Review of “QBOi El Niño Southern Oscillation experiments: Assessing relationships between ENSO, MJO, and QBO” by Elsbury et al., submitted to EGUsphere

This study investigates the extended winter MJO in an ensemble of QBOi climate models forced separately by perpetual El Niño and La Niña SST conditions. The results indicate that the simulated MJOs are largely insensitive to the stratospheric QBO, as the models fail to capture the observed lower stratospheric QBO-related changes that are critical for MJO modulation. However, the authors find that the simulated MJO and other convectively coupled equatorial wave activities differ markedly between the positive and negative phases of ENSO, with some deviations from previous studies.

Overall, the manuscript is well-written, and the methodologies are clearly presented. However, I have major concerns regarding the SST forcing used in the models. Given that the study’s conclusions heavily rely on these model experiments, I recommend that the authors reassess their experimental design and verify whether their results remain robust in light of previous literature before considering the manuscript for publication.

Major comment:

1. To enhance the atmospheric response to ENSO in their simulations, the authors scale the annual cycle of SST by two factors. Notably, the multiplicative factor of 1.8 for the positive ENSO phase is quite large, producing conditions similar to those observed during extreme El Niño events. This scaling also amplifies ENSO asymmetry by a factor of two, as illustrated in Fig. 1. Therefore, the conclusion that MJO activity is stronger during El Niño compared to La Niña is likely driven by these amplified SST perturbations in the models. Previous studies (e.g., Hendon et al. 1999) have reported that exceptionally warm Pacific SSTs can influence MJO activity by allowing the eastward propagation farther to the east. Similarly, the record-breaking strong MJO event in March 2015 (Marshall et al. 2016) was linked to warm equatorial central Pacific SSTs. Given the strong SST forcing in these experiments, enhanced MJO events are not unexpected but rather could be a reflection of selective amplification. Therefore, I recommend using composite SST anomalies without artificial amplification. This approach would provide a more representative assessment of ENSO’s overall impact on MJO activity rather than capturing only extreme conditions. Such a revision would ensure that the study’s conclusions are more robust and broadly applicable.

1. L167: Since the models are forced with artificially amplified SSTs, comparing their results directly with observed/reanalysis El Niño/La Niña composites is not entirely justified, as the simulations represent exaggerated ENSO conditions rather than typical observed events.
2. The objective of this study is not clearly discussed. While I recognize that ongoing research involves perpetual ENSO experiments in QBOi models, the specific aim of this work should be explicitly stated at the end of the introduction to provide clarity and context for the reader.

Minor comments:

1. Please include the horizontal resolution and the number of vertical levels for each model in Table 1 to provide a clearer representation of their configurations.
2. L146: The figure references for El Niño and La Niña SSTs appear to be reversed.
3. L240: How does ENSO influence the periodicity of the QBO? A brief explanation would be helpful.
4. Instead of showing wavenumber-frequency spectra for three different parameters in Figs.2-4, I suggest considering any usual variable like OLR/precipitation. However, a more detailed analysis of eastward and westward propagating modes could provide valuable insights. For example, comparing the ratio of eastward to westward modes in the low-frequency range between ERA5 and simulations might offer an interesting perspective.
5. 6: The ERA5 tropospheric temperature anomalies appear slightly stronger during El Niño compared to La Niña, potentially leading to stronger cold caps in the UTLS and modulating MJO convection. Are the differences shown in the right column statistically significant?
6. L467: How does the precipitation-based WF spectra capture the UTLS K-wave signal, which is most likely the dry K-wave?
7. Abhik et al. (2019) and Sakaeda et al. (2020) highlighted that many DJF EQBO events tend to coincide with La Niña conditions, whereas Niño 3.4 values are more uniformly distributed during WQBO. Did the authors find a similar nonlinear relationship between ENSO and QBO in their model simulations? A brief discussion on how well the models capture this observed relationship would be valuable.
8. Is the QBO Fourier amplitude calculation based on the zonal wind? Given that QBO-related temperature changes typically occur above and below zonal wind anomalies, cooler (warmer) temperature anomalies are expected to be located below the lowest level of QBO descent in EQBO (WQBO). Including the zonal-mean temperature profile from both the simulations and ERA5 would provide better clarity. The authors may consider following the approach of Son et al. (2017).
9. L562: In WACCM, the QBO descends to the lower stratosphere (~90 hPa), which could contribute to the weak sensitivity of MJO activity to the QBO. In MIROC, the weak positive MJO variability difference appears to be off-equatorial. Discuss these differences.

References:

Abhik, S., Hendon, H. H., & Wheeler, M. C. (2019). On the sensitivity of convectively coupled equatorial waves to the quasi-biennial oscillation. Journal of Climate, 32(18), 5833-5847.

Hendon, H. H., Zhang, C., & Glick, J. D. (1999). Interannual variation of the Madden–Julian oscillation during austral summer. Journal of Climate, 12(8), 2538-2550.

Marshall, A. G., Hendon, H. H., & Wang, G. (2016). On the role of anomalous ocean surface temperatures for promoting the record Madden‐Julian Oscillation in March 2015. Geophysical Research Letters, 43(1), 472-481.

Sakaeda, N., Dias, J., & Kiladis, G. N. (2020). The unique characteristics and potential mechanisms of the MJO‐QBO relationship. Journal of Geophysical Research: Atmospheres, 125(17), e2020JD033196.

Son, S. W., Lim, Y., Yoo, C., Hendon, H. H., & Kim, J. (2017). Stratospheric control of the Madden–Julian oscillation. Journal of Climate, 30(6), 1909-1922.

Review of QBOi El Niño Southern Oscillation experiments: assessing relationships between ENSO, MJO, and QBO

Overview: The study evaluates the simulated MJO in an ensemble of climate model experiments with perpetual El Niño or La Niña state. The study thoroughly examines the characteristics of the MJO, such as its lifetime, structure, and propagation. Given the lack of QBO-MJO coupling in the climate models, the authors can also attribute the difference in MJO activities between perpetual El Niño or La Niña simulations from the ENSO states. Overall, the manuscript contains some interesting information, but it can be improved by clarifying the goals and motivation for the study setup.

Major Comments:

1. I struggled to understand the main objectives of the study and how the presented analyses met the objectives. The authors need to clarify the goals of the study and present analyses that align with the goals. Lines 106-107 on page 3 state that “more process level understanding of how the ENSO and the QBO influence the MJO is needed, which we pursue here using unique coordinated model experiments”. Despite this statement, I found no “process-level” diagnoses of the models. The results mainly presented statistics on MJO properties such as lifetime, structure, and propagation. The lack of consistency in the stated goal of the work and presented results makes readers lost in what they should be getting out of this work.
2. Because I did not clearly understand the goals of the work, I also struggled to understand why a particular experimental setup was chosen. In particular, the authors should clarify the motivation for perpetual El Niño or La Niña simulations. To understand processes of MJO, QBO, and ENSO interactions, why were the perpetual El Niño or La Niña simulations needed? Why was it insufficient to instead examine how MJO varies with the simulated internal variability of ENSO? Why was it also necessary to amplify ENSO forcing in the simulations? These choices of experimental setup need to align with the clarified goals of the study.
3. Using ERA5 precipitation as an “observational” reference is inappropriate. ERA5 precipitation is not considered observations but rather considered a forecast since precipitation is not assimilated. The quality of reanalysis precipitation is particularly uncertain in the tropics (Gehne et al. 2016). The authors should use satellite-based precipitations to validate the simulations. The same applies to OLR in Fig. 3. The authors should use satellite OLR, not ERA5 OLR.
   • Gehne, M., T. M. Hamill, G. N. Kiladis, and K. E. Trenberth, 2016: Comparison of Global Precipitation Estimates across a Range of Temporal and Spatial Scales.  Climate, 29, 7773–7795, https://doi.org/10.1175/JCLI-D-15-0618.1.
4. In section 4 (discussion and conclusion), it would be helpful to expand the discussion on the causes of inter-model differences in the results and why the models still cannot simulate the MJO-QBO link. The authors seem to conclude that, on average, the simulations indicate that the MJO is sensitive to ENSO states. However, there were differences in how each model simulated such sensitivity. What causes the inter-model differences in MJO sensitivity to ENSO? While the models can simulate MJO sensitivity to ENSO, why can't they simulate MJO sensitivity to the QBO?

Minor Comments:

1. Throughout the manuscript, the authors refer to a few manuscripts in preparation (Kawatani et al. in preparation and Naoe et al. in preparation). I am unsure if it is a good practice to keep referring to manuscripts in preparation. Readers cannot access those and cannot obtain information that the authors refer to in those manuscripts. I suggest avoiding references to manuscripts in preparation.
2. In section 2.4 (page 7, lines 191-202), it is helpful first to clarify why two separate techniques are needed to obtain the power spectra. From my guess, the authors did so to separate the spectral signal in a particular season (Nov-April) vs the entire year. Please clarify.
3. Some tables were too big, and it was hard to find information referenced in the text. Instead, the authors can consider converting them into figures (e.g., Tables 2 and 3).
4. Page 17, Lines 411-413 (“Composites of the background SSTs associated with…). Which figure or table supports this statement? Please clarify.
5. Page 18, Lines 441-443: “the diversity analysis affirms that the fast and standing MJO archetypes are closely associated with El Niño and La Niña, respectively”. I struggled to find evidence for this statement. To support this statement, the authors should probably show how many standing, jumping, slow, and fast MJO events were found in each perpetual El Niño and La Niña simulations. Figure 5 does not support this statement. To my understanding, Fig. 5 only shows the pattern correlation of simulated and reanalysis OLR hovmollers, given that each propagation type occurs. So, Fig. 5 only shows the simulation skills of the MJO with a particular propagation pattern and how the skills differ between the El Niño and La Niña simulations.