Lorian and Dagan, 2023 uses idealized simulations in radiative-convective equilibrium to study the effects of changing aerosol concentrations on radiation and precipitation at different sea surface temperatures. The authors find that while the effects on radiation are qualitatively similar at all temperatures, the effect decreases with increasing SST. The manuscript attributes the decrease in SST-mediated changes in aerosol-cloud interaction to differential aerosol effects on warm rain formation and cloud liquid in lower parts of deep convective clouds. These become deeper at higher SSTs, leading to different responses. Another interesting result I want to highlight here is the anvil cloud fraction and ice water path decrease with increasing aerosol number, that is caused by increased latent heating and consequently increased upper tropospheric stability, following the "stability iris" hypothesis.

This manuscript presents several novel results on a very important and understudied topic. However, I am not convinced about some of the proposed explanations. In particular, I remain skeptical about the radiative relevance of changes in the warm parts of deep convective clouds. Can these really modify changes in radiative fluxes at the top of the atmosphere, given that they lie below very reflective thick anvil clouds? Furthermore, my main suggestion to the authors would be to go beyond the domain-average perspective and try to decompose changes in different cloud types (for example active deep convective cores, thick anvils, thin anvils, low clouds).

Overall, I found the manuscript to be an interesting read, full of noteworthy insights. However, I would strongly encourage that the authors to consider further clarification or even revision of some of their key findings to improve the overall quality of their work prior to final publication. For a more detailed discussion, please see my extended comments below.

General comments:

1. Regions dominated by deep convective lifecycle that are at/near conditions of radiative-convective equilibrium are typically dominated by anvil clouds, both in terms of coverage as well as radiative impact (e.g. Berry and Mace, 2014). The radiative importance of deep convective cores is, despite their strong LW and SW radiative effects, relatively small due to their small coverage. Changes in their properties/frequency are therefore unlikely to substantially modify climatological CRE. Moreover, given the large IWP of deep convective cores (see e.g. Fig. 1 in Sokol and Hartmann, 2020), I find it hard to believe that changes in the warm parts of deep convective clouds can directly influence TOA radiation. Couldn't changes in TOA radiative fluxes be entirely related to changes in anvil clouds?
2. On the other hand, even in disaggregated RCE simulations, there is a small but non-negligible population of stratiform low clouds as shown in Figure 1a. What fraction of changes in cloud liquid, rain, and their radiative implications can be attributed to changes in stratiform liquid clouds?

→As mentioned above, it is difficult to answer my two key questions without digging deeper into the model output, beyond the domain averages. My first thought would be to perform an analysis of the 2D SAM model output fields and subdivide the domain into, for example, areas of active deep convection, thick anvils, thin anvils, and low clouds. This could be done, for example, based on the diagnosed ice water path fields. In addition, it may be useful to look at changes in COD/in-cloud CRE to distinguish between effects on cloud fraction and cloud opacity.

1. Figures 3a and 6h make me wonder if the effect of aerosols on clouds also affects cloud height, cloud top temperature, and thus the radiative budget. High aerosol simulations seem to lead to higher, colder clouds. This may be a radiatively important mechanism that should be discussed. Is the decrease in cloud top temperature strong enough to significantly affect the LW CRE?
2. While this is clearly an idealized study, I am not convinced that using the N_a of 20/cc is the best choice. To my knowledge, the Na should be higher even under very clean conditions.
3. Could lower aerosol radiative sensitivities at warmer temperatures be explained simply by an overall reduction in clouds (both anvil and low cloud fractions decrease at warmer temperatures), and thus a reduction in the domain-averaged effects of aerosol-cloud interactions?
4. Intuitively, I would expect more smaller ice crystals in polluted deep convective clouds. These would, at least holding everything else constant, lead to an increase in the anvil cloud fraction. Do anvil microphysical properties change under high aerosol conditions?
   
   If the above is true, these changes seem to play only a second order role compared to the stability iris mechanism.
   
   Does this mean that the aerosol-cloud interaction community shouldn't overemphasize the direct effects of aerosols on cloud properties, but rather think about cloud adaptations?
   
   Moreover, the thermodynamic changes that drive these adjustments may be easier to understand (or at least more accessible to a broader community). Another implication may be that we shouldn't worry too much about getting all the microphysical details right.
5. Several plots should be improved:
   
   -Fig. 1 and 4: I suggest adding markers for better line visibility (or thicker lines, or both)
   
   -Fig. 2,3,5,8,9: add zero line where appropriate (where anomalies cross it)
   
   -Fig. 3,5: please use same x-axis limits if possible

Specific comments:

Page 1, line 5:

 “decline in TOA energy gain”

I think this can be written in simpler words

Page 2, line 57:

I suggest to remove  “general”

Page 3, line 80:

Something is odd/missing there.

Page 4, line 89-90:

Changes in ozone heating could also influence anvils (e.g. Harrop and Hartmann, 2012; Seidel and Yang, 2022)

General comment on the introduction: are there similar studies that looked at aerosol impacts in RCE? If yes, please mention them, if not, please state that this is the first one looking at it.

Page 4, Model description:

Anvils are ice phase clouds that strongly depend also on how freezing is parameterized in the model. Please add information also about the ice phase microphysics!

Page 4, Experimental design:

Is the ozone heating profile fixed? If so, I suspect anvils may already be influenced by it in the warmest experiment.

Page 6, line 141:

“clouds become thicker”  Do the authors mean thicker in optical sense, or simply that their vertical extent is larger?



Page 7 & 8, Fig. 2 & 3:

With the upward shift in clouds a kg of air covers a larger volume. The increase of cloud ice/upper tropospheric cloud water may therefore not be consistent with the changes in integrated amount of ice. Does ice water path change across the investigated range of SSTs?

Page 9, lines 193-194:

Doo changes in graupel production really explain the warming in the upper troposphere? Peak warming seems to be at higher altitudes.

Page 10, lines 205-208:

It may be useful to decompose the within atmospheric heating to the cloudy and clear-sky parts. Or just clearly state that the total heating and the anomalies described here are driven by clear-sky radiative cooling.

Page 11, Fig. 9:

How should we interpret the increase in advective tendency of the liquid/ice water static energy?

Page 16:

The authors could mention two more caveats:

- the assumed sensitivities do not include sensitivities related to changes in ice nucleation and freezing. A polluted environment would most likely lead not only to large numbers of CCN, but also to large numbers of ice nucleating particles and/or ice nucleating particles that freeze at warmer temperatures. This could lead to significant climate impacts that could easily match those described in this paper.

-Although tropical oceanic convection doesn't have as strong a diurnal cycle as land convection, its effects cannot be neglected. If the main aerosol-mediated changes in SW radiation are indeed those in deep convective cores, they may be muted in simulations with a diurnal cycle due to the early morning peak in deep convective activity (e.g. Nesbitt and Zipser, 2003; Gasparini et al. 2022).

References:

Berry and Mace, 2014; https://doi.org/10.1002/2014JD021458

Gasparini et al., 2022; https://doi.org/10.1175/JCLI-D-21-0211.1

Harrop and Hartmann, 2012; https://doi.org/10.1175/JCLI-D-11-00445.1

Nesbitt and Zipser, 2003; https://doi.org/10. 1175/1520-0442-16.10.1456  

Seidel and Yang, 2022; https://doi.org/10.1175/JCLI-D-21-0962.1

Sokol and Hartmann, 2020; https://doi.org/10.1029/2020JD033107

Best regards,

Blaž Gasparini

Please see attached file.