the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2026
Nocturnal tropospheric gravity waves in a double-jet duct over the southern Taklimakan Desert: lidar observations and WRF simulations
Abstract. Atmospheric gravity waves can transport momentum and energy and couple different atmospheric layers. Their propagation and dissipation have been widely studied in the stratosphere and mesosphere, whereas their excitation and trapping within tropospheric double-jet ducts remain less well constrained. Here we combine coherent Doppler wind lidar (CDWL) observations at Minfeng, ERA5 reanalysis, and WRF simulations to investigate how a nocturnal double-jet structure over the southern Taklimakan Desert excites and modulates tropospheric gravity waves. Three events in September 2021 occurred under a similar configuration, with a lower-level easterly jet below 4 km, an upper-level westerly jet at 6–7 km, and a weak-wind layer between them. Oscillations of 10–20 min occurred within the weak-wind layer, with maximum amplitudes near the strong shear zone at the base of the westerly jet, suggesting a role for shear-related dynamical instability in wave generation. ERA5 buoyancy-frequency profiles show a stable layer near 4–6 km that favored thermal ducting, and Scorer-parameter diagnostics further indicate concurrent Doppler ducting. As the easterly jet intensified, the waveguide compressed and wave frequencies increased, consistent with a shift of trapped modes toward higher frequencies. WRF reproduced the first two cases reasonably well, yielding dominant horizontal wavelengths of 7–14 km and eastward phase speeds near 6 m s⁻¹, but showed substantial discrepancies in the third. These results indicate that nocturnal double-jet structures can act as an effective tropospheric waveguide by trapping gravity waves through thermal and Doppler ducting and shifting trapped modes toward higher frequencies as the duct evolves.
- Preprint
(6602 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-2865', Anonymous Referee #1, 18 Aug 2026
-
RC2: 'Comment on egusphere-2026-2865', Anonymous Referee #2, 24 Aug 2026
This paper uses lidar observations, ERA5 reanalysis data and WRF model simulations to investigate the ducting of trapped waves over the Taklimakan desert by wind shear associated with low-level jets and a high-stratification layer in the troposphere. The subject matter is interesting, of scientific relevance to the meteorology community, seems free of serious errors, and is well written. In my opinion, it should be suitable for publication in ACP after minor revisions.
Specific points
Lines 39-47: It would be a good idea to include in this passage an allusion to the importance of gravity waves as a phenomenon that must be parametrized in numerical weather prediction and climate models, due to their small (often sub-grid) scale. This is done in the Conclusions section, but it should also be included in the Introduction.
Lines 64-65: "drag effects on the large-scale circulation". This goes some way towards fulfilling the request in my previous comment. However, it should not only mention the importance of this drag on the large-scale circulation, but also the need to parameterize it as a subgrid-scale process in numerical models.
Line 71: "CDWL". This abbreviation is only defined in the abstract. Probably it should be defined again in the main text (i.e., here).
Lines 101-102: Incorrectly phrased sentence. One way to correct it would be, for example, by replacing "ranking" with "ranks".
Lines 110-111: The text in the caption of Figure 1 should be on the same page as the figure itself.
Lines 148-149: Incorrectly phrased sentence. I suggest correcting it by replacing "reconstruct" with "was used to reconstruct".
Line 157: "convection-permitting inner domains". Is a 1.5 km grid spacing convection-permitting? Not for shallow convection, I think.
Line 175: "closely collocated". In what sense? Please explain in more detail.
Lines 183-184: "suggesting that the observed gravity waves were likely generated by dynamical instability induced by the strong vertical wind shear". To test this hypothesis, it would be useful to calculate the Richardson number. While the Ri calculated from WRF data should have a threshold for shear instability at values substantially higher than 0.25, it would still be instructive to see how low it gets. Alternatively, radiosonde measurements could be used to achieve a higher resolution.
Lines 190-191: "All altitudes are given as above sea level". Is this really the most useful choice? Would it not be preferable to express the altitude above ground level instead?
Lines 193-196: This caption should be located below Figure 3, in the previous page.
Line 203: "highlighting the dominant oscillation periods". This temporal analysis by design excludes stationary gravity waves. Somewhere in the text, at the beginning of the discussion, this should be pointed out and justified.
Lines 214-215: "consistent with the polarization relations described by linear gravity wave theory". Perhaps add ", as will be shown next", since this has not been shown yet.
Line 229: "Based on the linearized equations". It should be added that these equations assume two-dimensional symmetry, otherwise Eq. (2) would not be so simple. Additionally, this assumption of 2D symmetry should be justified.
Line 237: "A Hilbert transform". "A" should be in lower case, as it continues the sentence after a comma.
Line 256-257: "the wave behaviour has evolved away from the initial evanescent state and toward a more vertically propagating mode in this altitude range". Note that trapped waves are by definition characterized by vertical propagation in the trapping layer and evanescence outside it.
Figure 6: The colours of the text "u lags w" and "u leads w" could be misleading, as they may be confused with the colours denoting the different time periods. I suggest using different more neutral colours.
Lines 275-277: "The elevated N^2 within this layer implies a strong restoring force on vertically displaced air parcels, thereby inhibiting the vertical propagation of gravity waves". Although the rest of the interpretation in the paper is correct, this passage seems incorrect. The Taylor-Goldstein equation tells us the exact opposite of what is stated, namely that high N layers are layers where (vertical) wave propagation is possible, whereas layers with low N tend to make the waves evanescent. The confusion may arise because, although vertical propagation is promoted by high N, if the layer of high N is narrow, then there will be upward and downward propagation within that layer (and reflection at its edges) over a short vertical distance, which produces net downwind energy propagation.
Line 290-291: "defining a strongly stable layer that acts as a thermodynamic waveguide, inhibiting the vertical propagation of wave energy". As noted in the preceding comment, actually high N promotes vertical propagation, it is the low-N layers bounding the high-N layer above and below that prevent vertical propagation, thereby keeping the waves confined to the high-N layer.
Lines 303-304: The last 2 lines of this caption should be below Figure 7, in the previous page.
Lines 319-320: "near-surface fields show relatively higher pressure over the basin floor and lower pressure along the surrounding terrain". Does this pressure include the hydrostatic component? If so, its physical meaning is limited and even misleading, since it is dominated by terrain elevation. To have a dynamically significant pressure that can be interpreted in terms of wind patterns, this should be the pressure reduced to mean sea level (as is standard in weather charts). Please check this, and if appropriate correct it.
Lines 335-336: "with a plausible wave source being the intense shear zone near the interface between the two jets". As mentioned above, calculating the Richardson number, either from radiosonde observations or from WRF output, would be a good idea, to check how low it gets and test this hypothesis.
Figure 9: The arrows and text in the legends of this graph are probably too small for comfortable reading. Please consider magnifying them.
Figure 10: The arrows are a bit too small to be readable. Please magnify the figure panels.
Line 345: "surface pressure". As mentioned 4 comments above, this might not be the best variable to present. Mean-sea-level pressure is probably more useful.
Line 356: "Wave propagation" should probably be replaced by "Vertical wave propagation", as it is this aspect that Q^2 quantifies.
Figure 11: The wave signal in w seems considerably weaker than in Figure 2. It would be appropriate to explain why when describing this figure.
Lines 410-413: It would be useful to point out in the caption of Figure 13 that the horizontal axes are inverted (with larger values to the left and lower values to the right).
Line 432: "the" at the end of this line should be replaced by "to".
Lines 437-438: "a distinct sign-reversal from negative values below to positive values is evident near 5 km altitude". Readers might wonder what happens above? Q^2 should also become negative there, correct? Why is this not shown and discussed?
Figure 14(c): The slope assigned to the phase speed in the Hovmoller diagram seems somewhat subjective and not very well constrained (unlike in cases (a) and (b)). Was this done merely by eye, or some more objective method was used? This should be briefly clarified and discussed.
Figure 14(d)-(f): Why are these graphs truncated above 6-7 km, preventing us to see the region of the atmosphere above the wave duct? Please extend the graphs upward to include this region or, if this is not possible, explain the reason.
Line 462: "critical levels within the WJ". It would be useful to mark the critical levels in the graphs where they can be represented, to have a more intuitive view of the flow.
Lines 478-479: "the observed shortening of the dominant period is consistent with an evolving mode-selection process as the duct was compressed". Is there anything more quantitative (even if only in terms of increase/decrease) that may be said about this? For example, is the variation in the expected direction, and why?
Line 494: "terminating". "suppressing" might be more appropriate here.
Lines 506-507: "this shear zone provided favourable conditions for wave generation". As mentioned before, could this be shown using the Richardson number?
Lines 517-518: "the ducted layer became vertically compressed and the dominant wave period shifted from longer to shorter values". As mentioned 3 comments above, is this consistent with the physics of the flow?
Lines 551-552: "lower-tropospheric ducting in double-jet environments may represent an underrepresented process in gravity-wave parametrization". As mentioned before, this is a relevant aspect, but it should also be mentioned earlier in the Introduction.
In the reference list, there are a few entries where the volume number and/or page range/article number of journal articles are missing. Please correct this.
Citation: https://doi.org/10.5194/egusphere-2026-2865-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 154 | 46 | 20 | 220 | 18 | 22 |
- HTML: 154
- PDF: 46
- XML: 20
- Total: 220
- BibTeX: 18
- EndNote: 22
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
（1）The wave source is asserted, not demonstrated. Shear instability at the base of the upper westerly jet is invoked as the generation mechanism (and is the most plausible candidate), but no Kelvin–Helmholtz instability (KHI) diagnostic, no Richardson-number analysis, and no spectral energy budget is provided.
（2）*The "mode-selection" / frequency-shift interpretation rests on a misapplied theory. Lindzen & Tung (1976) treated a duct bounded “below by the ground”; the duct in this study is bounded “above and below by evanescent layers in the free atmosphere”, which is a different (and richer) boundary-value problem. The simplified H/L scaling is presented as the operative physics, which is misleading.
（3）Q2 is plotted but no formal trapping condition (e.g., number of trapped modes, mode shapes) is evaluated, and the result that Q2 flips sign near 5 km is presented without sensitivity tests on the (uncertain) horizontal wavenumber k and phase speed c. The Doppler-duct / Scorer-parameter analysis is qualitative
（4） The model is acknowledged to be ~1 km too low in altitude and to fail on case 3, but no quantitative skill metrics (RMSE, bias, correlation) are given, and the consequences for the Q2 diagnosis are not honestly discussed. WRF validation is incomplete and partly negative.
（5）The phase-difference analysis (Figs 5–6) is particularly underspecified.
（6）A few substantial references are missing, e.g., on KHI in the lower troposphere, on tropospheric ducting by LLJ, on Taklamakan/nocturnal jet dynamics.
(7) The paper uses "duct", "trapped", "waveguide" interchangeably. Strictly, a trapped mode in a 2-layer duct requires Q2 > 0 in the interior and Q2 < 0 in both bounding layers, with a decay scale in each bounding layer that is short enough to prevent significant leakage. A leaky mode (sometimes called "partial duct" in the older literature) can propagate substantial energy out of the duct before being reflected. A useful diagnostic would be to compute, for each case, the vertical decay scale of the evanescent solutions in the two bounding layers, and to estimate the leakage timescale T_leak. If T_leak is much longer than the observed wave duration (~hours), the duct is effectively trapping. If it is comparable, the wave is leaky and the "trapped" terminology overstates the case.
(8) Three cases is fine for a case study paper, but the paper should be careful not to over-generalize. The phrasing in the abstract ("nocturnal double-jet structures can act as an effective tropospheric waveguide") is appropriately cautious; some of the Section 3 and 4 phrasing slips into "double-jet structures act as an effective waveguide" without qualification.
(9) Lidar data are "available upon reasonable request". Please double check the data request of ACP, it's better to deposit the processed CDWL data (1-min CNR, u, v, w, in a documented format) in a public repository (Zenodo, NOAA, or a project-specific archive) and provide the DOI in the revised manuscript. WRF output: same comment. The Hovmöller-diagram and FFT inputs would suffice; the full model output is not necessary. WRF namelist.input: please include this in the supplement.