See the attached PDF for my comments.

“Investigating the impact of subgrid-scale aerosol-cloud interaction on mesoscale meteorology prediction” by Wenjie Zhang et al.

The authors implemented a double-moment microphysics scheme into the KFeta convective parameterization of the CMA_Meso5.1/CUACE mesoscale atmospheric chemistry model to investigate subgrid-scale aerosol-cloud interactions (ACI) during the summer over central and eastern China. The authors conclude that the forecast improvements in surface downward shortwave radiation, near-surface temperature and humidity, and precipitation are due to the effects of subgrid-scale ACI. There are concerns regarding the physics configuration and experimental design, which are somewhat unclear. After carefully reviewing the model configuration in this manuscript and its references, I believe that the differences between the ACIsub and NO-ACIsub simulations stem not only from subgrid-scale aerosol-cloud interactions but also from subgrid-scale microphysical processes. This may mislead readers into thinking that the results are solely attributable to subgrid-scale aerosol-cloud interactions. I recommend that the authors address the following major issues before the manuscript is considered for acceptance.

Major:

1). Model Configuration and Experimental Design: The authors claim that the difference between the NO-ACIsub and ACIsub runs represents the impact of subgrid-scale ACI in the first set of experiments. However, the control run (NO-ACIsub) uses the regular KF cumulus parameterization without microphysical processes in convection, while the ACIsub run incorporates both subgrid-scale microphysics and subgrid-scale aerosol-cloud interactions. In the title, abstract, results analysis, and conclusion, the authors attribute the differences between the ACIsub and NO-ACIsub runs solely to subgrid-scale ACI. However, some of the significant differences could be due to the subgrid-scale microphysics. This could mislead readers into attributing all the differences to subgrid-scale aerosol-cloud interactions. The authors should reconsider the experimental design to more clearly distinguish the effects of subgrid-scale ACI.

2). Contradictory Model Configuration: In line 120, the authors mention that sand/dust (SD) is available in the CUACE model. However, in lines 155-156, they state, “The current scheme does not include real-time ice nucleation because dust is not available in the CUACE model.” Following this, the authors apply an empirical formula for constant ice nucleation (IN). This contradictory description of the model configuration could confuse readers and should be clarified.

3). The ACIsub run results in more total precipitation and a higher CLWP. There should be some compensating reduction in vapor at certain levels of the troposphere in the ACIsub run. However, the ACIsub simulation also shows increased vapor moisture at each vertical layer of the troposphere (Fig. 9f). It is concerning that some hydrological processes might be double-counted, or there may also be significant differences in ice-phase hydrometeors. This should be discussed in the manuscript.

4). The authors attribute the increased total precipitation in the ACIsub run to “Subgrid-scale ACI further enhances precipitation, especially at grid-scale.” However, the results of large-scale and convective precipitation are not shown in section 5.2 of the manuscript. The partition of resolved and unresolved precipitation should be provided, at least in the supplementary material, to help readers better understand the precipitation physics in this study.

5). The ACIsub run shows stronger relative humidity throughout the vertical atmosphere compared to the NO-ACIsub run (Fig. 9). Additionally, there is a notable increase in CLWP in the ACIsub run (Fig. 5f) compared to the NO-ACIsub run (Fig. 5e). However, the difference in cloud fraction between the ACIsub and NO-ACIsub runs is not significant. The further analysis of high, middle, and low clouds might help.