Warm Core and Deep Convection in Medicanes: A Passive Microwave-Based Investigation

Round 1

Reviewer 1 Report

 

Review of the paper Warm core and deep convection in Medicanes: a passive micro-wave-based investigation by Panegrossi et al.

 

This paper presents satellite diagnostics, applied to six Medicanes where Ianos was used as an exemplary case. I read the paper with great interest and found the new methods presented here of excellent timeliness, well aligned with current discussions in the field and potentially very useful to advance the state of the art of Mediterranean cyclones. My general impression is certainly positive, I really appreciated the rigorous observational analysis and therefore I would highly recommend eventual publication.

 

I believe however that there is quite a room for improvements. I have three major comments and several ones of minor nature (most of them related to the major comments).

 

Major comments:

 

1) Scope, focus and organisation

The paper is of very technical nature. There are many acronyms and many different observational details. This is good but less attractive to non-experts on satellite observations and the specific instruments. It is quite hard to follow description of the results, digest information and even more difficult to relate to dynamics. I would highly recommend to split sections 3.1 and 3.2 (by the way there are two “3.1” sections) in smaller ones. Please establish a more efficient relationship to physical processes and dynamics with the observations and thus discuss results accordingly (see next major comment). In these regards, lines 139-146 provide three objectives:

 

Objective 1: This is well achieved, in much detail.

 

Objective 2: In my opinion this is a very ambitious objective. I am not sure if reaching the end of the paper I can relate WC with the intensification of the six cyclones. For instance, in Table 1 there are different intensities provided in terms of MSLP but I am not sure about the relationship to DC or WC diagnostics provided. Line 699 states that Ianos and Zorbas are the two most intense ones but table 1 shows Qendresa to be deeper than Zorbas. Also I am not sure whether it is demonstrated here that intensity (at least for Ianos and Zorbas) is a function of DC alone. Actually in Table 1 Ianos and Zorbas are classified as “A”. The other cyclones as “B”. According to Miglietta and Rotunno (2019), “A” means that cyclones fully comply with WISHE while “B” there is an important baroclinic forcing. It would be useful to provide more information how classification was done for all cyclones. If this requires much of analysis probably this objective should be omitted.

 

Objective 3: The paper is convincing that there is much potential. Nevertheless demonstration of insights into the relationship between WC diagnostics and cyclone dynamics are rather weak. I would say that objective 1 could be better accompanied with a brief analysis of dynamics. I would like to see overlayed fields of the atmospheric state from analysis or reanalysis to better understand how the observations connect with atmospheric dynamics.

 

2) Discussion on dynamics

Relationship to cyclone dynamics is largely based to phenomenological similarities with past papers mainly on TCs. The introduction assumes that Medicanes are all share similar dynamics with TCs. In these regards interpretations about e.g. absence of lightning is done in function of TC papers. Scales of convection and symmetries are demonstrated but do not seem to be comparable to the ones for TCs even for the cases of group “A”. The paper would benefit from better linking dynamics with the observations. In these regards some additional figures or overlayed meteorological fields would be very beneficial.

 

3) Balanced presentation of cases

I appreciated the detailed figures for Ianos (Figs 1 and 3) but to acquire deeper insights into the usefulness of the diagnostics and the variability met in different cases, it would be highly recommended to show similar diagnostics for all cases. Therefore, differences could be discussed with regrads to the location and spread of the different impressions of Wcs.

 

Minor comments:

line 41-44 : A small fraction of cyclones are presenting "visual characteristics" of tropical cyclones (cloudless "eye"). The use of cyclone phase space (CPS) might provide hints about axi-symmetry and WC structure but I am not aware of any study that provides a climatological overview of these criteria alone or combined with the arbitrary criterion of a cloudless "eye". Please rephrase with an attitude of hypothesis.

 

Line 44-45: "shallow WC structure" does it refer to CPS diagnostics? Please be more precise. Also "weak frontal activity" please rephrase to "weak fronts".

 

Line 52-53: The mechanism of tropical transition has been proposed or discussed for some case studies but here it is given as a fact. I suggest you start with "Medicanes have been suggested to develop… (References)".

 

Line 53: Please remove "Even today".

 

Line 55: Probably it is better if you cut-paste lines 49-51 after the phrase ending: "..presence of a WC during the TLC phase. Medicanes' definition...". Then you could start a new paragraph with "Model-based diagnostics..."

 

Up to Line 55: The terms of TLC and WC need to be explained in more detail. They are throughout considered as given knowledge but their physical description is of high importance for the results.

 

Lines 76-80: As before, please introduce these lines as suggested classification by other authors. The way you provide this information is as if it is common knowledge.

 

Lines 86-88: Therefore most systems might fit Category C as explained above. Please synthesize results.

 

Lines 89-108: This is a nice overview but seems disconnected to the dynamics-centered parts, discussed earlier. Please articulate better by explaining the connection to WC. Basically I miss a reply to a "so what" question.

 

Lines 114-116: Not sure if reference [25] refers to Medicanes. Instead, as also [24], they refer to tropical cyclones. In these regards, the peak of lightning activity comes while a tropical cyclone is already well formed and is quite close to reaching maturity (maximum intensity), i.e. 24 hours before. Given the much smaller temporal and spatial scales of Medicanes, a tropical cyclone is hardly formed 24h before maturity and probably is still in the stage of what is referred earlier in the introduction as pre-TLC phase (i.e. under the influence of baroclinic forcing?). Despite the phenomenological similarity concerning the lack of electric activity at maturity, in terms of dynamics, the similarity between Medicanes and TC dynamics (and probably relevant interpretations on DC) seems rather weak. In any case, I believe that this part should be more elaborated and discussed in more depth.

 

Table 1. How was categorization done in Groups A and B? Is this out of analysis of the authors based on past articles? If yes, please provide more information.

 

Lines 349-351: Wouldn't this finding suggest lack of symmetry?

 

Lines 381-384: Is the displacement of MSLP minimum respect to WC comparable to ERA5 grid spacing. If not, would this suggest that ERA5 has a bias in the location or that the WC is not aligned with the cyclone center. Actually the use of wind scatterometers (if available) would be helpful to provide insights into the alignment of the WC with the cyclone center. If it is not aligned, then how would you explain this inconsistency?

 

Section 3.1: How was a "well-defined, isolated WC" defined (line 413)? Is this an arbitrary classification? Given you only have six cases, please provide similar figures as for Ianos for all cases (probably only for the most important panels). Table 2 is rather difficult and too technical to understand. In addition, the discussion of the results relies much on technical aspects (grid points, number of overpasses etc.).

 

Line 428-429: I am not sure what is meant by "WC structure was captured by 59% of overpasses. Does this mean that the WC was not well defined in 41% of overpasses? Please discuss in function to dynamics and physical processes and any relevant uncertainty if linked with technical aspects.

 

Lines 491-492 state that "Qendresa DC activity" is very limited". This is quantified in terms of "pixels". It is hard for readers with no experience with these measurements to understand and link DC with cyclone dynamics. In these regards, it is rather unclear with what is meant by "limited DC activity". Does this mean that DC is not frequent in the cyclone lifetime? Is the DC spread or concentrated to a single cell? What does this mean for the cyclone intensity? How does this compare with conceptual structure of cyclone thermodynamics (tropical or extratropical)? Even a direct comparison with other cyclones in terms of pixels would be useful (e.g. tropical cyclones have typically 1000 pixels of DC within 3 hours).

 

Line 431: "All" refers to the six cases?

 

Lines 451-459. This seems speculative. Probably I am wrong but in reference [16] there is only the analysis of one cyclone with lightning data (Rolf, 2011) and here is compared with reference [21] that still on Ianos. I suggest you remove this part.

 

Lines 460-463: From table 2 it seems that the min/mean/max distance is close to ERA5 grid spacing (~30 km). I would say that this is more than reasonable.

 

Lines 481-483: The mean eyewall radius for TCs is typically less than 60 km even for the strongest systems. Figure 4a suggests higher distances (also in Fig. 1). In my understanding, a radius as it is shown here suggests that the cyclone is either much larger than a tropical one, or that you address a different kind of mechanism (other than a convective eyewall). Could you please comment on scales?

 

Line 493: What does is meant by this high variability in terms of dynamics? Again this comment goes on the similarity with tropical cyclones. What would it be expected in tropical storms about DC? I would guess a much larger spread in the boxplots of Fig. 4a, but the median would be close to 30-60 km where you get the eyewall. Actually It would be useful to compare with Fig. 7 of Flaounas et al. (2017):

 

https://link.springer.com/article/10.1007/s00382-017-3783-x

 

It seems that the findings here are not so much different from what was found in cyclones climatology (in terms of distances) from other intense cyclones. Could you please comment?

 

Line 513: It would highly useful to also compare DC with times before WC was reached and relate with different stages of cyclone intensity (mainly genesis, intensification and maximum intensity stage).

 

Figure 5: DC close to Ianos seems to range from ~0 to 2 degrees. This would roughly cause the distribution in Fig. 4a to range from close to 0 until 175 km. Maybe I miss something here? Probably Figs 4 and 5 could be discussed together.

 

Figure 5: How the lack of symmetry shown here compares to tropical cyclones? What would you expect in a tropical storm or even from a mature hurricane?

 

Lines 546-583: This last section seems to be of technical nature and disconnected from the previous part (after all it makes reference to figs 1 and 2). This part confirms previous discussions on the choice of variables to identify WCs and does not add much to the results. I would either shorten it -dramatically- and insert it to the methods section or remove it entirely. Actually the way it is discussed here does not seem (to me) to add much to the three objectives.

 

Lines 606-607: I agree that diabatic forcing might be much stronger in Ianos and Zorbas but I am not sure how strong is the impact of upper level dynamics on the cyclones intensity. For instance, Portmann et al. (2020) show that Zorbas location at 28 Sep 2018, 1200 UTC, is strongly dependent on the upper level dynamics. Their figure 2 suggests that Zorbas is the outcome of warm seclusion and while you detect a WC, there is a cut-off low over the cyclone center, obviously separated by a preceding PV streamer. Could this suggest that the WC detected here is of different origin rather than of subsidence as in TCs? In these regards, would it be possible that WC may not be related to WISHE as suggested in line 608?

 

Portmann, R., González-Alemán, J. J., Sprenger, M., and Wernli, H.: How an uncertain short-wave perturbation on the North Atlantic wave guide affects the forecast of an intense Mediterranean cyclone (Medicane Zorbas), Weather Clim. Dynam., 1, 597–615, https://doi.org/10.5194/wcd-1-597-2020, 2020.

 

Figure 3: In relationship to the previous comment, it seems that the trough in Figure 3c is related to a warm anomaly (widely spread to the west of the cyclone center). Could a cut-off low be related to a warm anomaly and therefore give the impression of a WC? IF yes could e.g. this explain the dislocated WC and cyclone center in ERA5?

 

Lines 585-616: seem to be more of a summary than a discussion. I would suggest to relocate this part or change the title of section 4.

 

Figures 2&3 please correct dates (2020 instead of 2022).

 

Lines 627-628: I do not understand what is meant by a "bottom-up WC development". I am not a WISHE expert but I think that WC in the upper troposphere comes out of subsidence and conservation of equivalent potential temperature. I understand that observation diagnostics are showing a WC development but this hardly demonstrates relationship to WISHE. It is certainly good to discuss to discuss similarities but uncertainties need also to be stressed.

 

Line 630: What is meant by "Cold low transition to a WC system"?

 

Lines 630-638. This is an interesting discussion. I highly recommend to link to your findings and discuss accordingly. For instance, how's latent heat (or at least precipitation as a proxy of it) in your cases? It would be nice to demonstrate that the WC is not related to latent heat here, or not.

 

Lines 642-644: I do not really understand this part. The "WC in lower channels" how can it be merged with a warm sector? It seems that you describe the merging of "a diagnostic" with a "physical structure". Please be more precise about what kind of physical structure would be merged with the warm sector and please provide an example of it with a reference.

 

Line 647-648: Why WC is meant to be the "actual center"? Deviations might be due to comparing different datasets, i.e. models and observations.

 

Line 673: From memory, Qendresa was found by DiMuzio et al. to be a cold core cyclone (using ECMWF analysis) and Carrio et al 2017 found that upper tropospheric dynamics were found to be important to cyclone development. You may want to compare results with their own (especially DiMuzio et al (2019) have analyses few of the cases listed here).

 

doi.org/10.1016/j.atmosres.2017.07.018

doi.org/10.1002/qj.3489

 

Line 680-682: How this result is deduced? Absence of lightning does not necessarily implies same mechanism as in TCs.

 

Lines 685-687: Please compare with cyclones climatology using Flaounas et al., 2017 (see comment above). This would put the six cases into climatological context.

 

Lines 701-705: This part insinuates that intensification is only a function of convection but we are in place from previous studies (cited in the paper) that baroclinic forcing might have large share to the intensification. The favorable location of convection could e.g. be related to specific tilting of a PV streamer. Probably this part needs to be revised.

 

Line 708-711: Could you please comment on how scales and symmetry of the six Medicanes compare with the results in reference [58].

Author Response

Revierwer 1 answers

Author Response File: Author Response.docx

Reviewer 2 Report

In the present paper passive microwave measurements are used to analyze the characteristics of six Medicanes. Warm core structure and deep convection properties of the Medicanes are examined using direct measurements or derived products (i.e. cloud top height and ice water path). Both the data and the methods are adequately analyzed and described (or relevant references are given). The results are presented and discussed in detail. The paper can be published as is.

Author Response

Reviewer 2 answers

Author Response File: Author Response.docx

Reviewer 3 Report

This paper uses passive ｍicrowave and lightning data and physical quantities calculated by the DEEPSTORM, a machine learning method to determine variable importance for warm core occurrence and warm core type of Medicanes. Despite the small sample size and the application limitations of the ML method, the results are very interesting.

I would like to ask several questions regarding the scientific basis of this paper. I would also like to request additional analysis using the ERA5 atmospheric reanalysis data. Thus, I would like to recommend major revision.

【Major comments】

・Even though the authors deal with 6 Medicane cases, the data that can be used for the analysis is limited to that passed over the targeted disturbance at a certain time. On the other hand, each Medicane varies from time to time, and then the Medicane phase is expected to change with each hour. Table 1 shows that the duration is a few days, and Figure 1 is a snapshot. It is not clear from the paper whether the durations in Table 2 were calculated based on the snapshot data like Figure 1. Also, if the best track exists for each medicate during the periods in Table 1, the authors should post the time variation of central pressure. In other words, please show why one medicane can only have one type (A,B or C) within a period of time.

・When considering the phase transition, a description on the relationship between a Medicane and  the upper trough is ambiguous in the current manuscript. Since the ERA5 atmospheric reanalysis data set is used in this study, the time evolution of the potential vorticity should be shown. This allows readers to understand the relationship between the time evolution of the Medicane and the vorticity evolution in a synoptic field, and also makes the relationship between the three A, B, C types clearer.

・I do not understand in the current manuscript how the results of the analysis on dynamics are presented. In my impression, physics and thermodynamics would be better. The enhancement of WC through adiabatic heating due to stratospheric intrusions is not understandable to readers unless the authors make additional analysis using ERA5 data.

Also, the term 'upshear' comes up in the current manuscript. Since the authors use ERA5 data, perhaps they should calculate the vertical wind shear in the analysis domain and discuss it in combination with the frequency of DC occurrences.

・Regarding the discrepancy in the center position of medicanes between ERA5 and the satellite observations, I believe that the difference in time must also be taken into account. In addition, it is necessary to describe whether the satellite data used in this study were assimilated in ERA5 or not.

・Regarding the analysis using a random forest, one of the machine learning methods, the distinct difference in the value of variable importance between the DEEPSTORM results and the TB results may be due to the lack of normalization and also due to the difference of the coverage of the output by DEEPSTORM. This suggests that the analysis may not have been performed perfectly. I would also like to see an explanation of how the random forest analysis was performed with a figure such as a tree chart.

【Minor comments】

L32 IPW should be replaced with "IWP".

L63-65: Please show the reference(s).

L72-: Are categories A, B, C in the Introduction a phase in general? Or are these categories to be considered a phase in this paper? What is the rationale for considering them as a phase?

L118: PMW should be written down here.

L123-125: The location of '(TC)' 

L144: The word 'intensity' means a central pressure of a Medicane?

L255, L276: Why do the authors set the domain of 7°x7° (or 5°x5°)?

L266: and

Figure 1: Information of vertical wind shear in a domain is desirable.

L306: Section number 3.1 should be replaced by 2.5.

L417: How do the authors calculate the anomaly of TB.

L444: How can readers understand the occurrence of the diurnal variation detected in the afternoon?

L457: The authors should show the evidence on the inhibition.

L613: potential vorticity

L639: The authors should show the evidence on the effect of vertical wind shear.

L670: TCs

 

Author Response

Reviewer 3 answers

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

 

Review of the revised manuscript entitled “Warm core and deep convection in Medicanes: a passive micro-wave-based investigation” from Panegrossi et al.

 

The authors made substantial effort to revise the manuscript. The content and discussions are greatly improved and I really appreciate the rigorous replies to my comments. Right below I would like to raise few issues which are related to the authors’ replies and that might need a bit more attention when discussed in the manuscript. I regard the following as issues of minor concern and as such I fully support the approval of the paper for publication.

 

1) Regarding displaced MSLP centre and WC, the authors state:

“Such inconsistency is due to two main factors: the uncertainty in the WC center definition linked to the low spatial resolution of the AMSU-A 50 GHz channels especially at the edge of the radiometer swath, and possible mismatch with ERA5 regular grid, and the fact that the channel used for the WC center location peaks around 400 hPa.”

It may also be related to model bias in the reanalysis. This is quite common for instance when comparing tropical cyclone best tracks and tracks derived from ERA5.

 

2) Regarding radii of convective eyewalls:

Thank you for pointing out Tsukada and Horinouchi, 2023. As far as I understand from their tables 2 and 3, there is only one case (shown twice) out of 52 where the “eye” radius is more than 100km where almost all of the rest cases is less than, or about, 60km. Please note that their study also included arbitrary distinction between cases with and without convective eyewalls where especially the case of Trami (2108) that was measured to have an eyewall of 160 km was quite a hard case to diagnose with objective methods (See Tsujino et al., 2021). In these regards, the following phrase seems quite misleading:

“However, according to [60], the size of the eye in TCs varies a lot from ~10 to ~160 km in radius. Especially for relatively weak TCs with “unclear eyes”, our observations of the size of Medicane WCs and the distance of DC activity around the WC center are comparable with TCs.”

Please revise accordingly or remove this phrase. Also the previous phrase reads:

“Such variability is an indication that baroclinic forcing might have a large share to their intensification [59].”

Is it meant that DC away from the cyclones centre is related to large scale ascent? If yes, please replace accordingly “baroclinic forcing”.

Tsujino, S., Horinouchi, T., Tsukada, T., Kuo, H.-C., Yamada, H., & Tsuboki, K. (2021). Inner-core wind field in a concentric eyewall replacement of Typhoon Trami (2018): A quantitative analysis based on the Himawari-8 satellite. Journal of Geophysical Research: Atmospheres, 126, e2020JD034434. https://doi.org/10.1029/2020JD034434

 

3) Regarding WC formation and the example of Zorbas:

I appreciate the discussion provided in the reply document along with the figures from Ntafis et al., 2020 and the commenting in the manuscript. I am not sure though I completely understand the arguments here:

“At a later stage, on 28 Sep. 2018, … strong DC had eroded the upper-level PV streamer suggesting that the observed WC at upper levels is due to DC activity close to the cyclone center.”

Figures from Ntafis et al. are provided in iso-potential temperature surfaces so it is quite hard to easily deduce any (potential) temperature anomaly from PV fields. With this in mind, from panel d (morning of 28 of September), it is quite evident that stratospheric air resides over the cyclone centre and this could be related to warmer air intrusion in higher altitudes. At 19z times of the same day (panel e) high values of PV are still over the cyclone centre probably also contributing at some extent to the WC. So WC could be partly due to “remnants” of stratospheric air. Given the scales and details in the panels, I guess that the figures here come from model simulation. Therefore, all argumentation here also entails uncertainty especially given the “pixel-level” details that you analyse here.

If however, the simulations are indeed very close to the observations, I would like to stress again that as far as I understand tropical cyclone dynamics, the WC is formed within the cloudless calm eye and and is not due to DC but due to subsidence (adiabatically heated air). Panel d of Ntafis et al. and the argumentation :

“In this case, a bottom-up WC development is indicated (i.e., upper-level trough has weakened but the WC has a more clear definition), which is a characteristic of Category A Medicanes in [16].”

both suggest that the WC is due to latent heat release. In this case, we have a system where convection has a positive feedback to the cyclonic circulation but its WC is due to latent heat release and not due to subsiding air. As a result, dynamics are different from TCs and thus fitting this case in “category A” should be discussed or revised accordingly. I am not aware of other studies showing so clearly the WC relationship to convection in the Mediterranean from the perspective of observations and I commend the authors for that. This is a very important result, it exceeds phenomenological similarities and needs to be stressed in the text after revising it accordingly (here and elsewhere). This should be also done in the conclusions probably where the text reads:

“In this regard, analogies with TCs could be further investigated in the future…...”

 

 

 

Author Response

We thank the reviewer for his/her comments and we attached a file with a point-by-point responses.

Author Response File: Author Response.docx

Reviewer 3 Report

Since the instruction from the editorial office was to respond within three days, and I did not have time for the review during this period except for today, now I would propose the recommendation after reading the manuscript several times.

Anyway, I thank the authors for conducting major revision. However, I am not satisfied with the authors' response regarding my previous comments. The essential point is the revised descriptions as follows. "The group for each Medicane according to the classification by [16] is also indicated. It is worth noting that the authors classify the cyclones based on whether they reach a milestone phase, thus, there is only one category for each Medicane. Such categorization is based on the cited previous articles and it was made possible after discussing these cases with the leading author of [16]".

 

The above description indicates that Categories A, B, C were determined by a several members and cannot be classified by any other researcher. In other words, it refuses to be verified by the other scientists. For example, could I make the same classification on the six medicanes as the authors by working according to the fullness of what is written in this paper?

 

In other perspectives, why don't the authors show in this paper that WC and DC characteristics allow to classify Medicane?

 

There are also typos scattered throughout this manuscript.

(Examples)

・Emanuel(1986) should be Emanuel[12]

・ L242, 245, 358, 381, 393: WWLLN

 

Although the number of references has increased and the content has become voluminous, the basic result of this paper is that WC and DC can be detected with the techniques used in TC analysis and that the relationship with lightning around medicanes is similar to that around TCs. If the paper is simply written in a manner consistent with the two research objectives (L155-159), it meets a level of Remote Sensing sufficiently. 

 

I am stepping down as a reviewer as of this manuscript.

 

Author Response

We thank the reviewer for his/her comments and we attach a file with a point-by-point responses.

Author Response File: Author Response.docx