See attached .pdf review file

Comments on “Experimental Protocol for Phase 1 of the APARC QUOCA QUasibiennial oscillation and Ozone Chemistry interactions in the Atmosphere) Working Group” by Orbe et al

Summary

This paper is a data description article which introduces the APARC QUOCA. Personally, I planned to anticipate the initiation ceremony of this project, but I was hampered by some temporary assignments. I am happy to review this paper. Generally, this paper is well written and constructed, and this project is clearly introduced. The relationship between ozone and QBO is worth exploring. However, I also find that this study emphasizes the possible relationship between QBO and ozone from a dynamical perspective. Actually, for interactive ozone runs, the chemistry processes might also be responsible for the interaction between ozone and QBO. That is, the relative contribution of the ozone transport and chemical reaction for ozone change and therefore the QBO changes is not well shown. I suggest a revision at the present time (mostly minor points).

Major comments

1. The introduction mentions that the SST changes can more directly influence the QBO amplitude via modifying the Brewer Dobson circulation strength and the non-orographic gravity wave drag forcing. However, it is not easy to quantify the relative contribution of SST and the direct impact of ozone to the QBO amplitude. In my understanding, the SST might be the leading driver for the QBO amplitude and cycle changes. The interactive ozone limitedly affects the QBO amplitude, although some individual models indeed simulate a change of 10-20% in QBO amplitude.

1. The largest obstacle for this project is to conquer the model bias in simulation of the QBO amplitude. First, models diverge in the simulate QBO amplitude, and the intermodel spread is much larger than the so-called improved change of the QBO amplitude with interactive ozone. Second, most CMIP5/6 models underestimate the QBO amplitude. Do you have any suggestion whereby the QBO amplitude bias and the so-called QBO amplitude change is well separated.

1. This paper also plans to store the data on the admin server, where most of the community has no access to this dataset. I suggest to provide the data freely to all users, especially for beginners of the QBO works. CMIP6 data are available for global users. Since this project provides similar variables as in CMIP project, why not follow the CMIP6 project to share the data?

1. To take an example of using the APARC QUOCA project experiments, only one model provides the related runs. The GISS model indeed simulate the QBO amplitude change, but the consistency with other models is unknows. If one or two more models also provide the experiments and consistent projection at least with the same sign in change, the significance of this project is robustly confirmed. Figure 1 is a study from Butchart et al more than 20 years ago, which is too weak to support this study, and the model the authors used is also not mentioned. If they use a different model from GISS, please also show the relative change of the QBO amplitude with interactive ozone in this different model to support the results from GISS.

1. The GISS model the authors used is different from the CMIP6 version. I checked the GISS models in CMIP6, but the QBO is not simulated. Please provide relevant information for the difference between this model version and the CMIP6 model version in CMIP6. A general assessment of the model skill should be shown before the project is initialized. Low-skill models might not well answer the questions Q1-5 that is based on the models with a reasonable QBO simulation.

1. The questions Q1-5 are repetitive in my understanding. Q1 is ozone => QBO and QBO teleconnections (including QBO downward impact to surface). Q2 is the identical to Q1 but for future scenarios. Ozone => QBO and QBO teleconnections (including large-scale circulation, BDC, and polar vortex). Q3 is repetitive for Q1 and Q2, but emphasizes the mechanism. What is the difference between mechanisms and dynamics? They are the same question. Further, the second question in Q3 is also biased from the main stream of the APARC QUOCA (comparison: CO2 => radiation, chemistry, and therefore circulation VS CO2 => SST changes => circulation). Q4 is 4xCO2 => BDC and polar vortices change => ozone feedback. Q5 is identical to Q4 but for the troposphere: 4xCO2 => troposphere => ozone feedback. Therefore, I suggest to focus on two or three key questions and do not list the question branches as the key questions.

Minor comments

1. L24: QBO can also impact the North Pacific pressure (geopotential height). Please refer to Rao et al. 2020a, 2020b (doi: 1175/JCLI-D-20-0024.1; doi: 10.1175/JCLI-D-19-0663.1).

1. L37: I am not sure if this feedback is seen in all CMIP6 models? Or this feedback is a result in very few models, and it is not extracted from observations.

1. L44: This QBO biases should be well reviewed before this sentence. The weak QBO and the weak QBO teleconnections of both hemispheres have been reported in Rao et al. 2023ab (doi: ). You must provide the background how the QBO is simulated. The ozone can be prescribed, but the QBO can also be prescribed. If the interaction is focused, the experiment from prescribed ozone to free QBO and from prescribed QBO to free ozone can be setup.

1. L48: Butchart et al. 2023 is frequently mentioned in this paper. The change in the QBO cycle is seen in this model, but inconsistent with other models (L53-54). So what can we learn from this inconsistency?

1. L79: Projected QBO amplitude has been explored in Rao et al. 2020c GRL (doi: 1029/2020GL089149). Please refer to this article for details. QBO is weakening, but the extratropical impact is strengthening.

1. L81-89: Some aspects of those three questions are repetitive.

1. L103: Can you show the results between 30-50 hPa where the QBO wind variability is largest?

1. L104-105: Here you mention INT and NINT, which mean interactive and non-interactive ozone chemistry. In Figure 2, LINOZ is also an interactive chemistry, although it is simple.

1. L110: All ozone variability except the annual cycle is removed. It is not only the ozone related variation that has been removed, the variation related to SST forcing such as ENSO is also removed. This sentence is not accurate enough.

1. L111-112: Here it is a problem that should be addressed first. If the model is not consistent with observations, the dynamics processes analyzed from models are also unreliable. If the modeled annual cycle is biased from the observed annual cycle, what is the significance of the modelling results?

1. L116: The ozone feedback is the difference between interactive and non-interactive chemistry models. I suggest not to include the role of CO2, which mainly determines the mean background circulation and climate.

1. Figure 2: The figure legend is misleading and hard to follow. In my understanding, OMA should be INT, to keep consistency with your description in the introduction section. LINOZ is also a type of INT, but you did not mention in the main part of the section 1.

1. Figure 2 caption: A typo here is picked out. FIXED is a type of non-interactive run, so it should be FT-NINT-1xCO2 (rather than FT-INT-1xCO2). Please clarify.

1. L124: I agree with this statement. The SST is the determinant factor for the BDC change. Other improvement such as interactive ozone contribute very little to the total change of BDC.

1. L128-129: I agree that the boundary conditions are of the first importance to induce the circulation change with global warming. The QBO changes might be very weak and is hard to extract.

1. L141: The dampened BDC response sounds like that the CO2 is more important than ozone for BDC response. Ozone only weakly dampens the total response. Ozone role in BDC change is indeed limited.

1. L157-158: Very repetitive when compared with Q1-3. I have suggested to reconsider the key questions of this project.

1. Figure 3c is repeating annual cycle of ozone without any QBO signals. What is the purpose for this experiment?

1. L173-174 “while …”: Here the CO2 response is mentioned. I disagree that the CO2 response exists. Since you have removed the role of global warming processed by using difference between two 4xCO2 runs. Here the 4xCO2 only creates a different mean state compared with the present-day climate. So it is none of the CO2 response business. That is, this sentence has entangled two questions. Please clarified.

1. Table 1: FT-NINT-4xCO2, O3 should be “Climatological PD-INT” or “Same as above”. I am not sure why the description is different from PD-NINT

1. L184: I did not see input4MIPs in Table 1.

1. L195: I wonder if the three climatologies are different from each other. If so, how do you explain this difference?

1. L196-197: This difference does not mean the improvement of those experiments. I still believe that the observational cycle shows some significance.

1. L214-215: SST changes the property of QBO, which has been reported in Zhu and Rao 2025 (doi: /10.1016/j.atmosres.2025.108241)

1. Figure 5: 30-yr and 50-yr PD NINT runs show nearly identical results. The cycle and amplitude of the QBO is nearly unchanged. What the mean QBO amplitude in the three runs if you use the method of Wang et al. 2025 (doi: 1016/B978-0-443-15638-0.00013-7)?

1. L228-229: The underestimation of the QBO amplitude should be included in the review of the existing literature in ample evidence (Anstey et al. 2019; Rao et al. 2020a, 2020b …)

1. L234-236: Both models and observations provide evidence. Add some observation evidence. Lu et al. 2024 CAWE (doi: 1016/j.wace.2023.100627), 2025 Comm. (doi: 10.1038/s43247-024-01812-x)

1. L238-239: The role of the chemistry should be mentioned in the introduction earlier. Please tell readers clearly where dynamics dominates and where chemistry dominates.

1. L245: As I point out in the major comments, why does the CMIP6 GISS model not simulate the QBO?

1. L266: Figure 2 has shown the linear ozone chemistry, but the introduction to the linear chemistry appear here, much later than expected.

1. Table 2, 3: I only have some knowledge that only four of those models can simulate the QBO (GEOSCCM, CESM2-WACCM, UKESM1, and MIROC6), if the model versions in CMIP5/6 is examined. Other models are unfamiliar to me and probably other readers.

1. L320: data is => data are

1. L325: JASMIN is only friendly to registered users. Do you have a plan B that share the datasets with more users?

1. L346: AMIP is good, but it lacks the air-sea interaction (or feedback from the sea). Can you have any better methods of assessing the role of the interactive ozone?

1. L358-369: Theme 1 + Theme 2 can be VS Theme 3.