Quasi-Biennial Oscillation Modulation of the Semi-Annual Oscillation in QBOi Models

M. Jaison, Aleena; Gray, Lesley J.; Osprey, Scott M.; Anstey, James A.; Andrews, Martin B.; Butchart, Neal; Chai, Zhaoyang; Hong, Dong-Chan; Huang, Kai; Kawatani, Yoshio; Knight, Jeff R.; Lin, Pu; Lott, Francois; Lu, Yixiong; Naoe, Hiroaki; Richter, Jadwiga H.; Rosenbloom, Nan; Serva, Federico; Smith, Anne K.; Son, Seok-Woo; Tang, Qi; Watanabe, Shingo; Xie, Jinbo; Yoshida, Kohei

doi:10.5194/egusphere-2026-4135

Preprints

https://doi.org/10.5194/egusphere-2026-4135

Preprints

22 Jul 2026

| 22 Jul 2026

Status: this preprint is open for discussion and under review for Weather and Climate Dynamics (WCD).

Quasi-Biennial Oscillation Modulation of the Semi-Annual Oscillation in QBOi Models

Aleena M. Jaison, Lesley J. Gray, Scott M. Osprey, James A. Anstey, Martin B. Andrews, Neal Butchart, Zhaoyang Chai, Dong-Chan Hong, Kai Huang, Yoshio Kawatani, Jeff R. Knight, Pu Lin, Francois Lott, Yixiong Lu, Hiroaki Naoe, Jadwiga H. Richter, Nan Rosenbloom, Federico Serva, Anne K. Smith, Seok-Woo Son, Qi Tang, Shingo Watanabe, Jinbo Xie, and Kohei Yoshida

Abstract. As part of the Quasi-Biennial Oscillation initiative (QBOi) phase 1, a model-intercomparison study explored the representation of the semi-annual oscillation (SAO) in participating models and found a common easterly bias of several tens of m/s compared to observations, mostly with weaker westerly phases and stronger easterly phases. QBOi phase 1 results also reported biases in model representations of the QBO, which dominates the region below the SAO, with most models displaying a westerly time mean wind bias, and generally weaker QBO easterly phases. Given that the SAO forcing terms can be influenced by the QBO, in this paper we explore the influence of these QBO biases on the representation of the SAO. A multi-model analysis is conducted here using QBOi phase 2 data from current state-of-the-art climate models to examine changes in representation of the SAO in response to corrections in zonal-mean zonal wind QBO biases. Most models show an improvement with a reduced easterly SAO bias in response to the corrected QBO bias. The extent of this response is found to vary significantly across models with SAO time-mean winds changing by 6 % to 403 % in simulations with a bias-corrected QBO relative to the control simulations. Changed gravity wave drag in response to the improved QBO is attributed as the forcing term having the largest overall impact.

How to cite. M. Jaison, A., Gray, L. J., Osprey, S. M., Anstey, J. A., Andrews, M. B., Butchart, N., Chai, Z., Hong, D.-C., Huang, K., Kawatani, Y., Knight, J. R., Lin, P., Lott, F., Lu, Y., Naoe, H., Richter, J. H., Rosenbloom, N., Serva, F., Smith, A. K., Son, S.-W., Tang, Q., Watanabe, S., Xie, J., and Yoshida, K.: Quasi-Biennial Oscillation Modulation of the Semi-Annual Oscillation in QBOi Models, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-4135, 2026.

Received: 10 Jul 2026 – Discussion started: 22 Jul 2026

Competing interests: JA serves as co-editor for the special issue to which this paper belongs. The remaining authors declare that they have no conflict of interest.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Preprint (PDF, 2409 KB)

Supplement (1154 KB)

Download & links

Status: open (until 23 Sep 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-4135', MATTHEW HITCHMAN, 19 Aug 2026 reply
This paper describes the effects of zonal wind nudging of the QBO on the amplitude and phase of the SAO (QBOi phase 2) in ~10 numerical models. It shows that models with better gravity wave drag parameterizations, higher model tops, and perhaps less diffusion near the model top can be more readily improved by QBO nudging, with regard to fidelity in representing the SAO. In the abstract it is stated that most models show an improvement with a reduced easterly SAO bias in response to the corrected QBO bias and that changed gravity wave drag in response to the improved QBO is attributed as the forcing term having the largest overall impact, which is parameterized unresolved gravity wave drag. This paper functions well as a medium of exchange of information for the community and it is useful and clear. It shows that gravity wave drag is of primary importance in explaining their results, but does not elaborate on the mechanism. I recommend that the authors include a description of what they think is happening in the models with regard to changes in gravity wave drag and recommend minor revision.

The primary conclusion in the abstract, restated on lines 109-110 and elsewhere, is robust, but the paper does not say much more about any physical process. I encourage the authors to provide a description of how the gravity wave filtering works, with regard to stating what happens to westward travelling waves and to eastward travelling waves for changes in a given wind layer imposed by the nudging. What are the gravity wave parameterizations used? Which models try to represent deep convective gravity waves from first principles? What are the ways that parameterized gravity waves can be absorbed in these models, going upward into the mesosphere, where both shear and reduced density can cause them to break?

l2-22: It would be helpful to briefly characterize vertical resolutions in the QBO and SAO layers. Does vertical resolution matter for parameterized gravity wave drag?

l148-149 and elsewhere: This is an important issue. I suggest that you describe a bit more about the upper layer in the models. Which models use Rayleigh friction? In your model standard, HadGEM3, what is the damping/numerical diffusion profile into the mesosphere?

l164 – I agree that nudging zonal wind is the right way to do it, in order to have the right relationship between temperature anomaly and vertical motion for a wave-driven MMC (e.g., downward / warm). If some models nudged temperature, then in the conclusions you could suggest more strongly that this might be a major reason why their results weren't as hoped. Vertical motion would be upward in an imposed warm anomaly in QBO W shear in a model with a radiative code, which is opposite to our understanding.

l180-190: You could introduce the idea here that another significant metric for evaluating model processes could be the difference in amplitude between the first and second phases of the SAO in each year. The reason that I mention this is that your definition of SAO blends both the first and second together. You could have a mean and a difference "PC" to intercompare models.

l215-220 - It may also be the case that vertical wind structure near 10 hPa is more important than near 50 hPa for filtering gravity waves, which might affect the more immediately-adjacent SAO layer more readily. But maybe that is not true. What do you think?

l245: You effectively treat HadGEM3 as the reference model in section 3.1, which seems fine. Do you think that the property of a higher model top is the primary reason? Suggest describe the damping properties of HadGEM3 in the upper layers a bit more.

l296-301 – In light of this discussion, what does this imply about the vertical profile of gravity wave drag? Perhaps add the thread of how it is that the models err here, which can be elaborated on in the conclusions?

Are there any models that represent the seasonal variation in deep convective gravity wave driving? If the greatest tropical deep convection occurs in DJFM, then a model which can represent that might be expected do better at the 1st v 2nd phase SAO amplitude asymmetry.

l357-358: The time mean vertical profile and how it is maintained is important. You could comment on quantification of GWD among models for the annual budget. It would give further useful information about GW behavior in the models.

l495-497: How are the gravity wave momentum fluxes for westerly and easterly gravity waves affected in each layer due to improving the QBO wind structure?

l505: Figure 8d shows the influence of the SAO on the QBO down to 50 hPa, due to the MMCs associated with the SAO. I think this is similar to what was shown by Coy (circa 1990).

l511-512: Please describe for the reader what this statement implies about what is missing in the models re GWD.

I notice that in this paper tropical GWD is deemed important for the westerly SAO phase, but I think it is also important for the easterly phase. The cross-equatorial flow is primarily a gravity-wave driven flow, a response to the large winter and summer gravity wave drag patterns. As air travels meridionally through the tropics it is exposed to a lot of deep convective gravity wave drag. In addition to GWD driving the pole to pole circulation during the SAO easterly phase, gravity wave drag integrated along the pathway from one hemisphere to the other also helps to explain why a maximum in easterly wind does not occur right at the equator, which would occur in inviscid flow starting with zero zonal velocity at a pole. Tropical westerly GWD, in addition to treatment of inertial instability parameterization in models on the winter subtropical side, can help to limit easterly advection.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-4135-RC1
RC2: 'Comment on egusphere-2026-4135', Anonymous Referee #2, 15 Sep 2026 reply

The paper "Quasi-Biennial Oscillation Modulation of the Semi-Annual Oscillation in QBOi Models" by Jaison et al. is an important contribution that compares the stratopause semiannual oscillation (SAO) simulated in a number of model simulations performed for the QBOi initiative. Model runs nudged to the QBO in the ERA5 reanalysis are compared with free model runs (Control) and improvements of the SAO are investigated. The SAO in the MERRA2 reanalysis is taken as a reference because MERRA2 assimilates MLS satellite observations and the SAO is therefore at least partly constrained by observations.

Two groups of QBOi models are identified. The SAO in the first group is improved by nudging to the ERA5 QBO. By removing a westerly bias of the QBO, the easterly bias of the SAO is ameliorated. The second group of models shows a mixed response to QBO nudging. The authors attribute the improvements of the SAO to a more realistic filtering of gravity waves by the QBO in the stratosphere below SAO altitudes.
Overall, this is a well-written and important paper that is recommended for publication in WCD after minor revisions.
A list of minor and technical comments is given below.

MINOR COMMENTS

(1) l.53: Please provide reference for these numbers, e.g. Garcia et al., JGR, 1997

(2) l.59: You should mention that from satellite observations it seems that significant driving of SAO easterlies by gravity waves seems to be linked to rare cases when the QBO wind filtering of easterly gravity waves by the QBO is less effective (Ern et al., Ann. Geophys., 2015).
Ern, M., Preusse, P., and Riese, M.: Driving of the SAO by gravity waves as observed from satellite, Ann. Geophys., 33, 483-504, https://doi.org/10.5194/angeo-33-483-2015, 2015.

(3) l.67: You should mention the reference Burrage et al., JGR, 1996 also here, as this paper introduces the original HRDI observations.

(4) l.84: you should mention that driving of the SAO westerly phase by gravity waves was also shown directly from satellite observation of gravity waves (Ern et al., ACP, 2021).

(5) l.89: You should include the reference Li et al., J. Clim., 2025 that investigates the performance of CMIP6 models for simulating the SAO. Further, you should explain what is common and what different between the QBOi Control run and the CMIP6 simulations described in Li et al., 2025.
Li, X., Han, Y., Xia, Y., Tan, X., Li, S., Wen, K., Wang, X., Li, Y., and Xie, F. (2025), Evaluation of CMIP6 Model Performance in Simulating the Stratospheric Semiannual Oscillation, J. Climate, 38, 4281-4297, doi:10.1175/JCLI-D-25-0045.1

(6) l.157: You should emphasize that nudging to ERA5 makes sense because on zonal average the ERA5 QBO should be relatively close to observations (e.g., Ern et al., ACP, 2023).
Ern, M., Diallo, M. A., Khordakova, D., Krisch, I., Preusse, P., Reitebuch, O., Ungermann, J., and Riese, M.: The quasi-biennial oscillation (QBO) and global-scale tropical waves in Aeolus wind observations, radiosonde data, and reanalyses, Atmos. Chem. Phys., 23, 9549-9583, https://doi.org/10.5194/acp-23-9549-2023, 2023.

(7) l.171: You should mention that there is a feedback between ozone and the QBO, and that interactive ozone will likely lead to changes in the simulated QBO (e.g., Ming et al., JGR, 2025).
Ming, A., Hitchcock, P., Orbe, C., and Dube, K. (2025). Phase and amplitude relationships between ozone, temperature, and circulation in the quasibiennial oscillation. Journal of Geophysical Research: Atmospheres, 130, e2024JD042469. https://doi.org/10.1029/2024JD042469.

(8) l.185-190: Did you check how sharp is the spectral peak of the SAO?

Is no significant spectral power lost by interannual variations of the SAO that would broaden the spectral peak?

(9) l.268/269, about Fig.2: My impression is that, except for HadGEM, at 1hPa near the zero wind line several models show downward propagation of the easterly SAO phases which seems to be less the case for MERRA2 and observations (e.g. Garcia et al, 1997). This is particularly seen for MIROC and LMDz and should be mentioned.

(10) Caption of Fig.2: mention that the figure refers to the "HadGEM3-similar group"

(11) Caption of Fig.3: mention that the figure refers to the "HadGEM3-distinct group"

(12) Fig.6: Optional comment: Would it make sense to include a distribution plot for MERRA2 (for a time period when MLS is assimilated!)?

(13) Fig.07: Optional comment: Would it make sense to include also MERRA2 TEM diagnostics in the figure?

(14) Fig.9: to better separate upper and lower row of panels, please introduce more white space

(15) l.479: You should mention that the finding of peak GWD a bit below the maximum change in the wind is as expected. It has been shown that eastward forcing of the stratopause SAO happens mainly in the zonal wind shear zones (e.g., Ern et al., 2021).

(16) l.619-621: You should include the information that already some evidence exists that high-frequency gravity waves are important for driving the SAO. As has been argued by Ern et al., Ann. Geophys., 2015 wave saturation that is not directly linked to critical levels seems to play also an important role

TECHNICAL COMMENTS
(1) General comment: please capitalize first letter of "figure", or "table" if followed by a figure, or table number. So far this is quite mixed throughout the paper.
(2) l.65/66: Barking Sands -> Barking Sands, Hawaii
(3) l.168: analyed -> analysed
(4) l.179:

have a parametrized non-orographic GW

->

use a non-orographic GW parameterization

(5) l.202: transition -> transitions
(6) l.249 is used -> are used
(7) l.322: remove blank after "models"
(8) Fig.4: Please check whether in the caption red and blue are interchanged for Fig.4d-f!
(9) In Table 1, please remove vertical bar in front of "-3.1" in the Control run wind column for MRI.
(10) l.561 pattern -> patterns
(11) l.575 they -> they are

Reply

Citation: https://doi.org/10.5194/egusphere-2026-4135-RC2

Supplement

https://doi.org/10.5194/egusphere-2026-4135-supplement

Viewed

Total article views: 262 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
187	58	17	262	43	21	19

HTML: 187
PDF: 58
XML: 17
Total: 262
Supplement: 43
BibTeX: 21
EndNote: 19

Views and downloads (calculated since 22 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	80	23	8	111
Aug 2026	41	15	3	59
Sep 2026	66	20	6	92

Cumulative views and downloads (calculated since 22 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	80	23	8	111
Aug 2026	41	15	3	59
Sep 2026	66	20	6	92

Viewed (geographical distribution)

Total article views: 240 (including HTML, PDF, and XML) Thereof 240 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 21 Sep 2026

Short summary

The tropical upper stratosphere has winds that switch between westerly and easterly every six months known as the semi-annual oscillation (SAO). Climate models commonly show an easterly bias in the SAO. We analysed data from many models to test if correcting the biases in winds lower in the stratosphere could improve SAO. We find correcting the lower-level wind biases reduces the SAO easterly bias, though further improvements, such as better representation of atmospheric waves, are still needed.


Total:	0
HTML:	0
PDF:	0
XML:	0