Comment to “Model-based insights into aerosol perturbation on pristine continental convective precipitation”

With WRF simulations, this study investigates the aerosol impacts on pristine continental convective precipitation over Naqu, China. It is found that under relatively polluted conditions, the onset of precipitation is delayed with increased cold rain intensity. The finding is interesting and worthy for publication.

In this study, different amount of aerosols have been considered and analyzed, I wonder if the surface topography (particularly TP region) along with aerosol types could play a role and analysis of these effects could be highly valuable. Of course, considering the focus of this study, these aspects could also be done in future by simply stating them in the final section.

Line 50, “Due to”

Line 50-52, Ma et al. (2018, doi: 10.1002/2017MS001234) might be able to serve as a support for this claim.

Line 54-57 and 88-90, A recent three-pole (Arctic, Antarctic and TP) aerosol characteristics study shows the aerosol type comparison results, including the aerosol types and sources over the TP, which is worthy to mention, Yang et al. (2021, doi: 10.5194/acp-21-4849-2021).

Line 66-67, This is true, while adding the reason could be more appreciated, such as “due to high sensible heat and low air density”.

Line 84-86, A reference is necessary for this result.

Line 103, “in the TP”

Line 142, “pristine continental environment”?

Line 145, “be regarded a a …” should be “be regarded as a …”

Line 205-207, It might be better to indicate the unit too.

Line 209-211, Actually, some other types of aerosols could also serve as INP, thus I would suggest adding “in model” here to constrain.

Line 231-276, What is the spin-up time for the simulations?

Line 304-306, Considering the potential uncertainties, one digit number might be enough for the increase.

Line 322, “occurs” should be “is”

Line 351, This is not accurate. Early description said 2 times concentrations, while here shows “2 times more abundant than …”

Line 365-366, Particularly over the Pearl River Delta region as a few studies indicated, which are worthy to cite. Note that this should be also related to aerosol types.

The authors have used model simulations driven by reanalysis to examine aerosol effects on convection over the Tibetan Plateau, an area with very low concentrations of aerosols normally. Their findings are consistent with other works, showing that generally with increased aerosols they see delayed warm rain production and then enhanced cold rain processes with some convective invigoration. I see no major flaws with the work, I just have some suggestions below for clarification and better readability. 

Overall: 

Throughout there are some typos and small language/grammatical issues that require some copy-editing. 

I would suggest adding subtitles to figures, it's easier to follow that way.

Specific:

Abstract is a bit confusing, what does this sentence mean?

"With the increase in the aerosol number concentration, the conversion of cloud water to rain in clouds is first enhanced."

First according to what? 

Line 127: The sentence starting with 'The aerosol impact in..' was confusing to read. I think you are just trying to state that impacts on convection can then effect teleconnections through the heat pump? Also, the reference is the 2007 paper, not 2016.

Line 146: 'This methodology could then be applied in other regions of the world with similar background environments.' I'm not sure which methodology you mean - just using model simulations driven by reanalysis? I'm certain this has been done numerous times in various regions of the globe. 

Line 236: Do you not have observations of convection and/or precipitation? Why do you need to compare the soundings to 'suggest' that convection developed in this time period?

Line 253: 1km and 60s are a bit coarse. (especially the 60s, was that truly used for the 1km grid?)

Is there a reason this particular date/case was chosen? 

What is the default value from the model and where does it come from? Why not perturb the observed aerosol value instead?

Line 299: You mention radar here, but why is this not shown anywhere? How do the simulated storms compared to what was observed?

Fig 5: I'm not sure how to interpret a cloud field that's averaged over 12 hours. Physically this doesn't make a lot of sense, as over this period of time advection, microphysics, and precipitation are all occurring. 

Line 302: You say 'indicating that the warm cloud process was dominant', but these are model simulations, and you should be able to easily check to see what processes are dominant. 

Line 340: You are repeating yourself from the previous paragraph when you describe the figure here. 

Line 356: I'm not sure I follow this argument. Might it just be the case that the extremely pristine environment is aerosol-limited?

Figure 9: It might be easier to see differences if this was just one plot. 

Line 401: These w averages are over a large domain, over 5 hours. There are different cloud/storm amounts in each simulation. It's not clear that you can interpret anything real from this plot, it would just be averaging over anything interesting. I would suggest looking only at updrafts, or use a condensate threshold, to be more sure this is actual invigoration. 

The conclusion feels kind of sudden and a bit non sequitur. There is talk of uncertainties in measurement, but not really in context of the study. It just seems kind of incomplete. 

This study investigated the possible aerosol impacts on the small-scale convection and related precipitation event in Tibetan Plateau (TP) by conducting modeling sensitivity simulations. The analysis on cloud response to aerosol perturbations are generally sound in this study and the evidences shown to support their conclusions are overall sufficient. Given that there are still limited studies on aerosol-cloud-precipitation interactions over pristine continental regions like TP, I recommend it for publication but with some further clarifications of aerosol initial condition settings (or assumptions) prior to publication.

General comments:

1. Lines 134-135: Could the authors elaborate a little more about the motivations on studying the small-scale convection and precipitation event which occurred in Naqu region? Was there any unique characteristics of the precipitation event selected for study? For example, was there aerosol intrusion during the precipitation event? How frequent and how important of the small-scale convections related precipitation in TP region?

2. Below I have some concerns about aerosol condition setups when conducting simulations:

Lines 267-271: How better the aerosol field adopted in the TP case can represent the real situation? And how was the aerosol vertical distribution set, which is also believed of importance in determining the CCNs involved in cloud development? It’s better to use available observations to validate the aerosol setups used in this study. One option the authors might consider is to use satellite retrieved AOD as reference.

Regarding the aerosol perturbation scenarios, Table 4 describes that the Clean case was set with 1/10 CCN number concentration at the basis of Control case, but from Fig. 4 it shows they are comparable in aerosol number concentration. So, could you explain the inconsistency between Fig. 4 and Table 4?

IN may also play a role in aerosol-cloud-precipitation. Are there any perturbations in IN loadings among different experiments?

3. Is there any influence on surface precipitation types due to either CCN or IN effects? That is, what are the relative contributions by rain and snow to surface precipitation and/or their temporal/spatial variations?

Minor comments:

Line 50: ‘Due’ -> ‘Due to’.

Fig. 3: What are A and B for? Please denote it in the caption.

Fig. 4: Please use the same color bar to better view the difference between the four cases.

Fig. 6: What are the possible reasons the TP case overestimated the precipitation amount observed?

Fig. 9: How about put all the Z-wind profiles in one panel so that the readers can more easily view the differences among different cases?