This paper investigates the occurrence of large vertical wind shear in the ECMWF ERA 5 reanalysis, taking advantage of its increased vertical resolution compared to earlier reanalysis products. The authors find that wind shear is enhanced in a region just above the lapse rate tropopause, named here the tropopause shear layer. The climatology and structure of shear around the tropopause is analyzed and the physical processes responsible for its generation are discussed, as well as implications for turbulence and stratosphere-troposphere exchange.

The paper is within the scope of WCD. Overall, I found it interesting and well-written and the referencing appropriate. I recommend publication subject to a few minor revisions (additions) detailed below.

Main comments:

1) The role of model vertical resolution is acknowledged in several instances but could be discussed more thoroughly. The authors mention in the introduction (p 3 l 67) that the previous generation of ECMWF reanalysis (ERA interim) was unable to describe the shear layers. How much improvement has ERA 5 brought ? The statement that it has ‘a sufficient resolution to realistically resolve central features in the UTLS’ (p 6 l 167-168) could be justified, although a comparison to radiosonde or other observations might be beyond the scope of this paper. The authors could refer to Figure 1 of Hoffman et al. (2019) which shows the respective resolution of ERA interim and ERA 5.

Similarly, on p 6-7 l 197-199, the issue of vertical resolution variations with altitude is raised but its impact is not estimated quantitatively. The authors may want to mention that the change in vertical resolution is slight in the region of interest, a few percent (how many ?) compared to the changes in shear occurrence frequency which varies by 2 orders of magnitude in their Fig. 3.

Finally, the authors could briefly comment on the performance of ERA5 compared to the ECMWF operational analysis which they analyzed in Kaluza et al. (2019).

2) Reading the paper made me wonder how much of the shear structure might be diagnosed/explained using the thermal wind relation (for instance the pattern in Fig. 4 and 5 a, the relationship with ridges in Sect. 4.2 or the seasonal variation of the EAJS in Sect. 4.3.1). The authors emphasize (p 24 l 547) that part of the co-location with the tropopause is related to thermal wind balance, as suggested for example in the cited study by Endlich and McLean (1965). They could test this hypothesis quantitatively. Sure, it does not directly explain how the temperature gradient are generated but this is documented elsewhere and a simple diagnosis could here help disentangle “balanced dynamics” from gravity wave effects.

3) If the resolution of ERA5 is good enough to distinguish the LRT from the cold point tropopause (CPT), it would be interesting that the authors determine which of the LRT or the CPT is closest to the enhanced shear layer in the tropics. This would be particularly relevant to the question of the stratosphere-troposphere boundary in the tropics (e.g. Pan et al., 2018) . I note that, in Fig. 5, the shear layer in the tropics is shifted upward by ~ 1 pixel (500 m) with respect to the LRT.

Other points:

p 2 l 30: ‘thermodynamic structure’ : do you mean because of mixing and heat exchange? If yes, this should be explained. Otherwise, ‘dynamic structure’ or just ‘structure’ would fit better (wind shear is strictly speaking not a thermodynamic feature).

P 3 l 61: Please convert feet to meters, following WCD guidelines

p 3 l 91: The authors may want to cite the recent paper by Trier et al. (2020). In this paper the occurrence of CAT around a mid-latitude cyclone is investigated with a special emphasis on its relation with gravity waves. This paper would also be relevant in the discussion, with the caveat that the small-scale waves are likely not resolved in ERA 5.

p 6 line 191 : There is a typo in the definition of Q, with an extra $\times$ which should be removed (the dot is the conventional notation for the scalar product). Also, is it the ‘full definition’ which is used here, with $\mathbf{\Omega}$ the vector of angular rotation? Although I imagine the differences will be small, I believe it should be replaced by $f \mathbf{k}$ where $\mathbf{k}$ for consistency with the primitive equations solved in the ECMWF model.

p6 line 196-197: I guess altitude is retrieved from the geopotential. Maybe state it explicitly.

P7 l 211: This paper is submitted within a special issue (WISE) and I guess the field campaign motivated the choice of the date. This could be mentioned here.

P10 line 251, figure 5a) and related discussion: Could you also show the equivalent of Fig. 4 a) on top of Fig 4c) which is shown here? This would emphasize the relevance of using tropopause relative coordinates and help understand lines 256-258.

p 13 line 305-306 and Fig. 6: Do you know how exactly this surface is defined in the ECMWF ? In particular, I am surprised that the PV=2 PVU surface crosses the equator in Fig. 6. If there is some adjustment at low latitudes in the ECMWF field it would be useful to mention it here.

P 18 l 384: low → lower . 40 m/s is not a particularly low wind speed even compared to the subtropical or eddy-driven jet.

P 18 l 391: I am not sure how the geographic distribution here can be compared with the radiosondes from 2 stations in Sunilkumar et al. (2015). Agreed, the stations are influenced by the TEJ but from two points it seems complicated to validate a geographic pattern.

A slight difference is that S2015 see this increase above the monthly mean CPT (their fig. 4), which might be 600 m (2-3 ERA5 levels) above the lapse rate (Sunilkumar et al., 2013; Munchak and Pan, 2014). Given the depth of the layer (1 km) used by the authors to investigate shear in tropopause relative coordinates, 600 m is significant. See also main comment 3.

p 19 l 409 : Could you provide the correlation coefficient ?

P 20 l 410: ‘neutral and La Nina conditions’: do you mean ‘ neutral and El Nino conditions’?

P 20 l 440: you might consider showing a scatter plot of N2 and S2 to demonstrate this

p 21 l 450: you might note that ERA 5 has been shown to represent realistically part of the gravity wave activity (e.g., Krisch et al., 2020; Podglajen et al., 2020), which justifies that Gws might indeed be responsible for the enhanced shear in the reanalysis.

p 21 l 459 and fig. 13: for comparison, you could depict the distribution of Ri for all values of shear in the same region as well as the distribution over a deeper layer, to determine whether or not the TIL is a region of low Ri number

Typos and suggested reformulations:

p 2 l 30: “an substantial” → a substantial

p 2 l 37: ‘linear wave theory’ → ‘linear theory’

p 6 l 185 : I think it is the pressure velocity $\omega$ rather than $w$ which is provided by ECMWF.

P10 l 255 : ‘compare e.g.’ → ‘compare with’

p 13 l 298 : ‘barclinic’

p 24 l 547: ‘fulfils’ → ‘fulfills’

p 18 legend of Fig. 11: “destails”

p 22 l 467: “e.g.” should be before the reference

p 24 l 544: I would remove ‘exceptionally’ since your analysis shows that this feature is not an exception

p 24 l 552 : operational analysis → ERA 5

References:

Hoffmann, L., Günther, G., Li, D., Stein, O., Wu, X., Griessbach, S., Heng, Y., Konopka, P., Müller, R., Vogel, B., and Wright, J. S.: From ERA-Interim to ERA5: the considerable impact of ECMWF's next-generation reanalysis on Lagrangian transport simulations, Atmos. Chem. Phys., 19, 3097–3124, https://doi.org/10.5194/acp-19-3097-2019, 2019.

Krisch, I., Ern, M., Hoffmann, L., Preusse, P., Strube, C., Ungermann, J., Woiwode, W., and Riese, M.: Superposition of gravity waves with different propagation characteristics observed by airborne and space-borne infrared sounders, Atmos. Chem. Phys., 20, 11469–11490, https://doi.org/10.5194/acp-20-11469-2020, 2020.

Munchak, L. A., and Pan, L. L. (2014), Separation of the lapse rate and the cold point tropopauses in the tropics and the resulting impact on cloud top‐tropopause relationships, J. Geophys. Res. Atmos., 119, 7963– 7978, doi:10.1002/2013JD021189.

Pan, L. L., Honomichl, S. B., Bui, T. V., Thornberry, T., Rollins, A., Hintsa, E., & Jensen, E. J. (2018). Lapse rate or cold point: The tropical tropopause identified by in situ trace gas measurements. Geophysical Research Letters, 45, 10,756– 10,763. https://doi.org/10.1029/2018GL079573

Podglajen, A., Hertzog, A., Plougonven, R., and Legras, B.: Lagrangian gravity wave spectra in the lower stratosphere of current (re)analyses, Atmos. Chem. Phys., 20, 9331–9350, https://doi.org/10.5194/acp-20-9331-2020, 2020.

Trier, S. B., Sharman, R. D., Muñoz-Esparza, D., & Lane, T. P. (2020). Environment and Mechanisms of Severe Turbulence in a Midlatitude Cyclone, Journal of the Atmospheric Sciences, 77(11), 3869-3889. Retrieved Mar 9, 2021, from https://journals.ametsoc.org/view/journals/atsc/77/11/JAS-D-20-0095.1.xml

This study presents a climatological investigation of vertical wind shear zones at tropopause levels in a 10-year dataset of ERA5 reanalyses. The authors systematically study the frequency of occurrence of high vertical wind shear (based on a threshold criterion) in a framework relative to the lapse rate tropopause. They find that high wind shear mostly occurs within the lowest kilometres above the tropopause. The geographical distribution, vertical structure and temporal variability is examined for different latitudes and jet systems, and the physical processes associated with the shear layer are discussed, as well as its potential impact on the tropopause structure, stratosphere-troposphere exchange and turbulent mixing. 

The paper is interesting and well-written, it is put into context of existing literature, and the study makes a relevant and new contribution to the understanding of the UTLS structure. I recommend publication subject to a few minor revisions.

General comments:

1) Generally, the existence of wind shear above the jet core and in the lower stratosphere is not a surprise as is expected from balanced dynamics, the exact structure of the shear zones however are more involved. The authors mention the relation of the shear layer to the thermal wind balance at several instances in the manuscript along with other mechanisms. How much of the structure of the shear layer can be explained by the thermal wind relation? It should be possible to quantify this based on the ERA5 fields. The possible role of gravity waves is mentioned in several sections and maybe this way the magnitude of their contribution could be narrowed down.

Furthermore, I would suggest to emphasize more clearly in the introduction, perhaps in a single summarizing sentence, what the main unknown aspects of the shear layer are (e.g. detailed structure, strength, vertical extent and occurrence in a statistical sense, formation mechanism) and which of these aspects are addressed in the study.    

2) I think the authors should reconsider some expressions and definitions related to the shear layer phenomenon. 

- The words "enhanced" or "exceptional" are used frequently. In what sense is the wind shear "enhanced", compared to what reference? The study shows that the layers of strong wind shear above the tropopause occur rather frequently and strong wind shear is certainly not exceptional near jet streaks.  

- The tropopause shear layer (TSL) is defined based on an occurrence frequency criterion. In this sense, it is a purely statistical feature. Since a layer of strong wind shear also seems to be physically present and nicely visible in instantaneous synoptic situations with a strong jet stream (see Fig. 13b), I find it unfortunate to define the "TSL" in a statistical sense rather than as a synoptic feature. It would be more intuitive to call the regions indicated by the red contours in Fig. 13b "TSL". 

3) The authors have chosen a well-considered threshold S_t^2 and the choice is sufficiently explained. However, it would be interesting to test how sensitive the results are with respect to the threshold. How would the pattern of the occurrence frequencies change if S_t^2 was even higher? 

4) I would be curious if the statistical analysis (for the midlatitudes) has also been done relative to the dynamical tropopause and whether there are any differences compared to the LRT-relative framework. This would be interesting e.g. in the context of many STE studies which focus on transport across the 2-PVU surface. 

5) The introduction is quite long, the authors might consider shorten it a bit if possible. 

Specific comments, suggestions and typos:

L13ff: Throughout the manuscript, the term "tropopause-based" vertical wind shear is used frequently. This expression is not very clear to me; does it mean "tropopause-relative", "near-tropopause" or "tropopause-level"?

L30: an --> a

L33: in return --> in turn

L36: to the --> its

L60 and all following occurrences: ° N --> °N, please remove the space between "°" and "N"/"E"/"S"/"W"

L61: for --> of

L67: ERA Interim --> ERA-Interim 

L67: data set --> dataset

L70: remove "on"

L93: data --> forecasts

L98: analysis data --> analyses

L105: presents --> constitutes

L109: causes --> cause

L173: analysis "of" a single day

L174: Sections --> Section

L178: Please check the date and time convections of WCD

L185: ERA5 provides omega in Pa/s, not w in m/s

L189: dynamic --> dynamical

L191: Please remove the symbol for the cross product and insert a comma after equation. 

L192: (with the angular velocity of the Earth, Omega). 

L196: How do you derive the vertical distance between the model levels, do you use geopotential?

L201: can not --> cannot

L202: for --> to

L203: It is a bit confusing to read about static stability in combination with the notation S_t^2.

L209: majorly --> mostly

Fig3: It would be interesting to see contours of S^2 in the snapshot vertical cross sections in addition to wind speed. This would illustrate not only the general position of the shear zones but also the spatial variability. 

L259-264: Here, it would be helpful to explicitly point out the different positions of the solid/dashed black lines in Fig. 5a. 

L267-270: While the schematic illustration in Fig. 5b is very helpful and easy to understand, I find the explanation in the text rather unclear. The authors might consider rewriting these sentences.

L298: barclinic --> baroclinic

L299: remove "is"

L315-316: I assume your background state is still latitude-dependent? From this sentence it is not clear if you also average over latitudes. 

Fig7a: What does the black dot indicate?

L326: Why did you choose exactly 51°N?

L355: DFJ --> DJF

L416: and "references" therein

L440-441: No co-location of TIL and TSL: Can you show this in a figure? Maybe a snapshot vertical cross section would do. Or perhaps something like a frequency distribution of N^2 in the lowest 2 km above the LRT, showing grid cells with S^2 > S_t^2 separately and comparing them to the N^2 distribution of all grid cells. 

L449-453: Are gravity waves (partially) resolved in ERA5?

Fig13b: Do the black circles indicate the LRT?

L487: I believe that throughout the summary the expressions "exceptionally pronounced/strong/enhanced vertical wind shear" and "enhanced tropopause-based vertical wind shear" are used synonymously. Perhaps consider using the same formulation throughout the paper (including introduction) to not confuse the reader. 

L528: dynamic --> dynamical