Comment to “Role of aerosol–cloud–radiation interactions in modulating summertime quasi-biweekly rainfall intensity over South China”

This study investigates the role of aerosol–cloud–radiation interactions in modulating summertime quasi-biweekly rainfall intensity over South China based on both reanalysis data and model simulations, with interesting results provided. Personally, I would like to suggest its acceptance for publication with minor revisions.

Line 31-33, Recent review studies regarding the aerosol effect on clouds and precipitation could be referred and mentioned, Zhao et al. (2023, doi: 10.1016/j.atmosres.2023.106899) and Li et al. (2019, doi: 10.1029/2019JD030758).

Line 35-36, Not always suppressing precipitation, it sometimes enhances precipitation, as indicated by recent studies.

Line 37-39, The semi-direct effect often refers the case absorbing aerosols within clouds.

Line 42-43, Actually, there are proposed mechanisms for this invigoration phenomenon, while debates exist.

Line 66-68, If possible, a short review about the existing studies over South China is appreciated.

Line 87, Why do not use the radiation from CERES?

Line 93-96, Similarly, why do not use CloudSat/Calipso observations?

Line 134-136, To be fair, limitations for model studies should also be acknowledged.

Line 144-145, Could this nudging reduce/remote some effects from aerosol-meteorology interactions? And what will this affect the analysis results?

Line 168-170, Why do the authors use so long time as spin-up, instead of 12 or 24 hours as used by many studies?

Line 235-236, One more 50-year observation based climatological study by Su et al. (2020, doi: 10.3390/atmos11030303) is worthy to refer here.

Line 316, cloud ice particles.

Line 380, I am not sure if we can use “verification” or not since these are not observations, but model simulations, while we could say “support”.

General Comments:

This manuscript presents a thorough investigation of aerosol effects on 8–30-day rainfall anomalies, integrating observational datasets and WRF-Chem simulations. The study addresses a timely and relevant topic with important implications for subseasonal precipitation prediction.

While the manuscript is generally well-structured, several sections would benefit from clearer explanations of methodological assumptions, diagnostic frameworks, and limitations.

Many analyses remain qualitative. For improved scientific rigour and clarity, the authors are encouraged to report more quantitative metrics throughout (e.g., % changes, W m⁻², correlation coefficients, σ-standardised anomalies). Figures should be enlarged where necessary and include confidence intervals or error bars where applicable to support statistical interpretation.

Lines 26–27: Could the authors clarify how dominant the microphysical effect is compared to radiative forcing in quantifiable terms (e.g., W/m², % contribution)?

Lines 46–48: Could the authors elaborate on what these uncertainties are? How do you isolate aerosol-induced variability (especially AOD) from synoptic-scale meteorological variability that might independently influence precipitation?

Lines 82–86: The study uses MERRA-2 for aerosol variables. Considering the availability of higher-resolution datasets such as ERA5 (0.25°) and CAMS, could the authors justify this choice? Would ERA5+CAMS provide better representation of subseasonal aerosol–meteorological interactions over South China?

Lines 94–99: ERA5 is used for cloud and circulation diagnostics. Could the authors explain why a single data source (e.g., ERA5+CAMS) was not adopted consistently throughout the study to avoid discrepancies in spatial/temporal resolution or reanalysis biases? Could the authors clarify why ERA-Interim was chosen as input for WRF-Chem, rather than ERA5, which offers significantly improved spatial, temporal, and vertical resolution?

Line 139: Is the 20 km horizontal resolution sufficient to resolve mesoscale convection associated with aerosol–precipitation feedbacks? Were nested domains with higher resolution tested?

Lines 159–160: WRF-Chem is known to underestimate dust emissions, particularly in East and Southeast Asia. Were adjustments made (e.g., emission scaling or tuning) to account for this bias?

Lines 203–205: The text implies that precipitation anomalies precede the composite reference point, but Fig. 2a suggests that AOD anomalies occur ~6 days prior to rainfall anomalies. Could the timing be clarified?

Line 223: The correlation between AOD and rainfall is relatively weak (r = 0.25). Could the observed relationship be influenced by shared drivers such as regional circulation patterns?

Lines 256–258: The term “stronger antecedent AOD anomalies” would benefit from quantification. What time window defines “antecedent”? Could the authors specify how much precipitation amplification is observed (e.g., % increase, mm/day)?

Line 264: Why were LA–SP cases with −0.4σ < AOD < 0 excluded? Could this exclusion skew the composite comparison?

Lines 289–298: Please provide numerical values (e.g., means, ranges) to support the interpretation of rainfall or cloud anomalies in this section.

Lines 298–300: Could the authors provide an example or reference to clarify what is meant by “other processes”?

Lines 308–311: Please quantify the increase in cloud water content discussed.

Line 314: The term “modest and insignificant” change in ice cloud fraction should be qualified—how small is the change?

Line 317: Are uncertainties or confidence intervals available for the vertical profiles shown in Fig. 4?

Line 357: The latent heating of >30 W m⁻² is substantial. How does this compare with known values for tropical or monsoon convection in prior studies?

Lines 356–360: The conclusion that microphysical effects dominate is based on indirect evidence. Is latent heating (green curve) being used as a proxy for microphysical contributions? If so, this should be explicitly stated. How much greater is latent heating compared to radiative components in numerical terms?

Lines 393–404 (Fig. 6): The model appears to overestimate precipitation and underestimate AOD. Could the magnitude of these biases be quantified? Could model resolution explain the discrepancies, and how might this affect attribution conclusions?

Fig. 7: Despite broad agreement between model and observations, timing mismatches are evident. Could the authors discuss the sensitivity of their results to these temporal offsets?

Line 414: What is the magnitude and direction of the rainfall bias noted?

Lines 452–465: Please quantify the differences between high-AOD and low-AOD regimes to support the interpretation.

Line 465: Sulfate dominance is mentioned—was this confirmed through emission data or WRF-Chem’s chemical output?

Lines 499–509: This paragraph would benefit from numerical estimates to support its conclusions.