Summertime Rossby waves in climate models: substantial biases in surface imprint associated with small biases in upper-level circulation

Luo, Fei; Selten, Frank; Wehrli, Kathrin; Kornhuber, Kai; Le Sager, Philippe; May, Wilhelm; Reerink, Thomas; Seneviratne, Sonia I.; Shiogama, Hideo; Tokuda, Daisuke; Kim, Hyungjun; Coumou, Dim

doi:https://doi.org/10.5194/wcd-3-905-2022

Articles | Volume 3, issue 3

https://doi.org/10.5194/wcd-3-905-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/wcd-3-905-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 3, issue 3

Research article

|

08 Aug 2022

Research article |

| 08 Aug 2022

Summertime Rossby waves in climate models: substantial biases in surface imprint associated with small biases in upper-level circulation

Fei Luo, Frank Selten, Kathrin Wehrli, Kai Kornhuber, Philippe Le Sager, Wilhelm May, Thomas Reerink, Sonia I. Seneviratne, Hideo Shiogama, Daisuke Tokuda, Hyungjun Kim, and Dim Coumou

Download

Final revised paper (published on 08 Aug 2022)
Preprint (discussion started on 06 Aug 2021)

Interactive discussion

Status: closed

RC1:
'Comment on wcd-2021-48', Anonymous Referee #1, 08 Sep 2021

The comment was uploaded in the form of a supplement: https://wcd.copernicus.org/preprints/wcd-2021-48/wcd-2021-48-RC1-supplement.pdf

Citation: https://doi.org/10.5194/wcd-2021-48-RC1
- AC2: 'Response_Letter_to_RC1_and_RC2', Fei Luo, 05 Nov 2021
  
  The comment was uploaded in the form of a supplement: https://wcd.copernicus.org/preprints/wcd-2021-48/wcd-2021-48-AC2-supplement.pdf
  
  Citation: https://doi.org/10.5194/wcd-2021-48-AC2
RC2:
'Comment on wcd-2021-48', Anonymous Referee #2, 28 Sep 2021
This study compares model errors with a set of nudging experiments using multiple AGCMs. They find nudging the upper-level circulation can significantly reduce surface biases, while nudging soil moisture can’t lead to much improvement. This is a very interesting study that definitely deserves to be published. However, the interpretation and presentation may need some extra efforts. Below I list several concerns, mostly minor, for consideration during the revision process.

Total field or anomalies. I get confused at different parts of the manuscript whether they were talking about the total field or the anomalies of v250. Wavenumber 5 and 7 may be more prominent and longitudinally phase locked in the total field, but it’s unclear whether this is also true for the anomalies. I am afraid that for subseasonal variability and extreme events, what matters more is the anomaly, not the total field. I can’t help wondering why they chose to only focus on events that project strongly onto wavenumber-5 and -7 of the total field.

It would be far stretched to use the current experiments to address Question 3, whether model biases originate from the atmosphere circulation or land surface-feedbacks. First, prescribing soil moisture does not necessarily mean the model can accurately simulate land-atmosphere feedback. Secondly, it seems that the nudging they applied to the atmosphere circulation (above 700 hPa) is far more than just “the upper atmospheric levels”. Therefore, they may need to rephrase the conclusion that “small bias in the upper atmospheric levels can result in big bias in surface weather conditions” (ln 305). See 4. for another concern on drawing this conclusion.

Why are the pattern correlation coefficients in Figs. 5,6 all positive? What’s the geographical domain used to construct the Taylor diagrams?

Why unlike other three models, CESM (Fig.B1) B_land is not much smaller than B_atm? In fact, why are the magnitudes of B_tot, B_atm, B_land and B_res rather similar?

Significant test for the composite maps in Figs.3,4 is needed.

The manuscript needs serious editing.
Citation: https://doi.org/10.5194/wcd-2021-48-RC2
- AC3: 'Response_Letter_to_RC1_and_RC2', Fei Luo, 05 Nov 2021
  
  The comment was uploaded in the form of a supplement: https://wcd.copernicus.org/preprints/wcd-2021-48/wcd-2021-48-AC3-supplement.pdf
  
  Citation: https://doi.org/10.5194/wcd-2021-48-AC3
AC1: 'Response_Letter_to_RC1_and_RC2', Fei Luo, 05 Nov 2021

The comment was uploaded in the form of a supplement: https://wcd.copernicus.org/preprints/wcd-2021-48/wcd-2021-48-AC1-supplement.pdf

Citation: https://doi.org/10.5194/wcd-2021-48-AC1

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Fei Luo on behalf of the Authors (18 Dec 2021) Author's response Author's tracked changes Manuscript

EF by Polina Shvedko (20 Dec 2021)

ED: Referee Nomination & Report Request started (23 Dec 2021) by Irina Rudeva

RR by Anonymous Referee #3 (18 Jan 2022)

For final publication, the manuscript should be

accepted as is

accepted subject to technical corrections

accepted subject to minor revisions

reconsidered after major revisions

rejected

Were a revised manuscript to be sent for another round of reviews:

I would be willing to review the revised manuscript.

I would not be willing to review the revised manuscript.

Suggestions for revision or reasons for rejection

The current paper investigates the representation of strong “circumglobal wave events” in three different climate models. A particular focus lies on the role of the upper-level flow versus soil moisture for model biases. First it is shown that the three models in a free-running mode do a pretty reasonable job in representing the anomalies associated with these “extreme circumpolar wave” events as compared with reanalyses. Second, the authors use a valuable set of numerical experiments that was performed in connection with the ExtremeX project in order to find out whether this success is primarily related to the correct representation of the upper-level flow or to the correct representation of the soil moisture conditions. As it turns out, the upper-level flow is considerably more important in this analysis.

The topic is highly relevant, the paper is overall well written, and the logic seems straightforward in large parts of the text.

However, the longer I think about this paper, the less I understand what I am supposed to learn. Essentially you investigate composites of events, and the events are selected on the basis of the magnitude of one particular zonal wavenumber. Arguably this is a somewhat artificial selection, because it does not necessarily represent any physical/meteorological situation. For instance, if you select a specific wavenumber and the amplitude of this wavenumber is much larger than the amplitude of all other wavenumbers, the situation corresponds to a circumglobal wavetrain. However, if you select a specific large-amplitude wavenumber and at the same time neighboring wavenumbers are of similar (large) amplitude (something which I assume to be rather common), the meteorological situation is more likely to be a large-amplitude Rossby wave in part of the hemisphere with much smaller amplitudes in the rest of the hemisphere. It transpires that the events that you select (based on one specific wavenumber exceeding a threshold) may be a collection of rather different-looking physical situations, and they are not necessarily always associated with a circumglobal wave train. Taken at face value, this would draw into question even the title of your paper, which promises a study of circumglobal Rossby waves. In any case, if you average over many such separate events (through compositing), it is unclear (to me) what this composite is meant to represent.

Possibly I have not fully understood your analysis. But maybe other readers have a similar problem. Therefore, the authors would have to be more explicit and more lucid in their interpretation and work out much more clearly what one really learns from the analysis of this set of simulations.

My second issue is somehow related the previous one. A superficial interpretation of your results suggests that soil moisture is really not important; rather, all you need is to get the upper-level flow right. Maybe this is not how the results should be interpreted, but there is the danger that this impression remains (unless your interpretation becomes much more detailed and explicit). In particular, one of the authors (Sonia Seneviratne) has shown multiple times in previous publications that soil moisture is important to determine summer surface temperatures at least over certain areas over the continents. Again, on a superficial level, the current results seem to contradict these earlier results. I would be interested in Sonia Seneviratne commenting on this issue, and the reader would certainly appreciate if you could clarify.

In my interpretation, your analysis points to a possible model bias in the sense that the models systematically under- or overestimate the spectral power in specific wavenumbers as far as the upper tropospheric circulation is concerned. In your analysis this is somewhat obscured by the fact that you focus on just a few wavenumbers. In reality, the lack of spectral power at a certain wavenumber may reappear as a surplus of spectral power at some other wavenumber. In other words, your selection of events is based on highly incomplete information in spectral space and may, therefore, obscure what is really going on. A more straighforward approach to analyse spectral model biases would be to consider the composite spectral power of all wavenumbers and determine related biases. In addition, to the extent that the meteorology in the lower troposphere is “slaved” to the dynamics in the upper troposphere, nudging the upper troposphere makes obviously a difference, but nudging the lower troposphere does not; this would imply that your results regarding the biases are somewhat trivial.

What’s non-trivial and what I find highly interesting is the fact that individual wavenumbers apparently have a preferred phase. However, this is not the topic of the current paper, and the current paper does not (aim to) further contribute towards an explanation.

Last, but not least, I have an issue with the title, and this mirrors the points that I raised above. I think the title is misleading for two reasons. (1) It suggests that you deal with “circumglobal Rossby waves”, which I would argue is not true. You select events on the basis of highly incomplete information in spectral space, and this does not guarantee that your events are characterized by a circumglobal Rossby wave (see my earlier remarks). (2) You say that “small biases in upper-level circulation create substantial biases in surface imprint”. To me as a reader this suggests that if you select a specific event and introduce just a small upper-tropospheric bias in that event, this results in strong changes in the surface meteorology. But again, this is not what you have shown. The concept “bias” in your analysis represents a comparison between two composites, e.g., one from a free-running model and the other one from reanalysis data. However, the underlying individual events (from which the composites are computed) may be very different and the analysis, hence, does not support the claim raised in the title (the way I understand it). I realize that you tried to summarize the whole content of your paper in the title, but with an analysis as complex as yours this is invariably going to be impossible.

In summary, I am aware that the current paper uses a quite special (and maybe somewhat artificial and unphysical) setup for the analysis: namely compositing only events that have a large amplitude of a specific wave number. This very special sort of analysis may have a strong imprint on its interpretation, but it is not straightforward (for me) to see and understand this imprint. As a consequence, I am not fully able to follow your interpretations and to appreciate the merits of your analysis.

Minor issues

Line 50: arguably, RRWPs are not the same thing as quasi-stationary Rossby waves, so the former should not be given as “an example” for the latter.

Line 88: “with the duration of sfc weather conditions….”: that’s a somewhat strange formulation. Do you mean “long duration” or “persistence” here?

Line 137: “This added tendency term….”: this sentence is not clear to me, as well as the ensuing sentence.

Line 152: Since your analysis is based on weekly averages, you could (and should) say that you are interested in rather persistent anomalies. For instance, in the abstract you could talk about “summertime persistent (or: quasi-stationary) wave events” instead of just “summertime wave events”.

Line 157: What is an “imprint”? Please define! Here you talk only about surface “imprints”, but in the following plots you also show composite upper tropospheric meridional wind. How are the latter defined?

Line 168: The concept of a “bias” is central to your paper (the word appears even in the title), but the “definitions” that you provide on page 6 are not clear to me. In my understanding, AISI represents a free running model run, and ERA5 represents a dataset. So what is the difference between a model run and a data set? This is way too sloppy, this must be specified much more clearly. (You probably mean the difference between the composites, not the difference between model runs or data sets, but this must be said explicitly).

Line 190: better “variance in wave amplitude” instead of “variance in wave activity”, because “wave activity” has a special meaning in dynamical meteorology (which you do NOT want to refer to here).

Caption Fig 2: not clear to me what the term “bandwidth” refers to here.

Line 204: Well, ERA5 shows a strong preference for one part of the hemisphere as opposed to the other part of the hemisphere. Therefore, since you have not clearly defined “phase locking”, it is not obvious to me that ERA5 is supposed to not show phase-locking for wave-6 and 8.

Lines 229/230: english? In addition, I appreciate your effort to determine statistical significance, but I was not able to extract from Fig. B12 and B13 what areas are significant. Please clarify. Where on the plots should I see “highlighted fuchsia color”?

Line 293: “almost fully….” seems somewhat overstated, I would say “to a large extent….”

Line 304: Ref to Fig 7a seems wrong.

Line 316: Ref to Fig 5b seems wrong.

Line 329: what do the correlation values given in parentheses refer to?

Line 330: better “climatology and variability”.

Line 343: I think that the vertical wind in your nudging experiments is only weakly constrained by the divergence/convergence of the horizontal wind. Rather, having the horizontal wind almost right in these experiments implies the correct forcing in the omega equation and, hence, a good representation of the vertical wind. In other words, I think your statement is correct, but the reason/explanation you give may not be correct.

Line 347: “can be disturbed in the models….”: that’s unclear. I agree that an error in the vertical wind may lead to an error in cloudiness, but there may be other sources for errors in cloudiness (e.g., moisture advection) which may be just as important.

Line 363: froced  forced

Line 371/372: do you mean that they yield cirumglobal waves “by design”!?

Line 373: What is the “stationarity of a data set”? also: I hope very much that the results do NOT depend on the method used to compute the FFT! Can you explain what you mean here?

Line 374: I do not understand this. If you provide only the zonal mean to an FFT algorithm, then the wave amplitude would be zero for all zonal wavenumbers except s=0.

Line 376: not “the data is bigger….”, rather “the amount of data is larger by a factor….”

Line 392: “slow to….”??? Also: a simple statistical connection between the upper-level flow and surface temperatures per so is not an “emergent constraint” (the way I understand that concept). Please clarify.

Hide

RR by Anonymous Referee #1 (31 Jan 2022)

ED: Publish subject to revisions (further review by editor and referees) (01 Feb 2022) by Irina Rudeva

Dear authors,

I appreciate the revisions made in response to the initial comments during the open discussion.

We have now received comments from two referees on the revised version. Both reviewers suggested that the manuscript might be considered for publication, however, needs major revisions.

Reviewer#1 asked for more accurate wording, particularly in the newly added parts the manuscript. I very much agree on this comment and would like to ask you to pay more attention to the wording.

Reviewer#2 made a critical comment on the method used to identify circumglobal wave events. I think the comment made by Rev#2 on existence of preferred phases for some wavenumbers is very important. They also raised a question on a disagreement between findings of the submitted manuscript, showing a little role of the soil moisture in simulations, and some previous studies.

Additionally, I made some comments that you will find below.

I encourage you to further revise your manuscript in response to these new comments. I am looking forward to a revised version.

Thank you for submitting your manuscript to WCD!

Best regards, 
Irina Rudeva

Comments from the editor:
Section 3.3: Discussion around Fig. 5 and 6 is very detailed. Since the numbers (correlations, n.s.d. and RMSE) can be seen in the plots and Tables, I suggest that you focus the discussion on the main message supporting the statement that prescribed soil moisture does not add much improvement for the anomalies in t2m and prcp. Please do not feel that you need to repeat all the information from the plots in the text.

Fig. 5 and 6: As the main purpose of these plots is to show an improvement made in AFSI experiments and little advantage in using SF, I suggest using color for different experiments and shapes for the three models. This should work well, especially for wave-7.

(lines in the version with tracked changes)
l. 221: ‘then times one hundred percent.’ I’d remove.
l.230-231: ‘The results imply that the model is able …’ - which model?
l.240: what’s the difference betweenFig. 3&4 and B10&B11?
l.461: V250 is not a surface variable
l.469:I thought improvements in AISF were marginal
l.490: remove ‘460’

Hide

AR by Fei Luo on behalf of the Authors (30 Mar 2022) Author's response Author's tracked changes Manuscript

EF by Polina Shvedko (31 Mar 2022)

ED: Publish subject to revisions (further review by editor and referees) (01 Apr 2022) by Irina Rudeva

ED: Referee Nomination & Report Request started (08 Apr 2022) by Irina Rudeva

RR by Anonymous Referee #3 (05 Jun 2022)

For final publication, the manuscript should be

accepted as is

accepted subject to technical corrections

accepted subject to minor revisions

reconsidered after major revisions

rejected

Were a revised manuscript to be sent for another round of reviews:

I would be willing to review the revised manuscript.

I would not be willing to review the revised manuscript.

Suggestions for revision or reasons for rejection

This is the revised version of a manuscript that I have reviewed before. My original concerns have been addressed to a large extent by the reply to my earlier comments as well as by the revisions.

The paper sort-of “validates” three state-of-the-art climate models in terms of the statistical behavior of selected episode during which a few selected Fourier components of the upper tropospheric meridional wind have strong amplitude. The motivation for those specific Fourier components stems from an earlier paper in which the authors have shown an interesting connection to summer heat extremes. The main point here is that the models represent these large-amplitude wave-5 and wave-7 events sort of OK in terms of the upper tropospheric circulation statistics, while there is a substantial underrepresentation (and partly misrepresentation) in terms of associated surface variables. The situation is analyzed in more detail with the help of model simulations in which certain model variables are nudged towards observed conditions. In particular, the underestimation close to the surface almost disappears when the upper tropospheric flow is nudged towards observations.

The analysis seems to be sound, the topic is relevant, and the results are intriguing. I have a few further issues which should be taken into account when preparing the final version of this manuscript.

Major issues

1.)The selection of episodes (both in reanalyses and in the free-running model simulations) is based on the large amplitude of a specific wavenumber of the upper tropospheric meridional wind. This procedure does not imply that the episodes in the reanalyses and in the model simulations represent the same “events”. The use of the term “event” is, therefore, somewhat misleading.

The situation changes fundamentally when the upper tropospheric flow in the model simulations is nudged to the reanalyses: now the different episodes, indeed, represent the same (or at least very similar) events. In particular, the associated surface anomalies now turn similar to the observed ones (from reanalyses) to the extent that the upper tropospheric flow determines the lower tropospheric meteorology, which is true to a considerable extent.

I think it would be helpful to make a clear distinction between “episodes” and “events”.

2.) Although the biases of the free-running EC model in terms of the upper-tropospheric flow are, in my eyes, quite substantial, they are referred to by the authors as “minor” or “small”. I think it would be fair to acknowledge that they are not just “minor” or “small”. This, in combination with the fact that the composite from the free-running model simulations is not composed of the same “events” as the composite from the reanalyses, renders the result somewhat less surprising: namely that nudging the upper tropospheric circulation improves the situation substantially.

Minor issues

Line 156, 209, 370: What do you exactly mean by the term “imprint”? Does it suggest a causal relationship? It seems that the latter is unproven at this point.

Line 605, caption of Fig 1: if the whiskers indicate the maximum and minimum values that have occurred, then all values are taken care of, right? In this case there should be no “outliers that are not shown in the plot”. Have I missed something?

Line 210: this sentence is somewhat non-sensical: even if the wave gets phase-locked in a not-preferred position, I would expect a prolongation of surface warm anomalies.

Line 212: Here you talk about “events”, suggesting that your method singles out specific meteorological situations. However, as argued above as well as in my previous review, this is not the case, rather your selection of “events” may contain a collection of rather different-looking meteorological situations. Therefore, the use of the term “event” is misleading. Rather, you single out specific “episodes”, not specific “events”.

Line 264 ff: The episodes that you study are selected using an upper-tropospheric criterion. For me, it therefore does not come as a surprise that the associated “bias” in the upper troposphere is less than in the lower troposphere. Can you comment, i.e., give the reader some guidance at this point?! Moreover, as stated above, I do not really find the biases in the upper troposphere to be “minor” or “small”.

Line 269: these two figures do not show biases, but composite anomalies. The biases can, at best, be inferred from these plots.

Line 285: unclear what the word “this” refers to.

Line 326: “individual Fouriermodes” would be better that “amplified Rossby waves”.

Line 333: again, you should not talk about “events” here (as detailed above). I you were more precise and talk only about “episodes”, the result (namely that surface variables do not necessarily show concomitant anomalies) would appear somewhat less surprising.

Similarly, the fact that these “biases” disappear when the (fairly substantial!) upper tropospheric biases are removed through nudging, turns less surprising, because you make a transition to the same “events” (as defined through upper tropospheric dynamics”).

Line 344: “… too weak in the models”: why?

Hide

ED: Publish subject to minor revisions (review by editor) (06 Jun 2022) by Irina Rudeva

AR by Fei Luo on behalf of the Authors (09 Jun 2022) Author's response Author's tracked changes Manuscript

ED: Publish subject to technical corrections (14 Jun 2022) by Irina Rudeva

AR by Fei Luo on behalf of the Authors (16 Jun 2022) Manuscript

Short summary

Recent studies have identified the weather systems in observational data, where wave patterns with high-magnitude values that circle around the whole globe in either wavenumber 5 or wavenumber 7 are responsible for the extreme events. In conclusion, we find that the climate models are able to reproduce the large-scale atmospheric circulation patterns as well as their associated surface variables such as temperature, precipitation, and sea level pressure.