44: Why do you choose this simulation to be the control simulation? Is this the most realistic one? Please explain.

P10 L257-260: Could you elaborate a bit more on ANOVA? Also, what do you mean with ‘maximum absolute relative difference’ and ‘mean absolute relative difference’?

Fig. 2: For areas in the east, it is hard to see if there is a low snow cover or no snow at all. This can be solved by providing a separate color for no snow cover. Furthermore, The figure might be easier to read if discrete colors are used.

Fig 3 and all other boxplots. Please also state that what is on the x axis; i.e., that fsno and SAR as a function of elevation are shown.

Fig. 4: Same as Fig 2. Also, ‘Same as Figure 3’ is written, while it is the same as Fig. 2

P11 L277: “… although their spatial patterns are similar.” This is somewhat hard to see in the figures. A figure with the difference between ELM and the observations could help with that.

P12 L290: “The difference may be due to the overestimation of snow grain size…”. Can you explain a bit more why this may be the case? As grain size has a strong impact on the albedo, modelling this incorrectly could lead to large differences with observations, potentially overshadowing any grain shape or mixing state effects. It seems to me like it may have a large impact on the results here and should be mentioned.

P14 L316: “Western regions show larger RFs induced by all LAPs than the eastern regions”. Looking at Fig. 2 there is almost no snow cover in the east, so logically barely any RF changes are visible there.    

P14 L324: I assume you mean Table 3, not Table 1?

P14 L328: “… which is identical to” --> which is similar to.

P15 L341: Are all shown changes in Fig. 6 significant? If not, please also illustrate significance on the maps.

Caption of Fig. S4 and S5: I suppose you mean α_sno instead of A_sno?

Fig. 7: A bit confusing that the color scale is now inverted compared to Fig. 6. To help solve this confusion, please explicitly state in the manuscript that the differences are now negative.

P19 L389: “… has larger effects on F_lat than F_sen”. Why?

P19 L391 - 392: “The differences in … than those for Koch snowflake shape”. Can you explain why?. Similarly on P19 L395 – 396, please explain why you would expect a sign change.

P21 L417 – 418: One of the figure references can be omitted.  

P23 L455 – 456: “… to the reported values in the existing study”. What do you mean? It is not clear to me what the authors want to say here.

P23 L459 – P24 L466: “However, different studies … or annual scales”. I am not sure if I understand what you want to say with this part.

P24 L470: “Mixing state of LAP-snow also has large impacts on SAR”. Looking at the results of this manuscript, it looks like the impact is considerable smaller than the impact of grain shape and TOP. Please rephrase.

This study incorporates several new features into the representation of snow albedo in the E3SM model, including non-spherical snow grains, internally-mixed dust and black carbon within ice grains, and sub-grid topographic effects.  The study assesses the impacts of these improvements on the simulated snow cover and surface energy budget, including impacts of individual effects as well as the combined effects of all model changes operating simultaneously.  Overall, the paper is well-organized and very well-written.  I have only a few minor comments, and I recommend that the manuscript be published after these are addressed.

Minor comments:

Section 4.3: In the sensitivity studies that apply internal mixing of LAPs, are all of the particles assumed to be internally-mixed, or only the proportion of deposited aerosol that was simulated to be internally-mixed?  (And is that proportion actually simulated?  My understanding is that the MAM aerosol model simulates the mixing state of aerosols, so that information could, in principle, be utilized, but it is not clear that such information is being extracted and utilized in the model experiments).  Either way, please expand the discussion on this issue, including any implications for the magnitude of impact assessed in the sensitivity studies.

line 115: "... control ELM simulation with the default settings..." - Please clarify whether the default settings are the old or new default settings, i.e. with or without all of the changes described in this study.  The use of "control" would generally imply the original configuration, before cay changes are implemented.

Line 318: "... because LAP-induced SAR is larger in spring ..." - While the model SAR is definitely larger in spring than winter, it is not totally clear from Figs 3 and 4 that the *observed* SAR is larger in spring.  (And I am surprised that it is not, given the likelihood of melt-induced surface accumulation during spring and higher frequency of springtime dust storms in this region).  Please comment on this.

Line 394: "Overall, the effects of mixing state of dust-snow are smaller than the effects of mixing state of BC-snow." - Why?  Please explain.

Line 511: "... but the real snow grain shape may be more complicated and irregular..." - I think you can safely say that the real snow grain shape *is* more complicated and irregular.