the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Dec 2022
Impact of the assimilation of water vapour isotopologues on diabatic heating and precipitation in the tropics
Abstract. The strong coupling between atmospheric circulation, moisture pathways and atmospheric diabatic heating is responsible for most climate feedback mechanisms and controls the evolution of severe weather events. However, diabatic heating rates obtained from current meteorological reanalysis show significant inconsistencies. Here, we theoretically assess with an Observation System Simulation Experiment (OSSE) the potential of the MUlti-platform remote Sensing of Isotopologues for investigating the Cycle of Atmospheric water (MUSICA) Infrared Atmospheric Sounding interferometer (IASI) mid-tropospheric water isotopologue data for constraining uncertainties in meteorological analysis fields. For this purpose, we use the Isotope-incorporated General Spectral Model (IsoGSM) together with a Local Ensemble Transform Kalman Filter (LETKF) and assimilate synthetic MUSICA IASI water vapour isotopologue observations. We perform four ensemble simulations, three where synthetic IASI isotopologue and humidity data are assimilated and another one where no observational data at all are assimilated. By comparing the ensemble simulations with data assimilation with water isotopologues to the one without any data assimilation we can, in contrast to the former study by Toride et al. (2021) where additionally conventional observations where considered, assess the direct impact of the IASI δD and water vapour data on the meteorological variables, especially on the heating rates and vertical velocity. The assessment is performed for the tropics in the latitude range from 10° S to 10° N. When the synthetic isotopologue data are assimilated, we derive reduced Root-Mean-Square Deviations (RMSDs) and improved skills with respect to meteorological variables (improvement by about 11–17%). When only IASI δD is assimilated the improvement in vertical velocity and heating rate is minor (up to a few percent) and restricted to the mid-troposphere. However, when water vapour alone or δD is assimilated additionally to water vapour, heating rates and vertical motion can be improved throughout the troposphere. The highest improvements, however, are derived when δD is assimilated additionally to water vapour confirming that water isotopologues hold different aspects of information than water vapour. Thus, in consequence δD offers, especially when assimilated additionally to water vapour, potential for improving diabatic heating and precipitation and thus meteorological analysis, weather forecasts and climate predictions in the tropical regions.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1375 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1408', Anonymous Referee #1, 19 Jan 2023
First review for authors
Impact of the assimilation of water vapour isotopologues on diabatic heating and precipitation in the tropics
Khosrawi et al. conduct a suite of experiments using the isotope-enabled IsoGSM climate model and assimilate synthetic observations of water isotopes (dD) in vapor to assess the impact on the reanalysis product produced by IsoGSM. They specifically focus on improvements in diabatic heating which houses large uncertainties in reanalysis products.
This paper examines the results of several data assimilation experiments to understand how assimilating dD changes the skill of the DA ensemble. The authors use DA ensemble means from 3 experiments and 1 control. The experiments are: a DA ensemble that assimilates only q (water vapor), one that assimilates only dD, and one that assimilates q and dD. Results are shown for the tropics, with focuses on the ascending and descending branches of the Walker circulation. Generally speaking, the DA ensemble mean with both q and dD performs better (more increased skill) than the DA ensemble with only q or with only dD. In a few cases, the DA ensemble with only dD results in a reduction in skill (a degradation), namely for the variables Q1, Q2, and w in the descending branches of the Walker circulation. The main result, that the highest skill improvement comes from the DA ensemble mean with both q and dD, suggests that “water isotopologues hold different aspects of information than water vapor”. The authors end by suggesting that DA-based meteorological analyses could be improved with the inclusion of dD in their forecasts, because water isotopes encode unique information about the dynamics and energetics of the troposphere that can’t be found in q alone.
This study makes use of water isotopes as a tracer for atmospheric processes, and applies this notion to a data-assimilation technique for meteorological forecasting. The authors build off previous work (Toride et al. 2021) that showed that using water vapor plus dD can improve the skill of a DA ensemble forecast, by specifically asking about the role of dD in improving diabatic heating rates and/or precipitation rates. This study seems to find generally agreeable results to Toride et al. 2021, i.e. dD plus q gives the best results, and the authors do a good job of outlining these agreements.
Unfortunately the presentation is low quality in terms of English grammar, style, organization, and structure, making it difficult to properly assess the manuscript. I edited the first few pages carefully for style and grammar but then stopped, as this really should not be the reviewers’ job. There are also several areas of the introduction in particular that are critically lacking citations. I recommend major revisions at a minimum, rejection with sufficient time to overhaul the manuscript at a maximum.
Major comments:
Introduction:
- This structure for the introduction does not clearly motivate what needs to be done and why, nor the gaps in the existing literature. The transitions between the first 4 paragraphs are quite choppy and do not clearly lay out the scientific need of the paper. Clearer topic sentences and transitions are sorely needed to better motivate the study.
- Transition between last two paragraphs in introduction is abrupt, and the introduction to the Walker circulation is out of place and confusing.
- The introduction should have a concise closing paragraph which describes the key questions and hypotheses that are investigated in the study, rather than a list of what each section says.
Novelty of work: I have serious concerns about the lack of novelty of this work. This could be partially due to poor presentation of the novelty of the study compared to the other studies cited throughout in the text. The authors need to do a better job clearly demonstrating what they are adding beyond the work of Toride, Tada, and the first author’s own manuscript in 2021. At present, this information is opaque/absent.
For example: 220-225, and throughout, these paragraphs make the reader wonder, what is this study adding? How do your results differ?
The authors state in their conclusions: “Our study shows that the assimilation of IASI δD and/or H2O data has the potential to improve meteorological analysis and thus also weather forecasts and climate predictions.” Again, I would argue that this has already been shown many times over.
To this point, I’m confused as to why the authors are using “mock” IASI data and not real satellite observations. Especially since this has already been done in Tada 2021? This study seems to be going backwards. The science has already progressed to assimilating real observations. A really strong and clear argument is needed as to why these oSSE experiments add information and/or help interpret the results in the real assimilation products. I can see how the water isotope retrieval uncertainties would need to be tested in a pseudo-framework such as the one presented here, but that is not discussed or laid out clearly in the text at all.
Reasoning: Several choices for the analysis seem random, like the assessment using the month of August. All such analysis choices should be clearly justified. I have marked these below in my specific comments.
Organization: Much of the text in the Discussion reads as Results, and the organization is choppy. What is the purpose of the Discussion as written?
A good Discussion/Conclusions section should have the following structure:
- What we did and its novelty/need in the field/ how it fills a critical gap
- What we found
- What are the uncertainties and caveats
- What is the future work required to move the field forward
- High-level conclusions (2 paragraphs) discussing how this work fits into the broader field of addressing the challenges posed by climate change in our warming future, and how these incremental improvements to our available tools will help us prepare/predict future outcomes.
Conclusions: the conclusions section is lacking. The authors need to discuss future avenues for how would we better elucidate the climate signals in this region? What additional data is needed and from where? If the models and the data don’t match, what do we do?
Heating Profiles: The work could benefit from a deeper dive into the variables Q1 and Q2, since it seems as though they are the variables of interest to this work, and these variables are a main way that this work is different from that of Toride et al. 2021. My understanding is that Q1 is the dry heat component of diabatic heating, while Q2 is the moist component? Equations 5 and 6 boil down to the material derivative of s (dry static heat) and L*q (latent heat of condensation). While this may be an appropriate method for quantifying diabatic heating, it could use some extra discussion beyond citing a paper from 1974. Have other authors done something similar more recently? Are there other ways we can think about diabatic heating? What do you intend for the reader to think of when they see Q1 and Q2 for the rest of the paper? I don’t think saying Q1 is “the apparent heat flux of the large scale motion system” is appropriate; Q1+Q2 is the total diabatic heating rate, you’ve just separated them into their dry and moist components. Also, I’m wondering if equation 5 is wrong (also mentioned this in the minor comments) – I *think* the third term on the right hand side should be w*ds/dp (as in the same term from equation 6, but replaced with s instead of q). This change would complete the notion that Q1 and Q2 are material derivatives of dry and moist static heat.
Skill Scores: Sometimes dD alone leads to a degradation in model skill, but dD+q is better than q alone (e.g. fig 2c, 2h, and 2i, and more). This is a very interesting finding! dD on it’s own sometimes degrades the forecast skill, but it also somehow manages to improve the skill of q alone (i.e. q+dD > q even when dD is degrading). I do not understand this, and I was hoping to see more of a discussion of the interaction between q and dD somewhere in the paper. Perhaps it is discussion material, but I was left really wanting to know your thoughts on why this is the case. Does Toride et al. 2021 have a suggestion? I think providing some thoughts on this would be a great addition to this work.
Control Run: It’s not clear to me, after reading the paper, what your control run is. The methods describe your Nature run, which is a climate model simulation using isoGCM. Then you describe a noDA experiment, which I think is your control (i.e. when you calculate skill, you use the noDA experiment as your control experiment, right?) I guess ultimately I am not sure what noDA is. You use the initial conditions from the Nature run to force your 96 ensembles for your DA experiments, right? But how do you create a DA experimental control (noDA) without assimilating any data? Is the noDA control just a DA ensemble assimilated with conventional climate data like T, u, and v? If so, “noDA” is probably not an appropriate name (maybe noWater instead?)
DA: A quick question about Kalman filters and DA algorithms in general – is it not a problem that q and dD covary quite closely in climate? Are you double-dipping on information when you use both q and dD ion an assimilation, or is it somehow worked out via the math of the Kalman filter? I know this is not a paper about Kalman filters, but providing some brief consideration about potential problems and why your method is acceptable would be good to see.
Grammar and style: Grammatical errors FREQUENTLY make the text hard to understand. There are some persistent and repeated grammatical mistakes, such as the use of “additionally to” when the correct phrase is “in addition to”. There are many, many areas where comma usage is incorrect and leads to an awkward reading experience (some examples listed in the minor comments, but not all.) This manuscript should not be resubmitted without a thorough rewrite for English grammar and style. The currently submitted version is unreadable.
Minor Comments:
Page 1 line 10: sentence starting with “By comparing” is long and convoluted and needs to be rewritten for clarity. The use of “WITH data assimilation WITH water isotopes” is repetitive and confusing. Please rephrase.
11: The interjection of “in contrast to …” comes out of no where is not appropriate materials for the abstract. Move this to the introduction or make it its own sentence.
15 improved “skills” in what? Please rephrase for clarity.
17 additionally = in addition to
18 can be = ARE improved
18 additionally – in addition to
19 “hold different aspects of information” is awkward and vague, rephrase.
20 this sentence structure is erratic and wrong, the comma statements are interrupting the main points.
There are basic English mistakes evident throughout the abstract, which I have corrected above, but will not correct for the full manuscript. Please carefully edit the text for English style and grammar before resubmission.
Introduction
25 – all these statements “e.g. representation of uncertainties” should be followed by a long list of references.
29 – “a high accuracy of these…for both…” what is the subject of this sentence?? What is “these”? These sentence structures are not well written.
29 comma splice
30 the phrase starting with “Diabatic heating” comes out of nowhere and should be a different paragraph.
31 citation?
38 contribution to what?
63 were = “where”
There are comma splices every other sentence.
The general organization of the writing is disjoint and hard to follow. In the introduction, for example, the structure appears to be (by paragraph)
- Weather forecasting has improved due to different advances
- Reanalysis products are crucial but diabatic heating is poorly resolved
- Water isotopes are sensitive hydrological cycle tracers (no transition from previous paragraphs)
- Water isotopes are also detectable in satellites (again, no connection to first two paragraphs)
- -6 are literature review on using water isotopes in data assimilation and weather forecasting…. But don’t draw us back to diabatic heating uncertainties…
- -9 describe the study focus and outcomes
This structure for the introduction does not clearly motivate what needs to be done and why, nor the gaps in the existing literature. The transitions between the first 4 paragraphs are quite choppy and do not clearly lay out the scientific need of the paper. Clearer topic sentences and transitions are sorely needed to better motivate the study.
73 “their latter issue,” no one will remember what that is…. Please be explicit
81 citations are needed…
Transition between last two paragraphs in introduction is abrupt, and the introduction to the Walker circulation is out of place and confusing.
The introduction should have a concise closing paragraph which describes the key questions and hypotheses that are investigated in the study, rather than a list of what each section says.
100 as A lower boundary condition.
130 again, misuse of additionally.
131 I’m not sure I understand what the ‘ensemble size’ means in this context. What are the details of the ensemble? How do the ensemble members differ?
165 what are the units of Rs?
176 why August?
177 “is prevailing” = prevails
179 (not shown???) what is the purpose of this statement? If you aren’t showing these results, don’t say this without proving it.
180 “This experiment” = What experiment? No idea what the authors are talking about, and this is a paragraph topic sentence.
185, 190, 204…. All of these paragraphs start with “Figure X shows….” And that is really not strong writing style. The paragraphs should NOT start with the figure description. The first sentence should be a topic sentence telling us what analysis is performed and why. Then, in the second sentence, or referenced at the end of the sentence, the figure should be presented.
195 In between (?) also what does this mean? Use scientific language, please.
200 what do you mean by “tighter?” Again, please use scientific language.
205 “This bar charts” = these bar charts…and thus summarize. Basic grammar mistakes continue throughout text.
Figure 1: labels are too small, increase all text labels. Some titles are cut off at the bottom.
220-225, and throughout, these paragraphs make the reader wonder, what is this study adding? How do your results differ?
Figure 3 all variables in the x-axis labels need to be defined in the figure caption.
249 awkward sentence “How large actually” – incorrect grammar, revise.
Figure 4 is not color-blind friendly, avoid Jet/rainbow colorbars, and labels on figure are far too small. Please increase size of all text on figures to appear as same size as printed text at minimum.
255 starting paragraph with figure reference. Please see comment above.
289 there is some error in the sentence structure here “dD has as water vapor….”
289 this has already been shown in many studies. What are you doing that is new?
293 what are the “conventional variables??” again this statement leads me to wonder how this study differs from Khosrawi 2021… and Toride….
Page 17 and the text around line 305 – this all reads as results, why is it in the discussion section?
330 these are results, not discussion.
339 this should also include a reference to Hu et al., JGR-A 2020, and several other papers. Again, this is not a novel result. The Role of IsotopeâEnabled GCM Complexity in Simulating Tropical Circulation Changes in HighâCO2 Scenarios
340 but you are also building on Khosrawi 2021, correct?
351 And generally, I’m confused as to why the authors are using “mock” IASI data and not real satellite observations. Especially since this has already been done in Tada 2021? This study seems to be going backwards. The science has already progressed to assimilating real observations. Thus a real strong argument is needed as to why these oSSE experiments add information and/or help interpret the results in the real assimilation products.
355 you are contradicting yourself in the same sentence here, saying that it can either slightly improve or degrade the results…. Doesn’t that cancel out any effect??? More specificity is needed here.
363 this has already been shown.
369 What uncertainties concerning modelling and observations? You cannot simply say “a lot” without providing a list of examples with citations. What uncertainties, specifically, would “hinder” the use of real IASI data?
Minor comments
Intro
Line 25: need an “and” before “representation”
Line 29: remove comma
Line 66: Sentence starting with “Comparing the three…” is confusingly worded. Suggestion: “Comparing the three assimilation experiments, they found that including conventional observations of water vapor, instead of dD alone, achieved higher improvements.”
The paragraph between lines 60 and 72 has some ideas that are stated multiple times. It could be cleaned up and made more precise.
Line 74: held should replace hold
Paragraph in lines 79-84: Could use a sentence about why the Walker Circulation is a useful area to focus on for this project.
General intro notes: Other DA projects that use water isotopes could be mentioned (e.g. the Last Millennium Reanalysis), but I don’t consider this crucial.
Data and Method
Line 95: remove comma
I still don’t really understand what MUSICA is – is it an instrument onboard the satellite? Is it software that processes raw satellite data?
So the initial conditions for the ensemble members of each data assimilation experiment comes from the first 96 6-hour time steps from the IsoGSM simulation?
Line 141: What do you mean by “experimental period”. You only evaluate the results of your DA runs for the month of August 2016, wouldn’t that make August 2016 your experimental period?
Lines 145-149: These equations don’t need to be defined – they are standard enough quantities.
Line 158: Same as above, you can just say skill or percent change. In my opinion those are generally understood quantities for readers of this material.
Line 167: The last sentence of this paragraph is either too long or it needs some commas or parentheses to separate the ideas in a more readable way. Suggestion: The apparent large-scale heat flux, Q1, is given by (insert formula for Q1), where s is the dry static energy. The apparent moisture sink, Q2 is given by (insert Q2 formula), where…. Etc.
Should the 3rd term on the right hand side of eq (5) be w*ds/dp, not s*ds/dp? As in the vertical velocity (w) times the vertical gradient in s (ds/dp)? That would make sense to me (i.e. the material derivative of s, similar to eq (6)), but I don’t understand the physical intuition behind this formula as it stands.
What is the difference between the Nature run and the noDA ensemble run? It is not clear to me how those are different.
Results
Line 191: The 500-600 hPa level of the satellite data should be mentioned before now, ideally in the methods section.
I need some more information on how to interpret Q1 and Q2. The standard climate variables (T, u, v, etc.) are all straightforward, but I’m not sure what Q1 and Q2 represent physically.
Figure 3: Make the lower limit of your y-axis closer to 0 so that there is not so much white space in this figure. (this applies to other similar figures as well)
Line 222: “…additional to conventional observations.” should be re-written to “…in addition to conventional observations.”
Line 225: See above comment
Line 222: I don’t understand the point of the sentence starting with “Thereby, we…”. Are you restating the previous sentence? If so, I don’t think you need this sentence. Is there something here about dD coinciding with regions of upward AND downward, whereas Q2 only coincides with upward? If so, make that more clear.
Discussion
Line 289: Not quite how I interpreted your results. You showed how assimilating dD plus h2o is good, but almost always dD on its own is not as good as h2o on its own. The key point being that the combination of dD AND water vapor leads to the best results.
Line 292: Sentence starting with “However…” is hard to read and has some general grammatical errors with commas and/or sentence structure.
Line 294: This paragraph needs a grammar review.
Conclusions
Line 340: help should replace hold
Line 347: “in addition to” should replace “additionally to”
Line 354: Use difference language to reduce confusion here: “…can be improved when assimilated with h2o alone, dD alone, or the combination of h2o plus dD…”
Citation: https://doi.org/10.5194/egusphere-2022-1408-RC1 -
RC2: 'Comment on egusphere-2022-1408', Anonymous Referee #2, 27 Mar 2023
Review of “Impact of the assimilation of water vapour isotopologues on diabatic heating and precipitation in the tropics” by Khosrawi et al. (egusphere-2022-1408)
The paper discusses the improvement of simulated meteorological fields through the assimilation of stable water vapor isotopes. The authors perform an Observation System Simulation Experiment (OSSE) to assess the effect of assimilating (i) only specific humidity (q), (ii) only δD, (iii) q and δD from a synthetic MUSICA IASI data set into the Isotope-incorporated Global Spectral Model (IsoGSM). They motivate their work through the inconsistency of diabatic heating rates in current meteorological reanalyses and assume that the assimilation of stable water isotopes leads to an improvement since stable water isotopes reflect the history of phase transitions. Their analysis is based on one month of data (August 2016) and focuses on the tropics (10° S to 10° N), specifically, the individual branches of the Walker circulation. They find that the assimilation of q and δD leads to the biggest improvements in the considered meteorological fields, followed by the assimilation of only q. The assimilation of only δD results in the smallest improvements and, in the case of three variables (ω, heat flux, and moisture sink), even a degradation (Fig. 1-3). They show that the assimilations affect the up- and downward branches of the Walker circulation differently (Fig. 5-7) and that for the vertical velocity and the moisture sink, the improvement/degradation depends on the vertical velocity itself (Fig. 9-11). Furthermore, they illustrate that the daily fluctuation of precipitation is only captured if q and/or δD are assimilated (Fig. 8) and that the improvement of precipitation is larger for regions with heavy precipitation compared to regions with little precipitation (Fig. 12).
I appreciate the authors' effort in creating a unique dataset of four assimilation runs with 96 members each. I also acknowledge the novel aspect of their work, which lies in restricting the assimilation to isotope variables. This approach allowed the authors to assess the effect of assimilating isotopes isolated from the effect of assimilating conventional variables (e.g., temperature, winds) as it was done, for instance, in Toride et al. (2021). However, the following two major issues need to be addressed thoroughly before the article is reconsidered for final publication. In my opinion, the two issues question whether the paper fits into the scope of WCD and does not instead belong in a journal that more specifically addresses the data assimilation community.
1) Lack of justification for the presented analyses
The current version of the paper resembles a technical report listing the results of an experiment without clearly stating the rationale behind it. As a reader, it is difficult to understand the sequence and relevance of the presented findings. The text should be written in such a way that it is immediately apparent why certain analysis steps were taken and what the gained insights are. To resolve this issue, the authors should explicitly answer the following questions in the paper:
1.1) Why do they perform an OSSE even though IASI observations are available? What is the benefit of this approach?
1.2) Why do they focus on August 2016, the latitudes 10° S to 10° N, and the branches of the Walker circulation? The authors should outline how they get from an inconsistency of diabatic heating rates in reanalyses (the initial motivation to perform the assimilation experiments) to a focus on individual branches of the Walker circulation during August.
1.3) It is clear from the introduction why the authors look at the heat source and the moisture sink during the evaluation of the assimilation experiment. However, the selection of the remaining variables, T, u, v, ω, q, δD, δ18O, and precipitation, remains unexplained. Are these variables selected because they are expected to be sensitive to the assimilation of isotopes? And, if so, what are the physical processes that link isotopes and these variables?
2) Insufficient interpretation and discussion of results
The presentation of the results is often merely descriptive and rarely addresses the underlying processes. In doing so, the authors miss an opportunity to advance our understanding of the processes that control isotopes. Furthermore, the physical consistency of the results remains untested. This shortcoming is particularly evident in the following cases:
2.1) What do the observed RMSDs (Fig. 1) mean in terms of processes? For instance: Do the different humidity profiles of the assimilation runs (Fig. 1e; especially the differences in the layer 800-400hPa) hint towards differences in the representation of the atmospheric circulation, the strength of convection, surface moisture fluxes, etc.?
2.2) I was surprised by the similar shapes of the profiles in Fig. 2. Since isotopes contain information about past phase changes (thus transport history), I would have expected that their assimilation has a spatially farther-reaching effect (i.e., affecting the altitudes above and below the altitude of assimilation) than the assimilation of q alone. Did the authors expect that the assimilation of δD and q would affect the same spatial region? Is the vertical resolution (28 levels) of the data sufficient to fully exploit the additional information contained in the isotopes?
2.3) How can the authors explain the degradation of ω, heat flux, and moisture sink through the assimilation of δD (Fig. 3, Fig. 5)? Isotopes are closely related to these variables (due to the strong vertical gradient in the spatial distribution of isotopes/the apparent involvement of phase changes). Therefore, I would expect an improvement of these variables through the assimilation of δD and not a degradation as observed in the paper.
Section 4 (entitled discussion, but in my opinion another results section since it contains new results) is to exclude from this comment. The analysis presented in this section is designed in such a way that it directly links to different processes, which the authors explicitly mention (L303-305). The results in Sect. 3 should be interpreted and discussed in the same way.
Minor/technical comments
- L76: isotoplogues (missing o)
- L80-88: inconsistency between Walker circulation vs. Walker Circulation
- L85: “In Sect. 3 we a assess the performance …” (missing comma after Sect. 3, and an “a” too much before assess)
- L88 + L288: “… in Sect. 4 we discuss …” Sect. 4 should be renamed, because the content belongs to the results and is not a discussion. Accordingly, the sentence on L88 should be adapted as well.
- L97-98: After describing the spatial resolution of your simulations, it would be helpful to learn about the temporal resolution. Please add this information.
- L123-145: I didn`t understand whether the nature run is one model with one member, or whether it also consists of 96 members.
- L135: “…from an free..” replace "an" with "a"
- L144: “… the mean difference between …” add (MD) after mean difference similar as you do for the root-mean-square deviation (RMSD) on L145
- L158: why do you introduce the abbreviation CTRL for the assimilation run which is otherwise referred to as noDA. I´d prefer if you stick to noDA also in the equation.
- L160: “(NWP, e.g., Bauer et al. (2015))” wrong format of citation, instead it should be (NWP, e.g., Bauer et al., 2015). This formatting mistake is repeated for other references.
- L186: “Figure 1 (left)” what does the (left) refer to?
- L323: “Figure 9 and Figure 10” should be Fig. 9 and Fig. 10 (since they are in the middle of a sentence)
- L309-311 + L332-333: Move/copy the bin definition into the caption of the respective figure (or even show the bin boundaries explicitly as thick labels). It is tedious to have to jump forth and back to understand the figure.
- L330 + L333: “Fig. 12” should be Figure 12 (since it is at the beginning of the sentence)
- Missing commas in several places
- Fig. 4: What do you intend to show with this figure? Elaborate in text. Also use the same abbreviations for the longitudinal boxes (Up1 Am, Up2 Af, …) as used in the remaining figures.
- Figures in general:
- Inconsistent use of abbreviations (e.g., iasi-δD-q vs. IASI δD+q vs. δD+q; use only one version throughout the whole document [figures, captions, texts])
- Inconsistent use of color code (Fig. 1, 2, 8 vs. Fig. 3)
- Incomplete units: often -10° to 10° or -10°-10°; add N for North
- Fig. 10 and 11: consider inverting x axis of vertical velocity to align up- and downward motion with bottom panels.
- Ensure sufficiently large font size (e.g., Fig. 4)
- Fig. 5-7: Up1, Up2, Up3, Up4 all have the same color; I assume that they are represented by the bars starting with Up1 to the left going to Up4 to the right? Explicitly write this in the figure caption (at least in Fig. 5) or chose different colors/fill patterns for the four bars. The same applies to Down1, Down2.
References
Toride, K., Yoshimura, K., Tada, M., Diekmann, C., Ertl., B., Khosrawi, F., and Schneider, M.: Potential of mid-tropospheric water vapor isotopes to improve large-scale circulation and weather predictability, Geophysical Research Letters, 48, e2020GL091 698, https://doi.org/10.1029/2020GL091698, 2021.
Citation: https://doi.org/10.5194/egusphere-2022-1408-RC2
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1408', Anonymous Referee #1, 19 Jan 2023
First review for authors
Impact of the assimilation of water vapour isotopologues on diabatic heating and precipitation in the tropics
Khosrawi et al. conduct a suite of experiments using the isotope-enabled IsoGSM climate model and assimilate synthetic observations of water isotopes (dD) in vapor to assess the impact on the reanalysis product produced by IsoGSM. They specifically focus on improvements in diabatic heating which houses large uncertainties in reanalysis products.
This paper examines the results of several data assimilation experiments to understand how assimilating dD changes the skill of the DA ensemble. The authors use DA ensemble means from 3 experiments and 1 control. The experiments are: a DA ensemble that assimilates only q (water vapor), one that assimilates only dD, and one that assimilates q and dD. Results are shown for the tropics, with focuses on the ascending and descending branches of the Walker circulation. Generally speaking, the DA ensemble mean with both q and dD performs better (more increased skill) than the DA ensemble with only q or with only dD. In a few cases, the DA ensemble with only dD results in a reduction in skill (a degradation), namely for the variables Q1, Q2, and w in the descending branches of the Walker circulation. The main result, that the highest skill improvement comes from the DA ensemble mean with both q and dD, suggests that “water isotopologues hold different aspects of information than water vapor”. The authors end by suggesting that DA-based meteorological analyses could be improved with the inclusion of dD in their forecasts, because water isotopes encode unique information about the dynamics and energetics of the troposphere that can’t be found in q alone.
This study makes use of water isotopes as a tracer for atmospheric processes, and applies this notion to a data-assimilation technique for meteorological forecasting. The authors build off previous work (Toride et al. 2021) that showed that using water vapor plus dD can improve the skill of a DA ensemble forecast, by specifically asking about the role of dD in improving diabatic heating rates and/or precipitation rates. This study seems to find generally agreeable results to Toride et al. 2021, i.e. dD plus q gives the best results, and the authors do a good job of outlining these agreements.
Unfortunately the presentation is low quality in terms of English grammar, style, organization, and structure, making it difficult to properly assess the manuscript. I edited the first few pages carefully for style and grammar but then stopped, as this really should not be the reviewers’ job. There are also several areas of the introduction in particular that are critically lacking citations. I recommend major revisions at a minimum, rejection with sufficient time to overhaul the manuscript at a maximum.
Major comments:
Introduction:
- This structure for the introduction does not clearly motivate what needs to be done and why, nor the gaps in the existing literature. The transitions between the first 4 paragraphs are quite choppy and do not clearly lay out the scientific need of the paper. Clearer topic sentences and transitions are sorely needed to better motivate the study.
- Transition between last two paragraphs in introduction is abrupt, and the introduction to the Walker circulation is out of place and confusing.
- The introduction should have a concise closing paragraph which describes the key questions and hypotheses that are investigated in the study, rather than a list of what each section says.
Novelty of work: I have serious concerns about the lack of novelty of this work. This could be partially due to poor presentation of the novelty of the study compared to the other studies cited throughout in the text. The authors need to do a better job clearly demonstrating what they are adding beyond the work of Toride, Tada, and the first author’s own manuscript in 2021. At present, this information is opaque/absent.
For example: 220-225, and throughout, these paragraphs make the reader wonder, what is this study adding? How do your results differ?
The authors state in their conclusions: “Our study shows that the assimilation of IASI δD and/or H2O data has the potential to improve meteorological analysis and thus also weather forecasts and climate predictions.” Again, I would argue that this has already been shown many times over.
To this point, I’m confused as to why the authors are using “mock” IASI data and not real satellite observations. Especially since this has already been done in Tada 2021? This study seems to be going backwards. The science has already progressed to assimilating real observations. A really strong and clear argument is needed as to why these oSSE experiments add information and/or help interpret the results in the real assimilation products. I can see how the water isotope retrieval uncertainties would need to be tested in a pseudo-framework such as the one presented here, but that is not discussed or laid out clearly in the text at all.
Reasoning: Several choices for the analysis seem random, like the assessment using the month of August. All such analysis choices should be clearly justified. I have marked these below in my specific comments.
Organization: Much of the text in the Discussion reads as Results, and the organization is choppy. What is the purpose of the Discussion as written?
A good Discussion/Conclusions section should have the following structure:
- What we did and its novelty/need in the field/ how it fills a critical gap
- What we found
- What are the uncertainties and caveats
- What is the future work required to move the field forward
- High-level conclusions (2 paragraphs) discussing how this work fits into the broader field of addressing the challenges posed by climate change in our warming future, and how these incremental improvements to our available tools will help us prepare/predict future outcomes.
Conclusions: the conclusions section is lacking. The authors need to discuss future avenues for how would we better elucidate the climate signals in this region? What additional data is needed and from where? If the models and the data don’t match, what do we do?
Heating Profiles: The work could benefit from a deeper dive into the variables Q1 and Q2, since it seems as though they are the variables of interest to this work, and these variables are a main way that this work is different from that of Toride et al. 2021. My understanding is that Q1 is the dry heat component of diabatic heating, while Q2 is the moist component? Equations 5 and 6 boil down to the material derivative of s (dry static heat) and L*q (latent heat of condensation). While this may be an appropriate method for quantifying diabatic heating, it could use some extra discussion beyond citing a paper from 1974. Have other authors done something similar more recently? Are there other ways we can think about diabatic heating? What do you intend for the reader to think of when they see Q1 and Q2 for the rest of the paper? I don’t think saying Q1 is “the apparent heat flux of the large scale motion system” is appropriate; Q1+Q2 is the total diabatic heating rate, you’ve just separated them into their dry and moist components. Also, I’m wondering if equation 5 is wrong (also mentioned this in the minor comments) – I *think* the third term on the right hand side should be w*ds/dp (as in the same term from equation 6, but replaced with s instead of q). This change would complete the notion that Q1 and Q2 are material derivatives of dry and moist static heat.
Skill Scores: Sometimes dD alone leads to a degradation in model skill, but dD+q is better than q alone (e.g. fig 2c, 2h, and 2i, and more). This is a very interesting finding! dD on it’s own sometimes degrades the forecast skill, but it also somehow manages to improve the skill of q alone (i.e. q+dD > q even when dD is degrading). I do not understand this, and I was hoping to see more of a discussion of the interaction between q and dD somewhere in the paper. Perhaps it is discussion material, but I was left really wanting to know your thoughts on why this is the case. Does Toride et al. 2021 have a suggestion? I think providing some thoughts on this would be a great addition to this work.
Control Run: It’s not clear to me, after reading the paper, what your control run is. The methods describe your Nature run, which is a climate model simulation using isoGCM. Then you describe a noDA experiment, which I think is your control (i.e. when you calculate skill, you use the noDA experiment as your control experiment, right?) I guess ultimately I am not sure what noDA is. You use the initial conditions from the Nature run to force your 96 ensembles for your DA experiments, right? But how do you create a DA experimental control (noDA) without assimilating any data? Is the noDA control just a DA ensemble assimilated with conventional climate data like T, u, and v? If so, “noDA” is probably not an appropriate name (maybe noWater instead?)
DA: A quick question about Kalman filters and DA algorithms in general – is it not a problem that q and dD covary quite closely in climate? Are you double-dipping on information when you use both q and dD ion an assimilation, or is it somehow worked out via the math of the Kalman filter? I know this is not a paper about Kalman filters, but providing some brief consideration about potential problems and why your method is acceptable would be good to see.
Grammar and style: Grammatical errors FREQUENTLY make the text hard to understand. There are some persistent and repeated grammatical mistakes, such as the use of “additionally to” when the correct phrase is “in addition to”. There are many, many areas where comma usage is incorrect and leads to an awkward reading experience (some examples listed in the minor comments, but not all.) This manuscript should not be resubmitted without a thorough rewrite for English grammar and style. The currently submitted version is unreadable.
Minor Comments:
Page 1 line 10: sentence starting with “By comparing” is long and convoluted and needs to be rewritten for clarity. The use of “WITH data assimilation WITH water isotopes” is repetitive and confusing. Please rephrase.
11: The interjection of “in contrast to …” comes out of no where is not appropriate materials for the abstract. Move this to the introduction or make it its own sentence.
15 improved “skills” in what? Please rephrase for clarity.
17 additionally = in addition to
18 can be = ARE improved
18 additionally – in addition to
19 “hold different aspects of information” is awkward and vague, rephrase.
20 this sentence structure is erratic and wrong, the comma statements are interrupting the main points.
There are basic English mistakes evident throughout the abstract, which I have corrected above, but will not correct for the full manuscript. Please carefully edit the text for English style and grammar before resubmission.
Introduction
25 – all these statements “e.g. representation of uncertainties” should be followed by a long list of references.
29 – “a high accuracy of these…for both…” what is the subject of this sentence?? What is “these”? These sentence structures are not well written.
29 comma splice
30 the phrase starting with “Diabatic heating” comes out of nowhere and should be a different paragraph.
31 citation?
38 contribution to what?
63 were = “where”
There are comma splices every other sentence.
The general organization of the writing is disjoint and hard to follow. In the introduction, for example, the structure appears to be (by paragraph)
- Weather forecasting has improved due to different advances
- Reanalysis products are crucial but diabatic heating is poorly resolved
- Water isotopes are sensitive hydrological cycle tracers (no transition from previous paragraphs)
- Water isotopes are also detectable in satellites (again, no connection to first two paragraphs)
- -6 are literature review on using water isotopes in data assimilation and weather forecasting…. But don’t draw us back to diabatic heating uncertainties…
- -9 describe the study focus and outcomes
This structure for the introduction does not clearly motivate what needs to be done and why, nor the gaps in the existing literature. The transitions between the first 4 paragraphs are quite choppy and do not clearly lay out the scientific need of the paper. Clearer topic sentences and transitions are sorely needed to better motivate the study.
73 “their latter issue,” no one will remember what that is…. Please be explicit
81 citations are needed…
Transition between last two paragraphs in introduction is abrupt, and the introduction to the Walker circulation is out of place and confusing.
The introduction should have a concise closing paragraph which describes the key questions and hypotheses that are investigated in the study, rather than a list of what each section says.
100 as A lower boundary condition.
130 again, misuse of additionally.
131 I’m not sure I understand what the ‘ensemble size’ means in this context. What are the details of the ensemble? How do the ensemble members differ?
165 what are the units of Rs?
176 why August?
177 “is prevailing” = prevails
179 (not shown???) what is the purpose of this statement? If you aren’t showing these results, don’t say this without proving it.
180 “This experiment” = What experiment? No idea what the authors are talking about, and this is a paragraph topic sentence.
185, 190, 204…. All of these paragraphs start with “Figure X shows….” And that is really not strong writing style. The paragraphs should NOT start with the figure description. The first sentence should be a topic sentence telling us what analysis is performed and why. Then, in the second sentence, or referenced at the end of the sentence, the figure should be presented.
195 In between (?) also what does this mean? Use scientific language, please.
200 what do you mean by “tighter?” Again, please use scientific language.
205 “This bar charts” = these bar charts…and thus summarize. Basic grammar mistakes continue throughout text.
Figure 1: labels are too small, increase all text labels. Some titles are cut off at the bottom.
220-225, and throughout, these paragraphs make the reader wonder, what is this study adding? How do your results differ?
Figure 3 all variables in the x-axis labels need to be defined in the figure caption.
249 awkward sentence “How large actually” – incorrect grammar, revise.
Figure 4 is not color-blind friendly, avoid Jet/rainbow colorbars, and labels on figure are far too small. Please increase size of all text on figures to appear as same size as printed text at minimum.
255 starting paragraph with figure reference. Please see comment above.
289 there is some error in the sentence structure here “dD has as water vapor….”
289 this has already been shown in many studies. What are you doing that is new?
293 what are the “conventional variables??” again this statement leads me to wonder how this study differs from Khosrawi 2021… and Toride….
Page 17 and the text around line 305 – this all reads as results, why is it in the discussion section?
330 these are results, not discussion.
339 this should also include a reference to Hu et al., JGR-A 2020, and several other papers. Again, this is not a novel result. The Role of IsotopeâEnabled GCM Complexity in Simulating Tropical Circulation Changes in HighâCO2 Scenarios
340 but you are also building on Khosrawi 2021, correct?
351 And generally, I’m confused as to why the authors are using “mock” IASI data and not real satellite observations. Especially since this has already been done in Tada 2021? This study seems to be going backwards. The science has already progressed to assimilating real observations. Thus a real strong argument is needed as to why these oSSE experiments add information and/or help interpret the results in the real assimilation products.
355 you are contradicting yourself in the same sentence here, saying that it can either slightly improve or degrade the results…. Doesn’t that cancel out any effect??? More specificity is needed here.
363 this has already been shown.
369 What uncertainties concerning modelling and observations? You cannot simply say “a lot” without providing a list of examples with citations. What uncertainties, specifically, would “hinder” the use of real IASI data?
Minor comments
Intro
Line 25: need an “and” before “representation”
Line 29: remove comma
Line 66: Sentence starting with “Comparing the three…” is confusingly worded. Suggestion: “Comparing the three assimilation experiments, they found that including conventional observations of water vapor, instead of dD alone, achieved higher improvements.”
The paragraph between lines 60 and 72 has some ideas that are stated multiple times. It could be cleaned up and made more precise.
Line 74: held should replace hold
Paragraph in lines 79-84: Could use a sentence about why the Walker Circulation is a useful area to focus on for this project.
General intro notes: Other DA projects that use water isotopes could be mentioned (e.g. the Last Millennium Reanalysis), but I don’t consider this crucial.
Data and Method
Line 95: remove comma
I still don’t really understand what MUSICA is – is it an instrument onboard the satellite? Is it software that processes raw satellite data?
So the initial conditions for the ensemble members of each data assimilation experiment comes from the first 96 6-hour time steps from the IsoGSM simulation?
Line 141: What do you mean by “experimental period”. You only evaluate the results of your DA runs for the month of August 2016, wouldn’t that make August 2016 your experimental period?
Lines 145-149: These equations don’t need to be defined – they are standard enough quantities.
Line 158: Same as above, you can just say skill or percent change. In my opinion those are generally understood quantities for readers of this material.
Line 167: The last sentence of this paragraph is either too long or it needs some commas or parentheses to separate the ideas in a more readable way. Suggestion: The apparent large-scale heat flux, Q1, is given by (insert formula for Q1), where s is the dry static energy. The apparent moisture sink, Q2 is given by (insert Q2 formula), where…. Etc.
Should the 3rd term on the right hand side of eq (5) be w*ds/dp, not s*ds/dp? As in the vertical velocity (w) times the vertical gradient in s (ds/dp)? That would make sense to me (i.e. the material derivative of s, similar to eq (6)), but I don’t understand the physical intuition behind this formula as it stands.
What is the difference between the Nature run and the noDA ensemble run? It is not clear to me how those are different.
Results
Line 191: The 500-600 hPa level of the satellite data should be mentioned before now, ideally in the methods section.
I need some more information on how to interpret Q1 and Q2. The standard climate variables (T, u, v, etc.) are all straightforward, but I’m not sure what Q1 and Q2 represent physically.
Figure 3: Make the lower limit of your y-axis closer to 0 so that there is not so much white space in this figure. (this applies to other similar figures as well)
Line 222: “…additional to conventional observations.” should be re-written to “…in addition to conventional observations.”
Line 225: See above comment
Line 222: I don’t understand the point of the sentence starting with “Thereby, we…”. Are you restating the previous sentence? If so, I don’t think you need this sentence. Is there something here about dD coinciding with regions of upward AND downward, whereas Q2 only coincides with upward? If so, make that more clear.
Discussion
Line 289: Not quite how I interpreted your results. You showed how assimilating dD plus h2o is good, but almost always dD on its own is not as good as h2o on its own. The key point being that the combination of dD AND water vapor leads to the best results.
Line 292: Sentence starting with “However…” is hard to read and has some general grammatical errors with commas and/or sentence structure.
Line 294: This paragraph needs a grammar review.
Conclusions
Line 340: help should replace hold
Line 347: “in addition to” should replace “additionally to”
Line 354: Use difference language to reduce confusion here: “…can be improved when assimilated with h2o alone, dD alone, or the combination of h2o plus dD…”
Citation: https://doi.org/10.5194/egusphere-2022-1408-RC1 -
RC2: 'Comment on egusphere-2022-1408', Anonymous Referee #2, 27 Mar 2023
Review of “Impact of the assimilation of water vapour isotopologues on diabatic heating and precipitation in the tropics” by Khosrawi et al. (egusphere-2022-1408)
The paper discusses the improvement of simulated meteorological fields through the assimilation of stable water vapor isotopes. The authors perform an Observation System Simulation Experiment (OSSE) to assess the effect of assimilating (i) only specific humidity (q), (ii) only δD, (iii) q and δD from a synthetic MUSICA IASI data set into the Isotope-incorporated Global Spectral Model (IsoGSM). They motivate their work through the inconsistency of diabatic heating rates in current meteorological reanalyses and assume that the assimilation of stable water isotopes leads to an improvement since stable water isotopes reflect the history of phase transitions. Their analysis is based on one month of data (August 2016) and focuses on the tropics (10° S to 10° N), specifically, the individual branches of the Walker circulation. They find that the assimilation of q and δD leads to the biggest improvements in the considered meteorological fields, followed by the assimilation of only q. The assimilation of only δD results in the smallest improvements and, in the case of three variables (ω, heat flux, and moisture sink), even a degradation (Fig. 1-3). They show that the assimilations affect the up- and downward branches of the Walker circulation differently (Fig. 5-7) and that for the vertical velocity and the moisture sink, the improvement/degradation depends on the vertical velocity itself (Fig. 9-11). Furthermore, they illustrate that the daily fluctuation of precipitation is only captured if q and/or δD are assimilated (Fig. 8) and that the improvement of precipitation is larger for regions with heavy precipitation compared to regions with little precipitation (Fig. 12).
I appreciate the authors' effort in creating a unique dataset of four assimilation runs with 96 members each. I also acknowledge the novel aspect of their work, which lies in restricting the assimilation to isotope variables. This approach allowed the authors to assess the effect of assimilating isotopes isolated from the effect of assimilating conventional variables (e.g., temperature, winds) as it was done, for instance, in Toride et al. (2021). However, the following two major issues need to be addressed thoroughly before the article is reconsidered for final publication. In my opinion, the two issues question whether the paper fits into the scope of WCD and does not instead belong in a journal that more specifically addresses the data assimilation community.
1) Lack of justification for the presented analyses
The current version of the paper resembles a technical report listing the results of an experiment without clearly stating the rationale behind it. As a reader, it is difficult to understand the sequence and relevance of the presented findings. The text should be written in such a way that it is immediately apparent why certain analysis steps were taken and what the gained insights are. To resolve this issue, the authors should explicitly answer the following questions in the paper:
1.1) Why do they perform an OSSE even though IASI observations are available? What is the benefit of this approach?
1.2) Why do they focus on August 2016, the latitudes 10° S to 10° N, and the branches of the Walker circulation? The authors should outline how they get from an inconsistency of diabatic heating rates in reanalyses (the initial motivation to perform the assimilation experiments) to a focus on individual branches of the Walker circulation during August.
1.3) It is clear from the introduction why the authors look at the heat source and the moisture sink during the evaluation of the assimilation experiment. However, the selection of the remaining variables, T, u, v, ω, q, δD, δ18O, and precipitation, remains unexplained. Are these variables selected because they are expected to be sensitive to the assimilation of isotopes? And, if so, what are the physical processes that link isotopes and these variables?
2) Insufficient interpretation and discussion of results
The presentation of the results is often merely descriptive and rarely addresses the underlying processes. In doing so, the authors miss an opportunity to advance our understanding of the processes that control isotopes. Furthermore, the physical consistency of the results remains untested. This shortcoming is particularly evident in the following cases:
2.1) What do the observed RMSDs (Fig. 1) mean in terms of processes? For instance: Do the different humidity profiles of the assimilation runs (Fig. 1e; especially the differences in the layer 800-400hPa) hint towards differences in the representation of the atmospheric circulation, the strength of convection, surface moisture fluxes, etc.?
2.2) I was surprised by the similar shapes of the profiles in Fig. 2. Since isotopes contain information about past phase changes (thus transport history), I would have expected that their assimilation has a spatially farther-reaching effect (i.e., affecting the altitudes above and below the altitude of assimilation) than the assimilation of q alone. Did the authors expect that the assimilation of δD and q would affect the same spatial region? Is the vertical resolution (28 levels) of the data sufficient to fully exploit the additional information contained in the isotopes?
2.3) How can the authors explain the degradation of ω, heat flux, and moisture sink through the assimilation of δD (Fig. 3, Fig. 5)? Isotopes are closely related to these variables (due to the strong vertical gradient in the spatial distribution of isotopes/the apparent involvement of phase changes). Therefore, I would expect an improvement of these variables through the assimilation of δD and not a degradation as observed in the paper.
Section 4 (entitled discussion, but in my opinion another results section since it contains new results) is to exclude from this comment. The analysis presented in this section is designed in such a way that it directly links to different processes, which the authors explicitly mention (L303-305). The results in Sect. 3 should be interpreted and discussed in the same way.
Minor/technical comments
- L76: isotoplogues (missing o)
- L80-88: inconsistency between Walker circulation vs. Walker Circulation
- L85: “In Sect. 3 we a assess the performance …” (missing comma after Sect. 3, and an “a” too much before assess)
- L88 + L288: “… in Sect. 4 we discuss …” Sect. 4 should be renamed, because the content belongs to the results and is not a discussion. Accordingly, the sentence on L88 should be adapted as well.
- L97-98: After describing the spatial resolution of your simulations, it would be helpful to learn about the temporal resolution. Please add this information.
- L123-145: I didn`t understand whether the nature run is one model with one member, or whether it also consists of 96 members.
- L135: “…from an free..” replace "an" with "a"
- L144: “… the mean difference between …” add (MD) after mean difference similar as you do for the root-mean-square deviation (RMSD) on L145
- L158: why do you introduce the abbreviation CTRL for the assimilation run which is otherwise referred to as noDA. I´d prefer if you stick to noDA also in the equation.
- L160: “(NWP, e.g., Bauer et al. (2015))” wrong format of citation, instead it should be (NWP, e.g., Bauer et al., 2015). This formatting mistake is repeated for other references.
- L186: “Figure 1 (left)” what does the (left) refer to?
- L323: “Figure 9 and Figure 10” should be Fig. 9 and Fig. 10 (since they are in the middle of a sentence)
- L309-311 + L332-333: Move/copy the bin definition into the caption of the respective figure (or even show the bin boundaries explicitly as thick labels). It is tedious to have to jump forth and back to understand the figure.
- L330 + L333: “Fig. 12” should be Figure 12 (since it is at the beginning of the sentence)
- Missing commas in several places
- Fig. 4: What do you intend to show with this figure? Elaborate in text. Also use the same abbreviations for the longitudinal boxes (Up1 Am, Up2 Af, …) as used in the remaining figures.
- Figures in general:
- Inconsistent use of abbreviations (e.g., iasi-δD-q vs. IASI δD+q vs. δD+q; use only one version throughout the whole document [figures, captions, texts])
- Inconsistent use of color code (Fig. 1, 2, 8 vs. Fig. 3)
- Incomplete units: often -10° to 10° or -10°-10°; add N for North
- Fig. 10 and 11: consider inverting x axis of vertical velocity to align up- and downward motion with bottom panels.
- Ensure sufficiently large font size (e.g., Fig. 4)
- Fig. 5-7: Up1, Up2, Up3, Up4 all have the same color; I assume that they are represented by the bars starting with Up1 to the left going to Up4 to the right? Explicitly write this in the figure caption (at least in Fig. 5) or chose different colors/fill patterns for the four bars. The same applies to Down1, Down2.
References
Toride, K., Yoshimura, K., Tada, M., Diekmann, C., Ertl., B., Khosrawi, F., and Schneider, M.: Potential of mid-tropospheric water vapor isotopes to improve large-scale circulation and weather predictability, Geophysical Research Letters, 48, e2020GL091 698, https://doi.org/10.1029/2020GL091698, 2021.
Citation: https://doi.org/10.5194/egusphere-2022-1408-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 430 | 132 | 32 | 594 | 24 | 39 |
- HTML: 430
- PDF: 132
- XML: 32
- Total: 594
- BibTeX: 24
- EndNote: 39
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1