This study uses environmental variables at different temporal resolutions to estimate data product–based air-sea CO₂ fluxes and demonstrates that conventional flux calculations based on monthly mean variables overestimate ocean CO₂ uptake by 0.12 Pg C yr^-1. This bias is substantial, representing approximately 5% of the total ocean CO₂ uptake and accounting for about 25% of the discrepancy between model-based and data-based flux estimates. The experiments are well designed, the results are clearly presented, and the manuscript is well written. I therefore recommend publication after the authors address the following minor comments.

I have no major concerns, but some minor comments for the authors to consider.

Minor comments:

Line 34: For a non-expert of models, the reader might don’t know that the model has a very high resolution for all the variables.  Probably first state that this resolution issue does not affect the model-based flux estimates.

Line 60: check if this is a typo, “fco2 atm” I think.

Line 61/141: CO₂

Line 62: maybe add “often”, …coefficient is often parameterized…

Line 64: “Thus” should be “The”?

Line 70-71, 80-82: I can understand what the authors want to argue here, but I am not sure if I agree on the statements themselves. I agree that the model can simulate the feedback process, but it does not mean the data-based flux estimate does not account for this process. Any feedback occurs in the ocean side should have already been reflected in the observed fCO2w values. I agree that the model is less sensitive to the K, but the model is sensitive to the vertical transport, which is uncertain in the model if I am right. If you want to stress the uncertainties associated the data-based and the model-based flux estimates, I guess a fairer way is to be honest to both methods. Similar comment for the next paragraph. I would suggest including some uncertainty analysis of the models, which does not harmful to any of the results presented in this study.

Line 112: missed the right part of the bracket.

Line 173: just a question, looks like the monthly Xco2 is used. If a higher resolution Xco2 is available? Do you think the Xco2 also response to the cyclone? I guess there will be a vertical profile of Xco2, do you think the cyclone will increase the vertical mixing of the atmosphere and results in change of Xco2?

Line 171: Since indicated in line 187 that the ERA5 data is used, the highest resolution of the SST product available (in ERA5) is hourly I believe.

Line 194: ok, this partially answers my question for line 173, but I still curious about your opinions on my question.

Line 224: can you add a sentence why they are different in the end years?

Fig. 1a: the GCB results represent the average of latitude from -60 to 60, right? Probably indicate this information in the caption. Moreover, in the Key points and abstract, look like the conclusion are relevant for the global ocean, but it is in fact for 60oS-60oN, right? I know the averaged flux beyond this range may not significant relative to the global average, but in somewhere of the manuscript, this should be discussed, I think. Fig 14 from Takahashi et al. (2009) could be an example.

Line 268: Two suggestions the authors can consider if they want to include in the discussion: 1) do you think the 14C inventory-constrained K has already include this pressure effect? 2) What do you suggest for the future GCB? Looks like this pressure effect is relatively constant in different years. Do you think we can still use the monthly pressure for the data product (to reduce the calculation), but add 0.12 Pg C yr-1 to the final estimated flux? Or for GCB, they can directly subtract 0.12 Pg c yr-1 from their data-based CO2 sink estimates?

Line 296: Please check if Rustogi et al. (2025) is a model-based study or a data product-based estimate. The recent study by Dong et al. (2025, Nature Communications) highlighted that the bubble-induced asymmetric transfer increases data-based CO2 uptake by 0.3-0.4 Pg C yr-1.

I have included my comments as a supplement PDF.