High above the tropics, winds in the stratosphere alternate between easterly and westerly roughly every 28 months, forming the Quasi-Biennial Oscillation (QBO), one of the dominant modes of variability in the tropical stratosphere. Although this oscillation occurs over the equator, its effects ripple beyond the tropics, influencing stratosphere–troposphere coupling, tropical convection, and global atmospheric circulation. Observations over recent decades show that the QBO in the lower stratosphere has gradually weakened, with two rare disruption events occurring in 2016 and 2020. Climate model projections suggest this weakening may continue under future global warming. These findings prompted us to ask: what is driving the QBO’s weakening?
Figure 1. Tropical SST evolution and QBO amplitude responses. (a) Multi-dataset mean tropical SST trends (30° S–30° N) for 1950–2020 derived from observational SST datasets (COBE-SST, HadISST and ERSST) together with the ERA5 reanalysis. (b–e) SST anomalies relative to the 1950–1980 climatology averaged over the tropical oceans (TO), tropical Indian Ocean (TIO), tropical Atlantic Ocean (TAO) and tropical Pacific Ocean (TPO), shown for observation-based datasets and CMIP6 multi-model means from historical simulations (1950–2014) concatenated with SSP2-4.5 and SSP5-8.5 projections (2015–2100). (f) Vertical profiles of linear trends in QBO amplitude (% decade⁻¹) using zonal winds averaged over 5° S–5° N from the ERA5 reanalysis and FUB observations for 1956–2020. (g) Vertical distribution of QBO amplitude using zonal winds averaged over 5° S–5° N, derived from 20–40-month band-pass-filtered zonal-mean zonal wind power spectra for ERA5 and the CMIP6 multi-model mean. Shading shows ±2 standard errors of the multi-model mean; error bars on black curves indicate ranges significant at the 95% level.
Tropical warming provides a clue
The sustained rise of tropical sea surface temperatures (SSTs) has provided an important clue. Warmer SSTs intensify deep convection, which alters the equatorial waves that drive the QBO. But under global warming, rising CO₂ and SST warming occur simultaneously, and observations alone cannot disentangle their individual effects. A deeper puzzle is that different tropical basins warm at different rates. The western Pacific and Indian Ocean have warmed more than the eastern Pacific and Atlantic, creating a La Niña-like zonal gradient. Could this spatial heterogeneity explain why the QBO is weakening? To find out, we used the CESM2–WACCM model to conduct targeted sensitivity experiments, warming the tropical Pacific Ocean (TPO), Atlantic Ocean (TAO), and Indian Ocean (TIO) individually and together, while holding all other conditions fixed.
Three oceans, three very different fingerprints
The results revealed a striking contrast. Warming the TPO produced the strongest QBO weakening: the mean amplitude fell from 19.9 m s⁻¹ to just 10.5 m s⁻¹, and the period stretched from 29.6 to 38.5 months. Easterly phases were particularly prolonged, and the downward progression of shear zones slowed markedly. This was the largest response we observed—but it was not the only one. In sharp contrast, warming the TAO actually strengthened the oscillation, pushing the amplitude up to 21.5 m s⁻¹ and shortening the period to 25.6 months. The TIO fell between these extremes, with a reduced amplitude of 16.6 m s⁻¹ and a faster 26.5 month cycle. The message was clear: the QBO does not respond to tropical warming as a uniform slab. The location of the SST anomaly matters profoundly.
When all oceans warm together, the signal gets blurred
The most instructive result came when we warmed all three basins simultaneously (TO). The QBO weakened to 17.8 m s⁻¹ and the period shortened to 26.3 months—consistent in sign with the individual responses, but notably muted compared with the TPO only experiment. Why? Because synchronized warming reduces zonal and inter-basin SST gradients. The sharp dynamical contrasts that make the TPO and TAO responses so different are smoothed out when the whole tropics warms at once. This non-additivity carries a warning: experiments that apply uniform warming may capture the right qualitative tendency, but they miss the basin-scale physics that governs the true magnitude of the response.
Tracing the dynamical pathway: waves and upwelling
To understand why each ocean leaves such a distinct mark, we diagnosed the zonal momentum budget using the transformed Eulerian-mean framework. In the control simulation, resolved wave forcing and parameterized gravity-wave drag alternate as the dominant momentum sources within the shear zones, balanced against vertical advection by the residual mean circulation. Under TPO warming, gravity-wave drag in the upper stratosphere collapsed: at 20 hPa, the easterly peak weakened by nearly three-quarters. Residual upwelling strengthened and persisted near 10–20 hPa, creating additional resistance to downward phase propagation. In contrast, TAO warming maintained stronger equatorial convection and a weaker Walker circulation, reducing wave filtering and allowing more effective momentum deposition. The TIO and TO experiments occupied intermediate states, where reduced upper-level wave forcing coincided with weaker low-level vertical advection, permitting faster descent despite smaller amplitudes. The QBO response, we found, is determined by the basin-specific competition between equatorial wave forcing and tropical upwelling.
From sea surface to stratosphere: two coupled pathways
The momentum budget tells us what is changing, but we wanted to know how the oceans communicate their signals upward. Two coupled pathways are at work. First, convective wave sources: in the TPO warming scenario, equatorial precipitation decreased while the intertropical convergence zone shifted northward, weakening the symmetric equatorial waves that project efficiently into the stratosphere. Second, the propagation environment: the intensified Walker circulation under TPO warming strengthened upper-tropospheric westerlies and modified critical layers, increasing background wave absorption before waves could reach the stratosphere. Conversely, TAO warming's weaker Walker circulation reduced such filtering, letting more waves through. These thermal and dynamical adjustments jointly determine where momentum is deposited and how fast QBO phases descend.
Why this matters for the future of prediction
This study provides a coherent framework linking spatially heterogeneous tropical SST warming to QBO structural change through adjustments of convective wave sources and wave propagation environments. Our results clarify a long-standing puzzle: why synchronized warming experiments consistently report robust QBO amplitude reductions, yet climate models disagree on period changes. Amplitude responds to the net weakening of wave momentum sources across most basins, whereas period depends on the delicate, basin-specific balance between wave-driven momentum deposition and residual upwelling. The broader implication is that differences in simulated SST warming patterns across models may be a major source of inter-model spread in QBO projections. Better representing how individual tropical basins force equatorial waves and modulate the Brewer–Dobson circulation could improve not only QBO simulations, but also the subseasonal-to-seasonal forecasts that rely on stratospheric predictability.
Figure 2. Schematic mechanisms of QBO responses to tropical ocean warming. (a–c) Responses to single-basin warmings over the TIO, TPO, and TAO; (d) pan-tropical synchronized warming (TO). All variations are shown relative to the CTRL run.
For details, refer to:
Wang, Y., Rao, J.*, Garfinkel, C. I., Ren, R., Osprey, S. M., and Lu, Y., 2026: Relative roles of different tropical oceans on the weakening of the stratospheric equatorial quasi-biennial oscillation. npj Clim. Atmos. Sci., 9, 83. https://doi.org/10.1038/s41612-026-01359-y.