the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 May 2026
Turbulence occurrence in the Tropical Tropopause Layer from superpressure balloon observations: distribution and sources
Abstract. Turbulence characteristics remain poorly documented in the Tropical Tropopause Layer (TTL), although turbulence may play a significant role in the vertical transport of tracers and momentum from the troposphere to the stratosphere. In the framework of the Strateole-2 project, we use pressure, temperature and GPS in situ measurements collected under stratospheric Super-Pressure Balloons (SPBs) during long-duration flights at quasi-constant altitude (~20 km) across the tropical belt. These quasi-Lagrangian observations are used to estimate high-resolution time series (~110 s) of the gradient Richardson number (Ri), allowing us to characterize the flow regime, turbulent or laminar, along the balloon trajectories.
Using a classical instability criterion (Ri<0.25), we estimate a mean turbulent fraction of about 0.18 in the tropical lower stratosphere and investigate its geographical variability. Turbulence occurrence is significantly enhanced in the vicinity of deep convection. Increased kinetic energy of short-period gravity waves is also observed close to convection, strongly suggesting that the breaking of gravity waves generated by convection constitutes an important source of turbulence in the TTL. We also identify turbulence events occurring far from convective regions, particularly over the Pacific Ocean, where they cluster in areas of enhanced vertical wind shear associated with the Quasi-Biennial Oscillation (QBO) amplified by the weakening with altitude of the upper branch of the Walker circulation.
These results highlight the important role of both convectively generated gravity waves and large-scale circulation patterns in controlling turbulence occurrence in the tropical lower stratosphere, with potential implications for tracer transport and mixing across the TTL.
- Preprint
(16249 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-2520', Anonymous Referee #1, 13 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2520/egusphere-2026-2520-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2026-2520-RC1 -
RC2: 'Reply on RC1', Anonymous Referee #1, 13 Jun 2026
I would like to make a correction to the Major Comment in my review, and offer a suggestion for another analysis that could address the issue it raises.
I meant to say that, if the trace of balloon zonal (or meridional) velocity versus altitude over one cycle forms a line, then the estimated shear should be correct. Again, if they trace an ellipse, the balloon and wind horizontal velocities would not be equal and the estimated shear would contain errors. My apologies for the incorrect wording in my original comment.
Another analysis that could address my comment would be to calculate the cross-spectrum of balloon altitude and zonal velocity, or of balloon altitude and meridional velocity. If the phase difference between the altitude and velocity is zero in the frequency range of the balloons’ oscillations about their neutral density surfaces, then it is possible that the balloon and wind velocities are equal.
Citation: https://doi.org/10.5194/egusphere-2026-2520-RC2
-
RC2: 'Reply on RC1', Anonymous Referee #1, 13 Jun 2026
-
RC3: 'Comment on egusphere-2026-2520', Anonymous Referee #2, 25 Jul 2026
Measurements from superpressure balloon (SPB) campaigns are used to deduce turbulence occurrence near the tropical tropopause. Specifically, the naturally occurring vertical oscillations of the balloon position are exploited to estimate vertical derivatives of temperature and horizontal wind components, which in turn can be used to estimate the gradient Richardson (Ri) number. The common threshold of Ri < 0.25 is used to designate flight segments to turbulence occurrence. The campaign data used are found to include an inferred mean turbulent fraction of 18%. Most turbulence is found in close proximity to deep convective activity and seems to be associated with enhanced high-frequency gravity wave activity. Enhanced turbulence rate is also found in large-scale shear zones associated with the QBO and upper branches of the Walker circulation.Estimates of turbulence occurrence near the tropical tropopause are important for improved understanding of tracer transport and its potential consequences for climate feedbacks. But unfortunately these estimates are rare, primarily because of the lack of sufficiently high-resolution observational data. So the present analysis based on unique SPB measurements is undoubtedly beneficial and helps to fill this gap. The paper is well written and mostly easy to follow.I have three general concerns that I think are important to be addressed in the paper and which may require extra work, which is why I selected "major revisions". I also have a few minor comments (see below), but otherwise support publication of this paper.General concerns:1The estimated vertical gradients are not instantaneous and therefore don't constitute true vertical gradients because they are affected by the horizontal drift of the balloon. This means that the inferred vertical gradient includes contributions from changes in time and horizontal distance of the field of interest (T, u, v). This potential issue needs to be discussed somewhere in the paper. Perhaps such a discussion is included in previous published work by the authors, in which case such a discussion could be brief. I did have a look at Wilson et al. (2023), where the effect of changes in time are discussed for the case of temperature gradients and the authors concluded such effects to be negligible. But I still wonder about the respective effect on the wind and more importantly the effect due to the horizontal drift of the balloon. For a 10 m/s horizontal velocity the balloon drifts 1 km in 100 s (roughly the measurement resolution used for vertical gradient estimates), so horizontal gradients at a km-scale would play a role. I wonder if situations with large-amplitude high-frequency gravity waves and/or turbulence could create strong enough horizontal gradients at this scale so that the inferred vertical gradients are "contaminated"? There could also be a corresponding sampling issue for turbulent patches if their horizontal extent is less equal 1 km.2The campaign data is certainly great, but sampling is still sparse over some regions. Fig. 8 indicates that over South America, the Atlantic and Africa only few flight tracks contribute to the total data. So the turbulent fractions necessarily suffer from sparse sampling there (perhaps most importantly in time). This sampling problem is a bit hidden by simply looking at the number of measurements (as opposed to contributing flight tracks) and should be discussed more.3The interpretation about the large-scale setup making turbulence more likely in regions associated with the QBO/Walker circulation shear zone (section 5) is quite interesting. Some quantitative evidence is shown in Fig. 14, but I wonder if it could be made more specific by showing a vertical cross section (longitude-height) of zonal wind and overturning streamfunction? There is quite a bit of discussion in terms of the schematic shown in Fig. 13, but such a discussion is necessarily somewhat speculative and should be made more quantitative where possible. The ERA5 data should include the required information about the large-scale circulation and allow such a quantification.Minor comments:line 5: I'd move the effective resolution "(~110 s)" to after "high-resolution" to make sure it's not misunderstood as referring to the length of the time seriesline 113: I suggest to provide your estimates for the distance from convection for the two example situations shown in Fig. 3line 132: please provide a comment as to why you didn't use the GPS altitude to compute the vertical displacements (I assume it's not sufficiently accurate?); also: is the hydrostatic assumption, which would not be needed for the GPS altitudes, sufficiently valid in turbulent regions?line 151: I suggest to move this comment about what is shown in Fig. 4 to its captionline 195: this is a common problem, also for radiosonde measurements, for which correction methods exist - please comment if any correction of this known bias was applied to the measurements hereline 252: please provide those different latitudes for referenceline 308: I think you need to provide justification for the choice of \Delta t = 3h (also relevant for section 4.3.1)line 337: this makes sense, but is this sufficient to claim that the obtained kinetic energy is really due to GWs? Could it include turbulent kinetic energy (which would be quite relevant for the results shown in Fig. 11)?line 344: what does the detrending refer to, i.e. what trend is being removed?line 399: do you have evidence for your hypothesis that the Walker circulation reaches above 100 hPa? Main convective outflow is between 200-150 hPa, so quite a bit lower (see also major comment 3)line 408/409: with typical ENSO time scale of 4-7 years there would be order of 10 La Nina's in 40 years, i.e., the "sixth strongest" would in fact not be particularly strong ... perhaps simply say "moderately strong" (which I think corresponds to the official NOAA distinction)Fig. 13: what are the color contours in the "Zoom over the Pacific"?Appendix A: why do you need to approximate the relations (e.g., Eq. A7)? Couldn't you simply compute the full relation based on the data?Appendix B: why only show T+, u+, v+ in Fig. B1 (i.e. why not also include their antisymmetric parts)?Citation: https://doi.org/
10.5194/egusphere-2026-2520-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 206 | 66 | 18 | 290 | 14 | 15 |
- HTML: 206
- PDF: 66
- XML: 18
- Total: 290
- BibTeX: 14
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1