The manuscript presents a well-designed and timely study on the correlation between compound climate extremes and rice yields in southern China, with clear relevance to climate change adaptation. The authors leverage growth-stage-specific physiological thresholds, multi-source gridded data, and compound severity metrics to offer new insights into how concurrent heat-drought and chilling-rainy events affect rice production. This work makes an important contribution in the construction of metrics for compound stressors. However, several points require clarification, and improvements in structure and presentation would significantly improve the manuscript.

=== Major Comments

1. Ambiguity in Drought Stress Severity Definition. The calculation of drought severity appears to exclude events shorter than 10 days, regardless of intensity. Please clarify whether severity is accumulated continuously or only calculated if a 10-day event threshold is met. For rice, a very low soil moisture period, even for a week, can be fatal. Justification for this duration cutoff should be provided, ideally based on physiological or agronomic evidence.
2. Clarification of Kernel Density Estimate. Figures 2a and 2c are labeled as Kernel Density Estimates (KDEs), but the x-axis represents time (e.g., 1981–2018), which is not standard in KDE applications. It is confusing what variable is being smoothed, and how the density values should be interpreted. If these are smoothed frequencies or rolling densities over time, the figure should be relabeled or revised accordingly. I recommend providing a more detailed explanation of the construction, including the variable used, kernel type, bandwidth selection, and the interpretation of density on a time axis.
3. Interpretation and Modeling. he analysis relating yield anomalies to compound severity lacks clarity. Both axes in Figure 5 are restricted to negative values, with no explanation for this truncation. Are positive yield deviations and low-stress years excluded? If so, why?
4. Final Yield Model. Additionally, the use of simple scatterplots without formal statistical modeling is insufficient, given the complexity of the stress indices. I encourage the authors to fit and report a statistical model or clarify the final equation for this analysis to formally characterize the relationship between yield anomalies and compound stress severity. This would substantiate the visual patterns and improve analysis.

=== Comments on Manuscript Structure and Flow

Table 2 is referenced in the manuscript but not included.

The manuscript is generally well-organized, but there are several ways the narrative can be improved:

• Abstract: Consider simplifying and using more intuitive phrasing to improve accessibility to the general scientific audience.
• Introduction: The rationale is well-motivated, but some repetition of literature gaps can be consolidated. Move technical details to Methods.
• Methods: While comprehensive, this section is very dense. I suggest creating a labeled subsections on “Compound Severity Metrics” that put together equations and definitions. A flowchart or schematic of the data-processing pipeline would improve readability.
• Results: Avoid overuse of code-like labels (C2R2, H3D3) in narrative prose; use descriptive names. Ensure all figures are introduced with clear interpretive framing.
• Figures: Improve color bar labeling and add interpretive guidance in captions. Figures 3 and 5 in particular would benefit from better explanation of axis ranges and unit meanings.
• Discussion: While informative, the discussion can be tightened.

===Summary

This is an important and data-rich study that addresses key gaps in our understanding of compound climate extremes for analysis of crop yields. I hope the authors find these suggestions helpful in strengthening the manuscript for eventual publication.

The paper has significantly improved compared to the earlier version. I thank the authors for taking the revision process seriously and applying the requested modifications.

In my view, the paper still requires more clarifications, particularly in the methods section and in how the results are contextualized within the broader literature:

1. Copulas are introduced but never mentioned in the results. Is the KDE introduced in Fig. 2 equivalent to the copula CDF? If so, the terminology needs to be harmonized. If the KDE represents something else, this should be clearly introduced in the methods section.
   
   In the copula section, the purpose of Lines 204–207 and Equation 6 is unclear. Isn’t the joint probability (i.e., P(x > X, y > Y)) the main quantity of interest? If so, why not introduce Equation 7 directly? You may refer to this article for inspiration on copula methods and joint return periods: https://wires.onlinelibrary.wiley.com/doi/10.1002/wat2.1579.

2. Section 2.6 is rather generic. What are B1 and B2? Please introduce them properly. If B1 refers to climatic conditions and B2 to non-climatic factors, then from Line 416 onwards, a direct inference about the impact of infrastructure on yields cannot be made.

3. Discussion section: Please revise the text to reflect the broader implications of your findings and include only points that can be directly deduced from your analysis.

Specific Comments:

• L14: "Hamper" doesn’t sound right.
• L116: Briefly introduce the two datasets at the end of this sentence before discussing them individually.
• L121 (and repeated elsewhere, e.g., L163): What is "QX/T 468–2018"? This terminology is unclear. If it refers to internal coding, it may be unnecessary to mention.
• L248: Use "The impact of … on yield" instead of "yield impact."
• Figure 4: I am not sure I understand what DC refers to. If it represents correlation, shouldn’t the boxplot range be limited to 1? Why does it go up to 1.2 in panel d1 C2r2 for DCtot?
• L384–396: This section needs thorough revision. The reference to Zhang is problematic. Additionally, suggesting a dominant factor may not be valid, as these relationships are likely highly location- and case-specific.  “Large” is not the right word here. Please remind the reader what "#3" refers to.
• L417: Use "Different impacts of … on yields" instead of "yield impacts."
• L421: Were these losses shown in any figures or derived from your analysis? If not, consider removing this sentence. Also, since the study does not directly assess the impact of irrigation, that discussion may not be relevant.
• L437: Replace "rainy stress" with "rain stress."
• L456: On what plots are these spatial shifts in concurrent events shown? If you refer to shifts over time, clarify this. If not, the sentence is unclear.
• L457: "Spatial difference in phenology" is unclear, please rephrase.
• L463 onwards (Conclusion): The conclusion is not the right place to introduce new references or discuss limitations. Consider revising this section and relocating these points to more appropriate sections in the manuscript.