Thank you for the opportunity to review this manuscript. The authors present a thorough comparison between ERatmos and ERforest methodologies in quantifying the exchange of O2 and CO2 above a forest canopy. They demonstrate that ERatmos could be significantly influenced by entrainment, which results in unrealistic values. Consequently, the authors recommend against using ERatmos for constraining O2 and CO2 exchanges at a local scale, advocating instead for measurements at multiple heights to more accurately derive ERforest.

Entrainment significantly influences atmospheric composition within the boundary layer and is a well-researched phenomenon. However, this study stands out as the first, to my knowledge, that specifically addresses the impact of entrainment on O2 and CO2 exchanges. This represents a notable contribution to the field. This study suggests careful selection of O2 and CO2 measurements at single heights is required to correctly represent the biological exchange between O2 and CO2 in forest setting. This consideration is equally important in urban and other backgrounds, particularly for studies focusing on exchange ratios over smaller spatio-temporal scales. Given its importance and novelty, I recommend the acceptance of this study after the following issues are addressed.

Major comments:

1.    Is it possible for the effects of advection to be counterbalanced by those of entrainment?  The observed discrepancies between ERatmos and ERforest might stem from both entrainment and advection processes (Equation 8). In Appendix A1, the authors analyze the influence of the entrainment coefficient (β) on ERatmos signals and discuss instances where ERatmos aligns with ERforest. However, the role of advection remains unclear. Can we rely on measurements taken at a single height when advection's impact is potentially neutralized by entrainment? This interaction might explain why ERatmos and ERforest yield similar results.

2.    Does entrainment exert a more pronounced impact during typical days? This modelling study is generally based on the mixed layer theory. In studies by Ishidoya et al. (2013, 2015), their analysis did not specifically distinguish between measurements on ‘typical days’ and ‘non-typical days’, and derive similar ERatmos and ERforest values. After reading this work, I am fully convinced the impact of entrainment should be considered on ‘typical day’. However, it remains uncertain how this applies to specific instances, such as heavily polluted urban days or during extraordinary events like COVID-19 lockdowns, where mixed layer theory may not always hold. It would be beneficial for readers to understand the frequency and significance of entrainment during these atypical periods. When assessing single-height measurements on non-typical days, can we still depend on ERatmos for accurate representation? While modeling these atypical days using the CLASS model might be challenging, I suggest the authors discuss these considerations, possibly in Section 5.2, to provide a more comprehensive perspective.

The O₂/CO₂ exchange ratio above a forest canopy is a valuable tracer for understanding carbon exchange at the air-sea and air-land interfaces. We conventionally used a constant for global application, but this ratio can change significantly on a regional scale, and the mechanisms are still unclear. This manuscript presents an insightful analysis of the dominant mechanisms that determine the O₂/CO₂ exchange ratio by either using observations from a single height or from multiple heights. This study highlights the complexity of only using CO₂ and O₂ measurements at a single height to quantify ecosystem carbon flux, pointing out the advantage of using measurements from multiple heights for a precise ratio estimate.

The manuscript provides a comprehensive study, and the line of thought is mostly clear. I believe this paper is of interest to the general audience of Biogeosciences. I only have several concerns regarding the model experiment designs and the readability of the paper. I recommend a minor revision before this paper can be considered for publication.

Main concerns:

1. It appears to me that the CLASS model was run for only one day with a prescribed initial condition. These runs clearly do not reach a steady state. In this case, the result strongly stands on the initial condition (e.g., the initial jump). I am curious about the rationale for the initial jumps used in this study. I am also curious if the authors have tried to run the model for multiple days to reach a semi-steady-state and check if there are different results.
2. I appreciate the detailed study on factors modulating the ER_atmos and ER_forest. The result part, however, contains too many details and reads more like a technical report. I suggest improving the readability, by either revising the leading sentence of each section to sharply focus on the main result or adding a leading paragraph summarizing the main findings. The abstract section also contains too many technical details. I suggest shortening it significantly.

Minor comments:

1. Figure 1 is a little bit unclear. Could you label thick arrows with F(O₂)s and F(CO₂)s. It is also not clear in the figure that ER_forest is actually calculated from the gradient. The two-sided arrow across BL seems to suggest that O2 and CO2 entrainment are of similar magnitude.
2. L61: Expand on 'small scale process' upon its first mention for clarity.
3. L74: Please elaborate on extreme conditions (i.e., low SMI, etc.)
4. L132: Modify Eq. 5 to reflect that ER_forest is derived from a gradient.
5. L139: It is not clear how you calculate DtO2 and DtCO2. Based on fit to high-resolution data? What’s the time window?
6. L336- 339: According to Fig. 4b, 2018 features a very low ER_forest during the night. Could you comment on whether it is related to elevated soil temperatures that only matter at night?
7. Figure 8: This figure needs extra details. Arrows indicate sunrise to sunset. Better labeled on the figure to make this point clear.
8. L527-528: The finding that ER_atmos can be so large compared to ER_forest stands on the assumption that there is no vertical gradient in CO₂ and O₂ within the BL. If the resolved ER_atmos is based on using data that is very close to the top of the BL, this ER_atmos can be more sensitive to entrainment, compared to other studies.
9. Figure A2: Why the model results are not extended toward the beginning and the end of the day
10. Figure A3: It seems like the model overestimates the BL height in the afternoon and around the sunset. The simulated O₂ concentration also seems to have a clear phase lag compared to observation. How would these affect the simulated ER?
11. The title reads like the paper is trying to falsify the idea of using single-height CO₂ and O₂ observations to constrain surface carbon exchange. However, this approach is still valid if it only uses nighttime data. Given the detailed model experiments conducted in this study, I suggest modifying the title to better represent the comprehensive details of this work.