General comment

The paper presents a comprehensive array of feedbacks on climate variables resulting from O₃-vegetation interaction during boreal summertime of the period 2005-2014. Utilizing the climate-vegetation-chemistry coupled ModelE2-YIBs model, this study compares a simulation assuming no damage to vegetation, with another incorporating high O₃ sensitivity. While some parts of methods and results need clarifications, some corrections, and further editing, the topic of this study is suitable for ACP and enriches the already existing literature.

Abstract

I think the abstract is good, presenting the scope of the paper concisely. I realize it’s a matter of personal taste, but I’d only suggest a less “number-intensive” report of the results in this section. Furthermore, “2010s” appears for the first time here, indicating the “2005-2014” period, while, commonly, it indicates the 2010-2019 period. If possible, I would suggest finding a different label for this time window throughout the article.

Introduction

In my opinion, the introduction provides the essential information to contextualize the paper, listing an adequately ample array of examples regarding ozone effects on vegetation and the possible types of feedback. Anyway, I believe it would be helpful to provide more details for the readers that are less familiar with this specific subject, and to improve the style for enhanced coherence.

Line 45: I believe that flux-based measurements are also statistical, at least to some extent, so I would avoid qualifying exposure-based indexes this way – I suggest either substituting or removing the adjective.

Line 45-46: it might be good to specify what the “dynamic adjustment of vegetation physiological processes” depends on.

Line 51: I’m not sure “inconsistent” is the appropriate adjective in this case. I’d reframe this in a different way.

Line 63: “but” introduces a contrast, whereas “consistent but stronger” seems to convey an idea of enhancement. I’d suggest choosing a different wording. Throughout the paper, this adverb was used in the same way (i.e. introducing contrast without need) and I believe that substituting it with other words, or reframing these sentences could enhance clarity.

Line 65-67: The sentence “Inclusion of O₃-vegetation… surface O₃” feels disconnected from the rest on the text and, just by reading it, it is unclear if it is a consequence of what has been said earlier, if it’s a claim lacking a reference, or if it’s introducing the next sentence (in which case I’d suggest to improve the whole 65-72 lines).

Line 73-74: I believe “In addition to… between land and atmosphere” would be a direct consequence to ozone affecting stomatal conductance and I would mention this to improve logical coherence.

Line 92: “aggregated” instead of “aggregate”? Anyway, the 91-93 sentence “which calculated…  functional types” could be improved in clarity.

Method

This section highlights adequately the general structure and workflow of the study. While it outlines the essential methods, I believe it would benefit with more details. Please add unit of measure when introducing some variable.

To my understanding, surface O₃ within ESMs refers to the O₃ concentrations at the lowest model level, which might correspond to about 15-25 m above the ground, or even more depending on the vertical resolution. Do you think that this affects the O₃ evaluation part, since measuring stations are typically at a lower height (2-3 m above the ground)? In any case, I suggest providing more details about this matter, in the section 2.1 and/or in the section 2.4.

2.1 Model descriptions

This section should mention the temporal resolution for the model.

Line 115-116: It might be a good idea to offer more details about , , and

2.2 The O₃-Vegetation damage scheme

Line 134: Calculation for  should be detailed further.

Line 135:  would need a reference.

2.3 Experiments

Line 139: “two sets of simulations” are mentioned. Does it mean that each experiment contains more than one run, or that it’s just “two simulations” (one run for each of the experiments)?

Line 140-142: the simulation labels meaning could be made explicit.

Line 142: please be more explicit with “high O₃ sensitivity”.

2.4 Data for evaluations

I’d include in this section how each variable is evaluated specifically, at least when the evaluation is different than what expected – for instance, the evaluation carried out later refers to the surface daily maximum 8-hour ozone, whereas I expected ozone to be evaluated at the same time resolution of the model (or at least over a temporal scale that is more related to ozone vegetation fluxes). However, in general, specifying the temporal resolution of the dataset used for evaluation might help clarifying things.

Line 161: which simulation are you evaluating? It is not clear.

Line 165-167: Was there any reason for picking the years 2009-2011, as opposite to including more years? Since the simulations you’re using cover 10 years, it might be more adequate to compare them with more years.

Line 170-171: could you be more specific about this data product? What do you mean by upscaled?

Line 177-179: I believe NMB to be useful when comparing different quantities, but the non-normalized mean bias might convey a better quantification for some variables (such as ozone, or temperature), and I’d advocate for it either in place or siding the normalized version (for instance substituting the map of the observed data with the map of the non-normalized mean bias).

Results

3.1 Model evaluations

I believe that claiming a certain evaluation metric (such as correlation or NMB) to be high or low, or a certain variable to be adequately replicated, is a loose statement when not framed contextually. Any statement regarding the quality of an evaluation should reference a specific context. For instance, specifying if the evaluation is good compared to the literature, if the evaluated variable is the best among the ones being evaluated, or if the quality of the simulated variable meets the needs of the study. Furthermore, any proposition in this regard should be adequately motivated. If it is not possible to do so, I suggest leaving the numerical evaluation, without any qualitative remark. I signaled in the specific comments which are the lines that I found improper.

I think that this section does not cover strictly model evaluation, but also describes and contextualizes variables, so a different title might be more appropriate.

Please be more specific as to which simulation you are evaluating. As far as I understand, it is “10NO3”, but it should be made explicit throughout the text.

Line 186: “adequately” (see general comment)

Line 191-192: Is this referred to the MDA8 average? Or just concentration averages? It would be better to refer always to the same quantity.

Line 195: “high R=0.77 and low NMB of -6.27%” (see general comment)

Line 211: “low NMB of 8.49%” (see general comment)

Line 212-213: both simulations are referred to, but there’s only one NMB value, and one correlation value, and the figure refers only to one of the simulations (10NO3).

Line 213: could you provide any context to “high values in the tropical oceans”?                  

Line 214-217: How do you define good performances? (see general comment)

3.2 O₃ damage to terrestrial ecosystems

In this section, I would mention the fact that there are some areas of the globe that display higher GPP/stomatal conductance/LAI under “10HO3”.

3.3 Global climatic responses to O₃-vegetation interactions

Line 249: could you be more specific about surface temperature? It is not clear what you’re referring to, since it is then compared it with canopy temperature (line 251). I’d be more explicit.

3.4 Changes of air pollution by O₃ -vegetation interactions

Line 285-286: “The enhancement of O₃ concentrations in polluted regions may exacerbate the warming effect of O₃ and cause additional damages to vegetation.” Should be elaborated further.

Line 296-297: there’s a citation of the paper title instead of the authors.

Conclusions and discussion

Should it be “Discussion and conclusions?”

Line 310: Does “surface warming” refers to air surface temperature?

Line 312: At first glance, the word “further” seems contradictory with the previous statements. For instance, I would suggest replacing the sentence with “However, the enhancement of cloudiness decreased surface temperature…”. Ultimately, just from this sentence, it is unclear if the net effect is an increase or decrease in air temperature, or if these contrasting effects involve different regions.

Line 326: Does “surface warming” refers to air surface temperature?

Line 327-328: Would you be able to provide any explanation for this difference.

Line 352-354: I suggest being more specific.

Line 356-357: Most readers would recognize that a coarse resolution would be a limitation, but I would be more specific as to why with respect to O3-vegetation interaction. For instance, later (line 360), it is mentioned that high-resolution improves simulation for extreme events: are extreme events relevant for O3-vegetation interactions?

Line 357-359: this sentence is unclear – it seems like “Ito et al. (2020) shows… that the model results are involved in the CMIP6 Coupled Climate-Carbon Cycle MIP (C4MIP)”, as if being included in the CMIP6 guarantees by itself that carbon fluxes are well represented.

Figures

Figure 1: there’s a mistake in the panel letters.

Figure 1-2: it might be more useful to show the mean bias maps instead of the observed values, as it allows for easier comparisons.

Figure 3: have the p-values been corrected to account for the multiple repetitions in space in some ways? For instance, with Bonferroni, or false discovery rate.

Figure 4a: you refer to this as “air surface temperature” in the text (line 241), but here you call it “Tsurf”, which I think generates confusion with land surface temperature.

Figure 3, 4, 6: I believe that the sentence “Please notice the differences in the color scales” is redundant when comparing quantities with different unit of measurement.

The authors use the coupled ModelE2-YIBs model to estimate climate and air pollution responses to ozone-vegetation interactions globally during boreal summers. This is an important and interesting topic that has been studied before by multiple researchers including some of the authors themselves, and surely falls within the scope of ACP.

While a lot of model evaluation work has been presented, as the authors noted, their results contradict with what have been reported previously and may be highly uncertain. It is also disappointing that, with multi-year long simulations, only period-mean results are shown. It’d be nice to see discussions on the temporal variability in their results, the drivers of that, and the (new) implications from temporally-varying sensitivities for estimating future environmental changes.

Specific comments:

L15: define ModelE2-YIBs

L18: delete “the”

L23: specify surface O3 concentration metric used

L25-27: quantitatively state the impact on aerosols, which is claimed as a highlight of this study

L43: there are quite a few concentration-based metrics used to assess O3 impact, not just AOT40

L60: please specify the study period by Gong et al.

L63: please specify tool used by Sadiq et al.

L89: Why is YIBs spelled out here, after its first appearance at L60?

L100 and section 2.1: In general the approach applied in this work is old with no major updates from the authors’ previous works on similar topics. ModelE now has version 4 (https://data.giss.nasa.gov/modelE/) that should address some of the deficits in previous versions of the model. The model was run at 2x2.5 deg/40 layer resolution which is far from sufficient to resolve processes that could impact weather states, chemical environments and feedbacks that are studied here. The accuracy of parameters in Table S1, based on Sitch et al., should also be extensively discussed. For example, the sensitivity parameters in Sitch et al. seem to be sensitive to life stages of trees and climatic conditions which is not accounted for/discussed in this study.

L140/142: Can the authors please come up with new experiment names that are more self-explanatory?

L143-154: The description on model configuration is very confusing. Was the 2010s anthropogenic/biomass burning emissions applied for 30 year simulations (including spin-up)? Does the model simulations include other natural emissions such as soil, lightning, BVOCs, etc, and if so, in later sections, could their sensitivities be shown? Was the PFT type input (shown in Fig. S1) temporally fixed throughout the simulation period and if so why would land use/land cover change not represented in the system?

L153: Why only boreal summers are focused on for a global (including the southern hemisphere) assessment? Also note that high O3 days are not necessarily high O3 flux days.

L156: no need to include “special”

L159: add model after “for”

Section 2.3: More descriptions on the used datasets and their respective accuracies (particularly for remote sensing and derived data) are needed. The different temporal coverages of these datasets are very confusing and hard to be linked to the results presented later.

L162-165: There are clearly O3 observation data in Africa and South America in Fig 1b. What are their sources?

L168: Which version of MODIS data? Does LAI data also come from MODIS?

L187: MDA8 is not necessarily the best metric for evaluating ozone flux and vegetation impacts. It is worth noting that the poor coverage of O3 observations can affect the global model evaluation.

Section 3.1: Evaluation is done on a global scale - this should also be done and summarized by various regions of the world (particularly, but not limited to the two hotspot regions). It is unclear why Case “10NO3” was evaluated and what the reported performance for this case means.

L198: There is no illustration of spatiotemporal variability in emissions that can support to this statement.

L265-269: Note that many of these processes discussed here may be well represented in models running at coarse resolutions.

L287-302: What about aerosol climate impacts that feed back to ozone?

L337-371: This long list of limitations make the interpretation of the reported model results harder. The authors may want to articulate what useful information can still be gained from this sensitivity analysis in spite of all these sources of uncertainty.

Fig. 1 caption: replacing upper, left, bottom and middle with letter labels; add “the” before 2010s (and throughout the paper). Why did the model fail to capture high O3 in the western US and the Middle East?

Fig. S2 caption: replacing upper, left, bottom and middle with letter labels

Fig. S3-S4 colors are very hard to discern. Can the color schemes be adjusted?