Review of “Warm conveyor belt characteristics and impacts along the life cycle of extratropical cyclones: Case studies and climatological analysis based on ERA5” by Katharina Heitmann, Michael Sprenger, Hanin Binder, Heini Wernli, and Hanna Joos

The paper presents both a climatological analysis and three detailed case studies of the evolution of warm conveyor belt (WCB) characteristics associated with extratropical cyclones in the wintertime North Atlantic. WCBs are first identified from Lagrangian trajectories computed in a systematic way in the ERA5 reanalysis, then different metrics are obtained for the characteristics of the WCB inflow, ascent and outflow using masks based on the trajectories. In agreement with previous studies, the results show a link between WCB ascent and cyclone intensification and precipitation at lower levels, and WCB outflow and ridge building at upper levels, where the cyclonic, anticyclonic and non-curved WCB branches are clearly distinguished.

The text is well written and the figures are of high quality. The merits of the study lie in its systematic and comprehensive approach using numerous metrics applied to about 5’000 cyclones and corresponding WCBs. This allows clear and robust results for the wintertime North Atlantic climatology, which are generalized to some extent to other oceanic basins. The downside of the approach is that the long description of complex methods and their application to both climatology and case studies may lose the reader. The main ideas and results are blurred in these sometimes lengthy descriptions, which involve many details but somehow miss the general picture.

Thus, I recommend the paper for major revisions before it can be considered for publication. General and specific comments are listed below to help improve the organization of the paper.

General comments

1. The main results of the paper are unclear. This is partly due to a strong focus on the methodology and to the lack of proper conclusions in Section 5, which discusses specific results without much hierarchy and misses more general statements (only one sentence about the results in Section 5.4).
2. The structure of the paper is imbalanced. On the one hand, the case studies are too detailed: the description of single panels at multiple times is repetitive and should be streamlined, and similarities and differences between case studies should be emphasized rather than each case described individually. On the other hand, case studies are helpful to illustrate the climatological analysis but it is unclear what should be learned from the case-to-case comparison beyond the existence of a case-to-case variability. In this regard, the comparison in Section 5.1 is too detailed and appears too late in the paper. An alternative structure would be to present the climatology first, then to (briefly) discuss the case studies in light of the climatology to emphasize their peculiarities.
3. The methods are complex, based on several steps and each involving some form of (arbitrary) criterion, which makes results hard to interpret. On the one hand, the methods would benefit from a general summary of the main steps and motivation. On the other hand, the complexity prevents easy interpretation and comparison with previous studies. The numerous metrics (e.g., number of trajectories, low/high-level PV) are defined in a too complex way to be informative per se, thus must be discussed to compare case studies or time steps only. Also, each and every criterion cannot be the subject of a sensitivity test but it must be clarified what is taken from previous studies and what is not (and why). These points are shortly mentioned in Section 5.3 but without much discussion and quite late in the paper.

Specific comments

Title The word “impact” has different meanings and is usually understood as casualties and damages; what kind of impacts is expected here?

l. 1 “global investigation”: although the approach is global, as illustrated by Fig. 1 and in the supplement, both case studies and climatological results focus on the North Atlantic only

l. 5 see above comment on impacts

l. 34–35 Is there a reference for the second part of the sentence, or is it a hypothesis?

l. 36–38 It is important to define the meaning of “characteristics” and “impacts” for the paper but this short paragraph is rather vague; many examples are mentioned in the next two long paragraphs, after which a clear definition for the scope of the paper would be helpful.

l. 84–85 The distinction between questions 1 and 2 is not obvious

l. 99–101 What is new compared to the WCB climatologies cited above?

l. 125 The resolution (6 hourly and 80 km) appears to be taken from ERA-Interim; this is fine but may deserve some comment.

l. 132–133 What is the difference between “at any time during the 48-hour ascent” and “strictly between the start and end of the ascent, 48 h later”?

l. 135–136 The sentence is confusing

l. 142–244 This is interesting but questionable, as several criteria are different, as well as the dataset

l. 159 Any motivation for this value?

l. 168–170 This sentence is disconnected from the rest

l. 175–176 Is the “enhanced frequency of WCB inflow in the region of the storm tracks” not merely a consequence of “a minimum of one trajectory per WCB bundle must at least once coincide with a cyclone mask during its 48-hour ascent”?

l. 176–181 The frequency values require some kind of calibration, otherwise they are hardly usable as such.

l. 181–187 This supports the use of vertical position instead of relative time but has little to do with the use of WCB masks, which requires more motivation considering the above limitations

l. 197 Lagrangian properties to contrast with the following Eulerian properties?

l. 231 The proportion of non-curved trajectories is quite high (two third of the total), while a number of them seems to follow the anticyclonic ones on Fig. 3

l. 223 Why the asymmetry?

l. 232 Altitude is not the best name for a pressure value

l. 261 The Bergeron unit is not defined

l. 288–289 Repetition of the reference

l. 303 The location of the developing cyclone is hardly seen on Fig. 3a

l. 310 Same comment as above, and is the cold front shown somewhere?

l. 320–330 “almost perfectly”, “considerably”, “most likely not yet strongly”: overstated

l. 337 This is hardly seen on Fig. 5f

l. 343 Remind the definition of ULPVA?

l. 350 What is “because of the low altitude and latitude of the WCB outflow in this region”?

l. 355 In what sense is it similar?

l. 380–382 The WCB impact on cyclone intensification is disputable, as both WCB intensity and cyclonic proportion are delayed compared to the deepening rate

l. 416–418 It is surprising to realize that the chosen case was illustrated above but not mentioned

l. 420 Of which trajectories?

l. 425–426 Cyclonic or anticyclonic branch in Martínez-Alvarado et al. (2014)?

l. 487–489 This sounds speculative

l. 492–495 This discussion breaks the flow and does not appear too relevant as PV is followed in the WCB mask but not along trajectories here

l. 503–507 This case study should likely be presented first, as it is discussed and illustrated in Sections 1 and 2 as archetypal WCB

l. 525–539 The described features (frontal wave, secondary airstream, trajectories ascending at lower latitudes) are interesting but not easy to identify on Fig. 10

l. 583–584 Unclear

l. 593 Why does “the movement of the WCB ascent region from low to high latitudes explain the decrease in the WCB ascent rate with time”?

l. 618 A comparison of the three case studies is expected here

l. 636 Why is it “intriguing”?

l. 671 The contrast looks quite weak

l. 675–676 Not sure what to learn from this and cyclone intensification lasts for longer than 6h

l. 683 “very likely”: is it or not related to intense convective precipitation?

l. 694–697 This questions the relevance of the ULPVA metric, which likely depends on the number (intensity) of WCB outflow trajectories but also on the extent of the corresponding mask

l. 698–717 This detailed description of supplementary figures likely belongs to the supplement

l. 723 Panels g-i in Figs. 5, 8, 11

l. 741 “lowest” is misleading for the highest pressure value

l. 748–750 This very short summary does not support the need for detailed case studies

l. 756–799 At that point of the paper, general conclusions are expected about what should be learned from the case studies, rather than a detailed listing of case-to-case comparison

l. 762 larger but opposite

l. 806–813 This is interesting but contradicts the WCB contribution to cyclone intensification by diabatic low-level PV production discussed everywhere else in the paper

l. 814–818 This is also interesting but is not mentioned before, thus does not summarize results

l. 833–834 Any explanation for this?

l. 835–840 This suggests that the latitudinal dependence of the Coriolis parameter is solely responsible for the LLPV evolution, while the WCB evolution discussed in this paper does not play any role

l. 848–850 This sounds speculative

l. 867–869 Why not try them?

l. 870 The purpose of this subsection is unclear, as it summarizes the methodology rather than the results (which are already summarized in 5.1 and 5.2)

l. 874, 878 novel vs new climatology

l. 882–883 positive PV and negative PV anomalies

Figs. 6, 9, 12 Changing scales between figures does not help comparison

Fig. 15 When two curves show the same variable, a common scale would be more appropriate

SUMMARY

This paper presents an analysis of warm conveyor belts (WCBs) in ERA5. Lagrangian trajectories are used to identify the WCB and a spatial mask is applied to associate the WCB to its impacts. Results are demonstrated in two parts. First, three separate case studies are analyzed, compared, and contrasted. Second, a climatology spanning the 44-year ERA5 data set is presented. The paper is dense and contains many interesting results, but one of the more robust results is that WCBs are typically most intense when the cyclone itself is deepening most rapidly.

Overall, this paper is well written, the figures are well presented and clear, the methodology is appropriate and clearly described, and the conclusions are supported by the evidence presented. This work represents the latest installment in a line of meticulously-conducted studies of WCBs leveraging LAGRANTO. I am eager to learn about follow-up work utilizing the same methodology but applied to climate model simulations.

I am pleased to recommend publication after a minor revision. While I have no major concerns that would merit extensive revision of this paper, I would like to make several points for the authors to consider in future applications of this methodology. These points are expressed below. The paper is long and ambitious in scope. I make this as an observation and not an implicit recommendation to break it up into several shorter papers. That said, this work could probably have been distributed over two papers, although I see no problem fundamentally with long papers.

MAIN COMMENTS

1. Diagnosing upper-level PV anomalies.

On lines 252 – 254, the method for diagnosing PV anomalies is described as follows. “ To quantify this impact, we first vertically average PV at all grid points inside a WCB outflow mask between 200– 375 hPa. The monthly 42-year climatology of vertically averaged PV over the same pressure range is then subtracted to get a PV anomaly. The subsequent upper-level PV anomaly (ULPVA) is defined as the median of the anomaly values of 255 all grid points inside the WCB outflow mask.” I am concerned that this method does not isolate the diabatic contribution (as implied on line 250). Would not an amplified ridge be guaranteed to host negative ULPVA? It is difficult to see how this metric could distinguish adiabatic Rossby wave amplification from diabatic enhancement. Some discussion and context would be helpful.

2. Masking technique

The WCB masking procedure (e.g., as illustrated in Figure 2) identifies the “impact” area to contain all points within a 100 km radius of particle trajectories. I support the rationale for defining an extended “impact” area to associate WCBs to precipitation and PV modification. I have concerns about the appropriateness of using a circular area drawn around trajectories, specifically for PV. Many particles in the WCB outflow are likely to accumulate along the edge of the tropopause (i.e., along the periphery of the downstream ridge). This is a region of very large PV gradient. Is there a concern that the circular mask encompasses a volume of air that is on the poleward side (i.e., above the tropopause)? Wouldn’t this create a very large positive bias in the estimated ULPVA? Have the authors experimented with smaller masks? How sensitive is the ULPVA to the radius of the mask? Perhaps the masking is more appropriate for precipitation and less appropriate for PV?

3. Variance in WCB characteristics

Even a small subset of cases, like that presented in Section 3, demonstrates a large case-to-case variability in WCB characteristics. Despite this variance, this paper also demonstrates that there are some robust similarities (e.g., in the relationship between storm intensification and WCB intensity). While this paper highlights those robust similarities, it devotes less attention to the variance. This is perhaps something for a future study, but I’d be interested to know more about the variance. For example, how much is explained by low-frequency modes of variability (e.g., PNA, NAO)? Is there any clustering of characteristics (e.g., are there distinct groupings of storms with similar cyclonic vs. anticyclonic branch structures)? This dataset is begging for such an analysis to be performed.