This manuscript compares global-scale simulation results of different ozone stress parameterization schemes in CLM5. The research topic is of strong scientific significance and value for model development. Using consistent meteorological and ozone forcing data, the study conducts a mechanism-consistent evaluation of three representative schemes—Sitch, Lombardozzi, and Li—and performs multi-scale validation against GPP data from MODIS and FLUXNET. Overall, the research framework is relatively comprehensive, and the results are reasonably convincing. In particular, the analysis of how ozone flux thresholds, response function forms, and memory effects influence simulation outcomes has practical significance for improving ozone stress processes in land surface models. However, the manuscript still has room for improvement in terms of the following aspects. The authors are therefore encouraged to revise the manuscript further to enhance its completeness and persuasiveness.

Specific comments:

1. It is recommended to further clarify the significance and scientific basis of adopting ozone stress parameterization schemes.
2. The current analysis mainly focuses on differences among the various ozone stress parameterization schemes. However, the results suggest that simulations without ozone stress sometimes exhibit higher accuracy. Therefore, it is recommended to include a comparative analysis between results “with ozone stress” and “without ozone stress.”
3. The zonal mean results are currently presented mainly as line plots (Figure 3), lacking quantitative comparisons among the schemes. It is suggested to add bar charts with quantitative metrics to more clearly illustrate and quantify performance differences across schemes.
4. It is recommended to further incorporate differences among biomes and potential directions for improvement into the discussion to enhance its depth.
5. In addition, the Li scheme does not consistently perform best for crop types and under autumn conditions. This should also be reflected in the results and discussion to ensure that the conclusions are more comprehensive and objective.

I enjoyed reading this manuscript providing a deep dive into alternative ozone damage functions and their performance. The original difference between the Sitch and Lombardozzi approach principally related to whether there is a decoupling in the impact of ozone on photosynthesis and stomatal conductance, e.g. there was empirical evidence of a decoupling under water stress, i.e. sluggish stomata. In the Sitch model the global parameter “a” (eqn 4) was calibrated for that DGVM and monthly chemistry-modelled concentration fields to reproduce the empirical dose-response function for a high and low sensitivity species within a plant functional type grouping. I would anticipate implementation of Sitch et al. scheme into another DGVM would require a recalibration of this parameter (can you confirm whether or not this was done?). For example, what was really done in the runs: Li + Lombardozzi thresholds and functions and Li+Sitch thresholds and function? (any retuning?) I like the Li approach as it considers the effect of cumulative dose on phenology/leaf turnover. Given Li “adopts a data-driven optimal threshold Y” and best form of equation to fit the data, one would anticipate its best performance (especially as the combination approaches may not have been retuned)? Nevertheless, it is interesting to see the comparison of performance against multiple datasets (the comparison was thorough). Given the above I am particularly interested in the comparison of Li and Lombardozzi parameterization schemes. Overall there are some interesting elements to this study and warrants publication although the justification and explanation around inclusion of the mixed schemes should be elaborated.