the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Evaluation of Transport Processes over North China Plain and Yangtze River Delta using MAX-DOAS Observations
Abstract. The transport of pollutants has a substantial impact on the atmospheric environment in megacity clusters. However, owing to the lack of knowledge of the vertical pollutant structure, our quantification of transport processes and understanding of their impacts on the environment remain inadequate. In this study, we retrieved the vertical profiles of aerosol, NO2, and HCHO using multi-axis differential optical absorption spectroscopy (MAX-DOAS) and analyzed three typical transport phenomena. We found as follows: (1) The main transport layer (MTL) of aerosol, NO2 and HCHO along the southwest–northeast transport pathway in the Jing-Jin-Ji region were approximately 400–800 m, 0–400 m and 400–1400 m, respectively. The maximum transport flux of HCHO appeared in Wangdu (WD), oppositely, the minimum transport fluxes of aerosol and NO2 also occurred in this station. (2) The North China Plain (NCP) was usually affected by severe dust transport. The transported dust suppressed dissipation and boosted pollutant accumulation, converting the vertical profiles into an exponential shape. Furthermore, dust can indirectly affect trace gas concentrations by weakening optical intensity. For stations with higher optical intensity, the reduced NO2 levels were closely associated with its heterogeneous reactions on dust and aerosol surfaces. Comparatively, for other stations with low solar radiation, the decreased optical intensity favored NO2 concentration increase by inhibiting NO2 photolysis. The reduced solar radiation favored local HCHO accumulation in Shijiazhuang (SJZ) due to the dominant contribution of primary HCHO. (3) A back-and-forth transboundary transport between the NCP and Yangtze River Delta (YRD) was found. The YRD-to-NCP and NCP-to-YRD transport processes mainly occurred in the 500–1500 m and 0–1000 m layers, respectively. This transport, accompanied by the dome effect of aerosol, produced a large-scale PM2.5 concentration increase, further validating the haze-amplifying mechanism by practical observations.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(3363 KB)
-
Supplement
(9868 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3363 KB) - Metadata XML
-
Supplement
(9868 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-653', Anonymous Referee #1, 27 Sep 2022
This manuscript  described three typical transport phenomena in megacity clusters using MAX-DOAS network in China. The authors elaborately discussed transport processes and their possible effects. It is significant to for us to deeply understand the physical and chemical processes of air pollution events. However, due to my concerns with this manuscript that I outline below, I believe there should be some minor revisions done. If the authors can address my concerns, I do believe this work can be a positive contribution to the journal of ACP.
Â
General comments:
- Section 2.2, Line 139: the spectra measured with a solar zenith angle (SZA) of >75° to avoid the strong impact of stratospheric absorbers. Please elaborate the impact clearly.
- Section 2.3, Line 154: the clouds have large impacts on the data quality. Please describe this procedure and put it into the Supplementary materials.
- What is the estimated measurement uncertainty?
- Section 3.1, Line 264: ‘After 16:00, the high-extinction air mass shifted MTL from to 300–1000 m toward the surface at SJZ, with the AEC gradually exceeding 0.5 km-1 (Fig. 5)’. Which shift do you want to emphasize, the shift of MTL caused by the high-extinction air mass or the shift of air mass? Please reorganize the sentence.
- 9: Why is the data missing at NC and XH stations during March 6-22, 2021?
Â
Technical comments:
Line 177, ‘the wind speed in the southwest-northeast direction (WS)’ → ‘the wind speed (WS) in the southwest-northeast direction’
Line 191-192, ‘Due to the large discrepancy in their vertical distribution, the MTLs of various pollutants were bound to have different varying characteristics’ → ‘Due to the large discrepancy in the vertical distribution of various pollutants, their MTLs were bound to have different varying characteristics’
Line 220, ‘semibasin’ → ‘semi-basin’, ‘intraregional’ → ‘intra-regional’
Line 264, add space between ‘MTL’ and ‘from’, the logic of this sentence needs to be reconsidered.
Line 302: ‘According to the selection standards described in Supplement Sect. S3, we confirmed that March 15 was a dusty day’ and ‘dusty day’ is used in the following paragraphs. However, in Supplement Sect. S3, the date when the dust storm happened is defined as ‘dust day’. Please use the unified definition between the manuscript and the supplementary materials.
Line 360, ‘two stations assigned to the dark group (DG) located on the right’ → ‘two stations assigned to the dark group (DG) are located on the right.’
Line 411, ‘four periods: west-to-east, YRD to NCP, transformation, and NCP to YRD’ → ‘four periods: West-to-East, YRD-to-NCP, Transformation, and NCP-to-YRD’. To keep in accordance with captions in Fig 12.
Fig. 11: the date format of ‘yyyy-mm-dd’ is different from that of other figures. Please take the unified date format.
Fig. S11-15: ‘surface, 500 m, 800 m, 1000 m and 1500 m’ → ‘surface, 500, 800, 1000 and 1500 m’
Citation: https://doi.org/10.5194/egusphere-2022-653-RC1 -
AC1: 'Reply on RC1', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2022-653', Anonymous Referee #2, 29 Sep 2022
In the manuscript "Evaluation of Transport Processes over North China Plain and
Yangtze River Delta using MAX-DOAS Observations" the authors attempt to use profiles of NO2, HCHO and aerosols, retrieved from MAX-DOAS measurements at 8 stations distributed over northern China during a period of several months to show a number of different transport processes between the different stations covering different sub-regions of northern China. As additional data, the authors use wind speed and direction profiles from WRF model data, as well as surface concentration data from the CNEMCs air quality station network, as well as satellite data from TROPOMI and Himawari-8. The data only partly supports their various claims of transport processes. I advice the manuscript to be accepted only after major revisions are performed.
general comments:
I will address the authors directly in this comments and hence use the personal pronoun "you" (and, correspondingly the possessive pronoun "your") instead of writing "the authors".ÂI am still not convinced, as already mentioned in the pre-review, that the statements you make, are really supported by the data. In the "details" section below, I go through the different occasions where I either disagree or simply cannot see your statement confirmed by the data. You should have many more data to choose from and I wonder if you made the best choice of data to show in order to demonstrate your "transport phenomena". I suggest to screen your data again and to see if you find occasions with more continuous (less missing data) and possibly more consistent data. For me, the data seems to very fit to discriminate between secondary formation and transport. Only the data used to show your point (2) is convincing. The other two examples do not convince me.
I positively note that you improved on some occasions your references. As detailed in the the section below, I would like to motivate you to keep working on improving your references.
I find it hard to remember all the stations and which stations belong to which region. This, combined with the fact that the station abbreviations in Figure 1 are not all very well visible, I suggest that you make a table of all stations, color-coded by "region" in the main text (essentially, move Table 1 to main text and improve the visual appeal a bit). Regarding the station abbreviations: You use both NC and NB, if the choice of abbreviations is yours, I recommend to change this. However if those are fixed, nothing you can do. Similarly, if you could avoid the abbreviation CAMS for one of your stations, that would be something to consider. In general, it might be good practice to use a two-letter abbreviation for stations and a three letter acronym for regions. That way, the text immediately becomes a bit easier to read.
The division of figures presented in the main paper and in the supplements seems very arbitrary. In many of the descriptions of the data, you refer very frequently to plots at different locations in both documents. This makes it very cumbersome and time consuming to find the relevant information. I encourage you to overthink the distribution of information in the main article and the supplement, as well as the general choice of figures.
You frequently jump in the description of figures, most notably for Figures 3 and 4 which makes it difficult or almost impossible to follow. More generally, in the description of the data, you neither strictly follow a certain molecule, nor a certain station/ region nor an order in the figures. All this makes it hard to impossible to follow and very much obscure the points you want to make. I suggest to make a bullet point list (or maybe better a numbered list) of points you want to make and then describe one by one how this statement is supported by the data.
Additional confusion is introduced by random use of past and present tense. It is never clear whether a statement is made and refers to something that was shown in a previous section or paragraph, or whether the following paragraph will contain the affirmation of the statement, based on the data. Since I am not a native English speaker, I refrain from giving advice here and instead suggest to consult a native speaker about the best use of different tenses.
Except for paragraphs around line 291 and the paragraph following line 386, I do not see a lot of support for the statements made. Maybe because I simply cannot follow the argumentation, or maybe because the data in fact does not support the claims. In any case, this is not good and both the suggestions above [regarding the organization and order of arguments] as well as the comments following [more on the presentation of data and some lacking analysis] will help to improve this.
The quality and presenting choices of the figures should be improved. Especially the choice how to display wind fields is not well made. It is absolutely impossible to see the actual orientation, the arrow ends are not visible at all. Due to the size, the actual speed is also unclear. For the latter, I suggest to use an underlying semi-transparent color map layer. For the former, I suggest to use larger (and thicker) but more sparsely placed arrows. However, also other figures need improvement, e.g. choice of color scale or combination of colors and ordering of line plots, details see below.
Regarding the error analysis: You do now include a section on integrated column and surface error. I note this positively, thank you for taking up my critics of the pre-review. However, neither the quality of that, nor the extend are very satisfying. I cannot follow how you get to the percentage values you quote. Please include some equations you used to calculate those values.
Further, for the error analysis, you concentrate on integrated column and surface errors, however you mainly use profiles in your analysis. Hence, it is of uttermost importance to discuss the reliability of the profile shape. This is absent in your analysis. Since you also often argue that the data shows that the presence of aerosol triggers the formation of certain trace gases, it is important to discuss retrieval artifacts of aerosols leading to possibly incorrect ("too peaked") trace gas profiles, important in this context is also the frequent underestimation of dSCD errors and the effect on the trace gas profile (double peak, oscillations).It would also be good to include more information about typical degrees of freedom (for the valid data), to make a comment on the percentage of data filtered out by the RMS and DOF criteria and to show at least an example of an averaging kernel. Additionally, since this (trace gas and aerosol profiles) is in some occasions not your "final" quantity you use for the interpretation of the data, you should also include further error analysis (the contribution of the model error on the wind and what this means for the flux).
Somewhat related to this: it is also not clear how exactly you treat the different height grids from the retrieval and the model wind. I think it is best to include a sentence or a small paragraph on this. Further I am not convinced that fluxes should consider the mixing ratios, I think they should be calculated using concentrations (just as you actually state you would do), more on this below in the detailed comments.
Regarding molecule notation: HCHO: check journal guidelines whether you should use HCHO or H2CO and whether or not you have to introduce the chemical formula (i.e. writing "Formaldehyde (H2CO)...." at the first occurrence or not.
Apart from the strange use of time (which I decide not to comment on more than I already did), I do not have many comments regarding the use of language. The few I have are listed together with the detailed comments.
detailed comments:
l.27: "..oppositely...also occurred in this station" I do not follow here, how can something be "oppositely" and "also" at the same time?
l.32: "Comparatively" to what?
l.37: What are "practical observations"?
l. 46: "remarkably contributes" --> "contributes remarkably"
l.46: Maybe add some references?
l.46: I do not follow the phrase "transportation directly deteriorates the environment through the production and emission..."
l.51/52: What is the difference between cross-regional and inter-regional transport? And is intra-regional transport the same as regional transport?l.53: "local contributions was": plural or singular
l.55/56: "interact with the planetary boundary layer (PBL) and create an environment favorable for direct emission accumulation" what is meant by this?
l.59: "The movement of warm and humid air masses..." How does this fit to the rest of the paragraph?
l.60: "Hence, ..." what does the "hence" refer to? To the movement of the warm and humid air masses?
l.66: Add "To" before "characterize"
l.66: "monitoring" --> "monitored"
l.66: "ground level" --> "surface"?
l.70 add "to" between "used" and "investigate"
l.71: what is "technological support"?
l.74: Are those references the best fitting references here?
l.75/75: "The chemical transport model" --> "A chemical ..."
l.77: What are "hypothetical conditions"?
l.80: What are "technical methods"?
l.80: Maybe remove Wang from the reference list here or make clear why it is important to add Wang here
l.92: This has been used by many many many groups, please add more representative references
l.96: What do you mean by "hyperspectral stereoscopic"?
l.98: "technical support"?
l.101: "...impacts on and between regions"?
l.104f: I think this belongs to conclusions.
l.107: Is it not more the other way around: Since you mainly analyzed data from NCP and YRD, you concentrate on transport phenomena between those regions?
l.112: This is unclear: Is the BTH region also called JJJ or what?
l.113: What characterizes the continental monson climate?
l.113: "The regional transport of pollutants is prevalent within the JJJ region" seems to refer to a specific one, "The". Which?
Figure 1: Add a scale, enlarge the color bar, black on dark green/ blue is not well visible. Blue on dark green/blue is not well visible. magenta and read on such a "colored" plot are not very well distinguishable. I suggest to make the underlying map semi-transparent (the orography color scaling only, not the region contours) and to additionally use a different line style to indicate the regions.Â
l.124: This is a skyspec 1D? Please specify
end of page 6: There is a loose Table caption here.Â
Table 1: For the fitting interval for HCHO, are the two NO2 and O3 cross sections orthogonalized?Â
l. 139: "measured DSCDs") maybe retrieved since you do not measure them directly?
l.140: Please check your statement about the ring spectrum. It seems confusing.
l.142ff: regarding the choice of retrieval windows, do you base this on some reference? If so, please add.Â
l.147: Which fraction of data (approximately) does not pass your RMS criterion?
l.148: Can you specify "slowly"? How do you implement this statement "we only use data with slowly varying..."
l.152: I think "maximum" should not directly be used as adjective to "posteriori state vector" but possibly to something like likelihood or so.
l.154: add surface albedo and aerosol properties here in brackets or change i.e. to e.g.
l.155: You come back to how you construct Sa later. However, instead of having this (theoretical concept, practical implementation and construction) at two different places, I would put all of this together here. The same for the a priori profile. Additionally, you do not include information about the trace gas a priori profile (or I missed it). For the aerosol profile, it is unclear (l.164) whether the value you state for the surface, is really the surface or whether it is the value used in your lowest layer and hence at 50 m. This is likely not very different, but I think you should be specific. It might also make sense to actually include which AOD this corresponds to since often the integrated value is stated for the choice of a priori instead of the surface value.
l.157: "our first" what?
l.158: Which previous study? (move reference form l 160 to l-158) Also, if you use this here, maybe highlight the difference to the "usual way".Â
l.159: what is meant by "semi-quantify"?
l.163: add "the" between "For" and "aerosol".
l.163: add "an" between "selected" and "exponentially"
l.163: add  "profile" after a priori (or reformulate as : decreasing profile .... as a priori). Also everywhere: put a priori italic.
l.166: what do you use to convert asymmetry factor to phase function moments?Â
Set. 2.4: Please add the used formulae.Â
l.175: Smoothing error is related to AK and hence on Sa. How does this refer then directly to the DOAS fit?
l.177ff: You start arguing about local minima ("imperfect minima") but then continue talking about which elevation angles hold more info on which profile height (that should be characterized by your gain matrix, right?), however, I don't see the direct connection here. Please elaborate.
As a more common reference for O4, I know Wagner et al. 2009 who quote 10% accuracy which is substantially higher than the quoted 4%, however the quoted reference is also more recent (2013), maybe double check?
l.191: Please check the sentence, something goes wrong here
l.192: What do you mean by "temperature gap"?
l.194: I very strongly recommend, as already mentioned in the general comments, to do such investigations. Given the importance of the trace gas profile shape in this study, it is important to have a really good understanding of the effect of imperfect aerosol retrieval on the trace gas retrieval.
Eq.2: Or just sqrt(0.5)(va + ua)
Eq 3: The dimension of flux should be "quantity over area over time". If you use, as indicated by this equation, in fact the concentration (in e.g. "molec/cm3"), then the dimension is correct, since you get: Â molec/cm3 * m/s --> Â 100 molec/cm2/s. However, this seems not to be what you actually did, considering Fig. 3. Given that the same concentration in terms of molec/cm3 corresponds to a very different mixing ratio in terms of ppb at different heights, I would think that the former would be the better quantity to use (also, it is more in line with the classic definition of a flux as quantity per area per time).Â
Eq.4: In order not to introduce "per unit width" which is somewhat confusing, I would recommend to divide by sum/H_i, that way you keep the correct dimension of flux.
l.221: high --> highest
l.222: Eq. 4 is simply the definition, it does not demonstrate anything
Sect. 2.5: You need to include a discussion here about the model error on velocity and the effect on the total flux error. You do not put any error on your flux, c.f. comment in general comments.
Sect. 2.6: Please make clear how and for what you use this ancillary data. Would it not be better to use reanalysis data for the wind? Why did you perform your own simulations? What is the time resolution of the data from CNEMC?
l.228: add "See" before "Supplement"
l.229: add "of" between "details" and "the model"
l.228ff: This refers to horizontal?
l.231: What is Himawari-8?
l.237: A correlation of ~0.6 -- 0.7 is not really good. Can you relate the differences you find to the errors you quote in Sect. 2.4?Â
l.238: You do not just exclude "some" stations, you exclude exactly half of them. While I agree on your criterion, saying "some" is not ok if it's actually 50%. While 10 km seems like "arbitrary", your group of 8 stations has actually a clear division in terms of "closeness" to a monitoring station: 4 of them are closer than 5 km, the other 4 are further away than 15 km. Maybe there is a way to make it more clear that 10 km is actually a good choice, and choosing any number between 5 and 15 would not have changed anything.
l.242: influenced --> influences
Figure 2: Can you comment on the vast difference in valid data (factor 3)? Please make those correlation plots use equal aspect ratio. Consider using a different estimate such as Theil Sen, by "eye", your linear fit looks like a bad fit. Have you considered including also AOD comparison to aeronet stations? Do you get vastly different values considering the 4 stations separately? Regarding the in-situ data: I assume that those are available on a very fine time resolution (it is not described in sect. 2.6). It is not clear to me whether you use the closest in time or some time average. Please state. If you use the closest in time: Is this a good choice? How does this actually compare with the time resolution of a scan? Similarly: Do you consider a single station or do you maybe have several stations located in the line of sight from your corresponding instruments? How does the distance relate to the area you probe with your instrument?
l.249 ff: I wonder if this paragraph should not better go to the introduction.Â
l.252: "simulation" --> "simulations"
l.257: "According to the TROPOMI results" is a weird formulation. Please reformulate. Also, you could rever your statements (the ones connected with "whereas") and it would still be true. I do not see the difference here which you clearly want to highlight using "whereas".Â
Figure S2: This is double, and Figure S1 is missing in the supplement.Â
l.260: For me it looks much more W --> E than SW --> NE. That is the impression I get from Fig. S3.
Fig. S3: Please enhance the arrow thickness. This plot is not yet too bad. I can see the direction of arrows, but I cannot see the orientation. I do get an impression of the wind speed here, for later wind-field plots this is not the case. This one is still ok in this respect.
l.266: "also": Where did you additionally identify this?
l.267: You just refer to Fig. 3, and I do not see the previous statement in Fig. 3. It is maybe intended to say that you will further explain this in the following paragraph, but this is not clear.
l.268: How do you deal with the different layer distribution between retrieval grid and model layers?
l.266 -- 303: This description jumps all the time between Figures 3 and Figures 4. It is impossible to follow and I cannot see the statements made in this paragraph (it is also far too long) confirmed in the data, or at least, in the way the data is presented. As already suggested in my pre-review: try explaining your points to a not-involved co-worker and take notes of additional information/ plots that were needed to present in order to convince your co-worker that the statements you make are supported by your data.
l.271: The wind field you provide is from 13:30 and it shows W --> E at SJZ not SW --> NE.Â
l.272: high AEC over 300 m at WD around 11:00: I do not see this in the plot. Before 11, it seems higher at the surface. What is the distance here? Does it at all fit together with the wind speed and the time distance here?
l.273: I can maybe see an increase at 11, but at 12? Also again at 13 h but then to > 0.3 / km. Please check.
l.274: "at 200--800 m at CAMS at 12:15", for me it looks more as this is a raising aerosol layer than a transported one. How do you exclude this?
l.275: "After 16:00", is this really 4pm? For me it looks as 5pm? (for CAMS, but it isn't even clear with respect to which station you make this comment).Â
l.276 "from to 300 --1000 m"?
l.277: "similar": similar in which way?Â
l.280: Surely the fact that the emissions are from traffic cannot be seen in Fig.4?Â
l.281: "because of the heavy traffic flow in Beijing" how do you know this from Figure 3?
l.283: What is meant by "variation in high-concentration"?
l.283/284: How does it agree "well with the shift in the corresponding MTLSs (Fig.4)"? I don't even know what I should be searching for in Fig.4.
l.284: When and where do you show this, you haven't yet talked about HCHO or have you?
l.285: Can you actually exclude that this is a retrieval artifact?
l.285: But it is actually intensifying, should it not be diluting if it is transported?
l.287: It is unclear to me how the satellite can capture a transport and not simply a presence.
l.288 "much more likely" sounds like you made some statistical analysis here. Where is it? What is this statement based on?
l.289: What is "strong"? By which measures is it "strong"?
l.290: "Likewise" likewise what? "increased to" increased from what or from which value?
l.291ff: I would start the whole paragraph with this sentence here: "High HCHO concentrations tend to appear at higher altitudes than those of NO2 and aerosols. ...". This is one of the few things in the description which is clearly visible and which has a whole set of sound explanations following the statement.Â
l.297: Can you exclude a retrieval artifact here?
l.298: If it is at the same time, how can it "follow"?
l.299: If I am not mistaken, then you argue on line 274 that the high AEC at CAMS is due to transport? Now you attribute it to formation due to NO2?
l.300: Or you can use the argument from line 272 and argue it is transported from SJZ. How do you decide when you attribute it to transport and when you attribute it to secondary aerosols? This is not clear.
Figure 4: This relates to my comment to Eq.2: I think you should really use concentrations and not mixing ratios. Apart from really relating to the physical quantity (molecules instead of molecules fraction) it has the dimensions of a classical flux. I am not sure how I should treat a flux with dimensions m/s. It is not clear to me here how you treated missing data. Compared to Fig.3, there seem to be more data points "filled". E.g.: No2 at SJZ shows many gaps inn Fig. 3 but not in Fig.4.
l.313ff: Could this also be a bias introduced by having data at different times of the day? E.g.: CAMS has no NO2 data from 8 --1, a time where values for e.g. WD are especially low.
l.318ff: This should be moved to conclusions. However, since it is unclear to me how you discriminate btw transport, local primary production and rising, and secondary production, I do not see your statement supported.Â
l.323: which satellite? what are the satellite results that reveal things?
Fig. S5: I am very lost here. Can you please indicate the regions introduced in Fig.1? The legend is not readable. What is the source here? A description is missing. Also: why is the source of the dust important? In which way is this relevant for the rest of the analysis? What was the main point of doing this back trajectory analysis? This is not clear to me.
Figure 5: Why such an inconsistent choice of time? (minimize missing data I presume. But since you have many more months of data, don't you have better data than this?) Maybe use a different color in the label for the dusty day? Maybe use a separate x-axis scale for the different days (although I also do of course see the point of using the same and if you had used the same, I would maybe say use the same... always difficult. It is just that you comment on some details of aerosol profiles on 2021.03.06 that are hardly visible in the plot. Maybe you could add an insert?
l.334: I agree on the AEC part, but for HCHO, data is mainly missing except for DY?Â
l.334: "while" I do not see any discrepancy here with the statement before, so why "while"?
l.335/336: Fig. 5 does not show a classification, it just shows the profiles but no classification.Â
l.337: It's hard to see this because the curves are on top of each other: see comment Figure 5. Maybe make an insert?
l.339: Do you really consider these profiles very different? If you consider the AKs, aren't they rather almost the same profile? In fact, we cannot know, you don't show a single AK.
l.342: This was not shown. But maybe you mean that you are going to show this in this paragraph? Not clear...
l.345: "On dusty day", maybe on "the" dusty day?
l.350: what's the dof? How does it compare to the a priori?
l.352: really? Seems rather flat if not surface peaked? It seems as if the surface concentration doubled?
l.355f: I don't follow your argumentation.
l.358: I think this should refer to Fig. S8. at NC, there is hardly any valid data. Why not directly refer to Fig. 5= NO2 (the few valid data that is there) at the same times look rather similar to the same times at the 22nd, so the big difference is not between the dusty day and the clean days, but between the clean day and the dusty day together with the clean day after.
l.358: This is not true any longer if you take March 22 as the benchmark. In a later Sect. you make a more detailed comparison, comparing separately to both "benchmark" days. I'd rather keep the more detailed one in the later section. And completely skip this description here.
l.360: XH seems to actually show higher SC in the evening so I'm not sure I agree with your statement here.
l.361: what is meant by "optical variation"?
Figure 6: I would show this together with the histogram from the supplements. In the box plots, the division line you chose seems fairly arbitrary, especially because you would also need to explain why you consider the mean and not the median. I find the distinction more clear directly from the histograms.Â
l.368: It is not clear what the "optical signal intensity" is. The total integrated spectrum? Please specify.
Fig. S10: It is not clear to me at which days this is. All days?
l.372ff: I do not follow this definition (later it makes total sense, but the description is weird)
l.377: Where are the "proofs" of this statement?Â
Fig. S11: Are these means? medians? How do you take into account that the distribution of time during day (in terms of valid measurements) is different? Put labels: BG and DG (also, I think the text is already very acronym heavy, I would just use BRIGHT and DARK).
l.386: "with" --> "to"
l.386: I thought it was just argued that there was a considerable difference also for AEC?
l.386ff/ Fig.7: I do not see the added value of Fig.7 over Fig.S8. Figure S8 also clearly shows that there is a lack of data to compare at NC and that, using the same times than XH should show actually higher SC. (my previous argument: can you really compare the different stations if you compare at vastly different times of day?)
l.386ff: I like this paragraph (in fact both paragraphs) in general and I find it easier to follow than all the rest of the manuscript (with the exception of line 291ff which is also very sound). Given that large parts of HCHO are not at the surface, how good is it to use tracers from surface measurements? Of course I am aware that you have to use what you have. But maybe you can discuss a bit more the effect on e.g. the correlation, what part of the "unexplained" HCHO could be attributed to a location mismatch of tracers and HCHO? How does the fit translate to errors in the division between primary and secondary? I would maybe also build up the paragraph differently: Dust and aerosols can have different effects on the trace gas concentration: On the one hand, it limits the received radiation and hence prevents NO2 destruction. In the same direction (concentration increase) acts the effect of reducing turbulence and hence the diminishing of mixing. However, the aerosols and dust particles act as surface for heterogeneous NO2 destruction processes and this leads to a diminishing of NO2. The received total light intensity anti-correlates with the NO2 concentration and hence.Â
Supplement Sect. 4: This refers to Fig. 10 a which is the histogram? This should probably refer to Figure S13a? Please explain the color coding in the Figure caption of Figure S13a.Â
l.402: Why is this owning to the worsening meteorological conditions?Â
l.405: How did you "note" it?Â
l. 407: Maybe quickly say that you used CO as tracer for primary HCHO and Ox [I am not a chemist, but I have never seen this notation as Ox=O3+NO2, I have only seen NOx.] for photo chemical production (So you use O3?).
l.424: "in" --> "of"
l.424f: You mean this is what you are going to show in the following paragraph?
l.425: "overpassing" --> "exceeding"
l.425ff: according to map 1, only parts of NCP. I cannot see the southward moving very well. The east part of YRD is much more covered in the satellite images already at the start? For me it looks more as if the AE moves west, also it looks more "growing".Â
l.427: How does this south-to-north transport fit together with the southward mentioned in line 425?
l.426: I do not see this in the data
l.429: I just cannot see this in the plots, I cannot see wind directions at all or wind speeds in this plot. Please work with colors for the speeds and reduce the number of arrows but make them way thicker and longer (i.e. the 8 ms arrow should also grow of course).
l.436: "wind gathered towards"??
l.437: Do you base this comment for the whole region on just the four stations? Why don't you use the network of large number of GB stations? (In fact you, but later. I think I would present it earlier)
l.438: I can clearly see a region exceeding 120? "average" means time average? spatial average? Not clear.Â
Fig. S20: add date headers here.
l.439: How do you define and how do you identify a transport belt?
l.440: How do values of 6.1 fit to the color scale (which ends at 2)?
l.440: The value of 1.41 quoted seems to be rather the value in the lowest layer and not in the 100--700 m layer?
l.441: Where do you show this?Â
l.442: I do not see this.
l.446: But the surface AEC seem to increase tremendously?
l.446 vs. line 451: This seems to be a contradiction?
l. 449: XH decreased after the 19th? (c.f. FIg. 9)
l.461 : 4.42 in Fig. 9 with color scale ending at 2
l.466: The station in both regions should be HNU. But HNU drops earlier than DY, not later?
Fig.9: As mentioned above, maybe color scale maximum of 2 is not a good choice if you have values of >4 and > 6. If you fear losing detail at lower aerosol depths, maybe consider a logarithmic scale?
Sect. 4: I do not see much of the conclusions really supported by the data. And exception are lines 493 --lines 502 which I do see supported. The paragraph following that, (lines 503 -- 510) I cannot say whether I agree or not, because I cannot see anything in the provided wind field plots. I could give you the benefit of doubt here, but I would prefer to see this more clearly in the plots. I cannot agree with your summary (line 511 to 515) because I don't think that the data was clearly supporting your statements.Â
l.481-482: "attributed" is repeated
l.486: remove "the"
l.490ff: Is it maybe that clean and dust days are mixed here? You showed this for the dusty days, not for the clean days.
l.491: when is it maintaining a Gaussian shape?
Citation: https://doi.org/10.5194/egusphere-2022-653-RC2 -
AC2: 'Reply on RC2', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Chengzhi Xing, 13 Nov 2022
-
RC3: 'Comment on egusphere-2022-653', Anonymous Referee #3, 06 Oct 2022
General Comments
The manuscript entitled "Evaluation of Transport Processes over North China Plain and Yangtze River Delta using MAX-DOAS Observations" by Song et al investigates the transport patterns and vertical distributions of NO2, HCHO and aerosols using a number of instruments from a MAX-DOAS network in China. The temporal variation of the air mass composition has been investigated using modelled wind fields, which allow to identify air masses moving from the region of one instrument to the other. This is a very useful approach that allows to investigate the dynamical and chemical processing of individual air masses.Â
A main problem of the manuscript is that there are many occasions where no clear distinction has been made between (1) conclusions evidently inferred from the measurements, (2) findings from other studies that support the measurements and (3) hypotheses based on the observations. Many conclusions drawn from measurements are highly speculative, for example that the presence of NO2 and HCHO enhance the AEC, and that secondary aerosols are present (Section 3.1). Another example is the statement "We discovered that secondary aerosol generation always accompanied the regional transport process" (L302), for which there is neither direct evidence from the measurements nor any other study mentioned that would support this finding.
I suggest to remove the discussion on dust properties inferred from the measured intensity. MAX-DOAS instruments are usually not radiometrically calibrated. Even if the spectrometers are of the same type, the signals from different instruments cannot be directly compared to each other since they depend on many parameters, such as the gain of the amplifier, as well as on the adjustment of the telescope optics, the length of the fibre bundle, etc. I therefore suggest to remove the corresponding paragraphs (Fig. 6 and L368-374), which anyway do not provide much extra information compared to the retrieved extinction profiles.Â
Finally, the manuscript appears to require substantial revision regarding of the usage of the English language. I have mentioned only a few in the technical corrections below.
Â
Specific Comments
L46: How does the transport of pollutants lead to the production and emission of pollutants? Please explain.
L73: This sentence is not only too general, but also incorrect. Satellite remote sensing data is certainly extremely useful to monitor variations in the atmospheric composition (although with no or only limited vertical resolution in the troposphere).
L76: I guess the statement "Large uncertainties remain in pollutant distribution estimation" only refers to model simulations. Please clarify.
L77: What do you mean with "hypothetical conditions"?
L86: Describing DOAS as a "a cutting-edge and promising method" seems inappropriate. DOAS is a well established and well validated technique that has been applied for the measurement of atmospheric trace gases since decades.
L93: I think this statement is not correct, since LIDAR has a much better vertical resolution than MAX-DOAS (at least for aerosols).
L96: MAX-DOAS is a not a hyperspectral method since spectral information is only obtained from a single viewing direction at one time. It is also not clear why MAX-DOAS should be a stereoscopic technique.
Section 2.3: Please discuss the fit errors and detection limits for the retrieved species. The optical density of HCHO and HONO shown in figure S2 are very weak. Can these trace gases be detected reliably, and is the signal-to-noise ratio sufficient for a useful retrieval of the vertical distribution of HONO and HCHO?Â
L140: Here it is not clear what you mean with "we calculated the ring spectrum as the measured spectrum, considering the contribution of the stratosphere to the DSCDs".
Section 2.4: Given that the signal-to-noise ratio apparent in Figure S2 seems to be very low, I am surprised that the smoothing and noise error components of the HCHO profiles are similar to those of the NO2 profile. It is not clear what you mean with "algorighmic error". Is this error due to inaccurate model parameters b or due to general incapabilities of the forward model to realistically represent the underlying physics?
L208: Explain why the wind-speed in north-easterly direction (and not in any other direction) is of relevance here.
Figure 2: Converting the in situ NO2 from µ g/m3 to ppb would allow for a much better quantitative assessment of the agreement between both datasets.Â
Figure 3: Why are there so many missing profiles? Is this due to outages of the instrument or has the profile retrieval failed in these cases?
L270ff: I feel that the description of the temporal and vertical distribution of aerosols at the different stations is not representing the overall picture appropriately. For example, it is stated that there is a "subtle increase" in aerosols above NC around 12:00, but it is not mentioned that this is just the onset of the presence of a strong aerosol layer throughout the afternoon. The finding that a persistent and elevated aerosol layer is first present at SJZ, and later at NC and CAMS, is not explicitly discussed. Are the times at which the aerosol layer reaches the different locations in agreement with the transport times from station to station as estimated from the wind speed? This would give further evidence that long-range transport has indeed occurred. What could be the reason for the much lower AECs at WD than at the other stations?
L297: How exactly do NO2 and HCHO enhance the AEC?Â
L300ff: How do you know that secondary aerosols were generated? To my knowledge, this cannot be inferred from MAX-DOAS measurements. I cannot find any evidence for your statement that secondary aerosol generation is always accompanied the regional transport process. Does that come from model calculations or other measurements? If so, please explain in detail. Is NO2 really the main precursor for organic aerosols? Please cite relevant publications that support this statement.
L277: It appears from Fig. 3 that the decrease in MTL of aerosols at NC already occured at 14:00, not 16:00. I do not think that the decrease in the aerosol layer height is related to the formation of a nocturnal boundary layer, which is formed much later right before sunset, and is initially very shallow. Aerosols present aloft would reside in the residual layer above the nocturnal surface layer. It seems much more likely that the increase in aerosol layer width is instead caused by increased vertical mixing due to a heat up of the surface in the course of the day.
L285ff: According to Eq. 3, shouldn't the unit for trace gas flux be ppb·m·s-1, and for aerosol extinction km-1·m·s-1?
L323: What kind of satellite results are your referring to? This should be explained in the main text, but is not even clear from the caption of Fig. S5.
Fig. S6: It is not clear what is shown here. Are these trajectories at different times or at different heights? The trajectories should be colour-coded for different heights/times.
Technical Corrections
L50: "driven by the southwest wind" -> "driven by south-westerly winds"
L66: "pollutant concentrations monitoring" -> "pollutant concentrations monitored"
L67: "Characterize" -> "Characterizing"
L74: "The chemical transport model" -> "Chemical transport models"
L75: "pollutant distribution" -> "pollutant distributions"; "is" -> "are"
Section 2: The title "Method and methodology" is a tautology. Use either "Methods" or "Methodology"
L124: "We operated a commercial MAX-DOAS instrument" -> "We operated seven (?) commercial MAX-DOAS instruments"
L140: Please explain what SCDs are (integrated concentrations along the light path).
Section 2.6: According to the ACP guidelines, all variables should be named according to the IUPAC conventions, with all variables being named using only a single lower-case letter. For example, in Equation (2) the expression va would be by convention interpreted as v·a, which is not what you mean here. I would suggest to use u and v for the meridional and zonal wind, respectively, and to replace Fluxc with Fc, and WSi with wi.Â
L221: Do you mean "layer with highest transport"?Â
L223: "discrepancy" -> "differences"
L257: Two times "that".
L258: "continuously" -> "homogeneously"
Citation: https://doi.org/10.5194/egusphere-2022-653-RC3 -
AC3: 'Reply on RC3', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Chengzhi Xing, 13 Nov 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-653', Anonymous Referee #1, 27 Sep 2022
This manuscript  described three typical transport phenomena in megacity clusters using MAX-DOAS network in China. The authors elaborately discussed transport processes and their possible effects. It is significant to for us to deeply understand the physical and chemical processes of air pollution events. However, due to my concerns with this manuscript that I outline below, I believe there should be some minor revisions done. If the authors can address my concerns, I do believe this work can be a positive contribution to the journal of ACP.
Â
General comments:
- Section 2.2, Line 139: the spectra measured with a solar zenith angle (SZA) of >75° to avoid the strong impact of stratospheric absorbers. Please elaborate the impact clearly.
- Section 2.3, Line 154: the clouds have large impacts on the data quality. Please describe this procedure and put it into the Supplementary materials.
- What is the estimated measurement uncertainty?
- Section 3.1, Line 264: ‘After 16:00, the high-extinction air mass shifted MTL from to 300–1000 m toward the surface at SJZ, with the AEC gradually exceeding 0.5 km-1 (Fig. 5)’. Which shift do you want to emphasize, the shift of MTL caused by the high-extinction air mass or the shift of air mass? Please reorganize the sentence.
- 9: Why is the data missing at NC and XH stations during March 6-22, 2021?
Â
Technical comments:
Line 177, ‘the wind speed in the southwest-northeast direction (WS)’ → ‘the wind speed (WS) in the southwest-northeast direction’
Line 191-192, ‘Due to the large discrepancy in their vertical distribution, the MTLs of various pollutants were bound to have different varying characteristics’ → ‘Due to the large discrepancy in the vertical distribution of various pollutants, their MTLs were bound to have different varying characteristics’
Line 220, ‘semibasin’ → ‘semi-basin’, ‘intraregional’ → ‘intra-regional’
Line 264, add space between ‘MTL’ and ‘from’, the logic of this sentence needs to be reconsidered.
Line 302: ‘According to the selection standards described in Supplement Sect. S3, we confirmed that March 15 was a dusty day’ and ‘dusty day’ is used in the following paragraphs. However, in Supplement Sect. S3, the date when the dust storm happened is defined as ‘dust day’. Please use the unified definition between the manuscript and the supplementary materials.
Line 360, ‘two stations assigned to the dark group (DG) located on the right’ → ‘two stations assigned to the dark group (DG) are located on the right.’
Line 411, ‘four periods: west-to-east, YRD to NCP, transformation, and NCP to YRD’ → ‘four periods: West-to-East, YRD-to-NCP, Transformation, and NCP-to-YRD’. To keep in accordance with captions in Fig 12.
Fig. 11: the date format of ‘yyyy-mm-dd’ is different from that of other figures. Please take the unified date format.
Fig. S11-15: ‘surface, 500 m, 800 m, 1000 m and 1500 m’ → ‘surface, 500, 800, 1000 and 1500 m’
Citation: https://doi.org/10.5194/egusphere-2022-653-RC1 -
AC1: 'Reply on RC1', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2022-653', Anonymous Referee #2, 29 Sep 2022
In the manuscript "Evaluation of Transport Processes over North China Plain and
Yangtze River Delta using MAX-DOAS Observations" the authors attempt to use profiles of NO2, HCHO and aerosols, retrieved from MAX-DOAS measurements at 8 stations distributed over northern China during a period of several months to show a number of different transport processes between the different stations covering different sub-regions of northern China. As additional data, the authors use wind speed and direction profiles from WRF model data, as well as surface concentration data from the CNEMCs air quality station network, as well as satellite data from TROPOMI and Himawari-8. The data only partly supports their various claims of transport processes. I advice the manuscript to be accepted only after major revisions are performed.
general comments:
I will address the authors directly in this comments and hence use the personal pronoun "you" (and, correspondingly the possessive pronoun "your") instead of writing "the authors".ÂI am still not convinced, as already mentioned in the pre-review, that the statements you make, are really supported by the data. In the "details" section below, I go through the different occasions where I either disagree or simply cannot see your statement confirmed by the data. You should have many more data to choose from and I wonder if you made the best choice of data to show in order to demonstrate your "transport phenomena". I suggest to screen your data again and to see if you find occasions with more continuous (less missing data) and possibly more consistent data. For me, the data seems to very fit to discriminate between secondary formation and transport. Only the data used to show your point (2) is convincing. The other two examples do not convince me.
I positively note that you improved on some occasions your references. As detailed in the the section below, I would like to motivate you to keep working on improving your references.
I find it hard to remember all the stations and which stations belong to which region. This, combined with the fact that the station abbreviations in Figure 1 are not all very well visible, I suggest that you make a table of all stations, color-coded by "region" in the main text (essentially, move Table 1 to main text and improve the visual appeal a bit). Regarding the station abbreviations: You use both NC and NB, if the choice of abbreviations is yours, I recommend to change this. However if those are fixed, nothing you can do. Similarly, if you could avoid the abbreviation CAMS for one of your stations, that would be something to consider. In general, it might be good practice to use a two-letter abbreviation for stations and a three letter acronym for regions. That way, the text immediately becomes a bit easier to read.
The division of figures presented in the main paper and in the supplements seems very arbitrary. In many of the descriptions of the data, you refer very frequently to plots at different locations in both documents. This makes it very cumbersome and time consuming to find the relevant information. I encourage you to overthink the distribution of information in the main article and the supplement, as well as the general choice of figures.
You frequently jump in the description of figures, most notably for Figures 3 and 4 which makes it difficult or almost impossible to follow. More generally, in the description of the data, you neither strictly follow a certain molecule, nor a certain station/ region nor an order in the figures. All this makes it hard to impossible to follow and very much obscure the points you want to make. I suggest to make a bullet point list (or maybe better a numbered list) of points you want to make and then describe one by one how this statement is supported by the data.
Additional confusion is introduced by random use of past and present tense. It is never clear whether a statement is made and refers to something that was shown in a previous section or paragraph, or whether the following paragraph will contain the affirmation of the statement, based on the data. Since I am not a native English speaker, I refrain from giving advice here and instead suggest to consult a native speaker about the best use of different tenses.
Except for paragraphs around line 291 and the paragraph following line 386, I do not see a lot of support for the statements made. Maybe because I simply cannot follow the argumentation, or maybe because the data in fact does not support the claims. In any case, this is not good and both the suggestions above [regarding the organization and order of arguments] as well as the comments following [more on the presentation of data and some lacking analysis] will help to improve this.
The quality and presenting choices of the figures should be improved. Especially the choice how to display wind fields is not well made. It is absolutely impossible to see the actual orientation, the arrow ends are not visible at all. Due to the size, the actual speed is also unclear. For the latter, I suggest to use an underlying semi-transparent color map layer. For the former, I suggest to use larger (and thicker) but more sparsely placed arrows. However, also other figures need improvement, e.g. choice of color scale or combination of colors and ordering of line plots, details see below.
Regarding the error analysis: You do now include a section on integrated column and surface error. I note this positively, thank you for taking up my critics of the pre-review. However, neither the quality of that, nor the extend are very satisfying. I cannot follow how you get to the percentage values you quote. Please include some equations you used to calculate those values.
Further, for the error analysis, you concentrate on integrated column and surface errors, however you mainly use profiles in your analysis. Hence, it is of uttermost importance to discuss the reliability of the profile shape. This is absent in your analysis. Since you also often argue that the data shows that the presence of aerosol triggers the formation of certain trace gases, it is important to discuss retrieval artifacts of aerosols leading to possibly incorrect ("too peaked") trace gas profiles, important in this context is also the frequent underestimation of dSCD errors and the effect on the trace gas profile (double peak, oscillations).It would also be good to include more information about typical degrees of freedom (for the valid data), to make a comment on the percentage of data filtered out by the RMS and DOF criteria and to show at least an example of an averaging kernel. Additionally, since this (trace gas and aerosol profiles) is in some occasions not your "final" quantity you use for the interpretation of the data, you should also include further error analysis (the contribution of the model error on the wind and what this means for the flux).
Somewhat related to this: it is also not clear how exactly you treat the different height grids from the retrieval and the model wind. I think it is best to include a sentence or a small paragraph on this. Further I am not convinced that fluxes should consider the mixing ratios, I think they should be calculated using concentrations (just as you actually state you would do), more on this below in the detailed comments.
Regarding molecule notation: HCHO: check journal guidelines whether you should use HCHO or H2CO and whether or not you have to introduce the chemical formula (i.e. writing "Formaldehyde (H2CO)...." at the first occurrence or not.
Apart from the strange use of time (which I decide not to comment on more than I already did), I do not have many comments regarding the use of language. The few I have are listed together with the detailed comments.
detailed comments:
l.27: "..oppositely...also occurred in this station" I do not follow here, how can something be "oppositely" and "also" at the same time?
l.32: "Comparatively" to what?
l.37: What are "practical observations"?
l. 46: "remarkably contributes" --> "contributes remarkably"
l.46: Maybe add some references?
l.46: I do not follow the phrase "transportation directly deteriorates the environment through the production and emission..."
l.51/52: What is the difference between cross-regional and inter-regional transport? And is intra-regional transport the same as regional transport?l.53: "local contributions was": plural or singular
l.55/56: "interact with the planetary boundary layer (PBL) and create an environment favorable for direct emission accumulation" what is meant by this?
l.59: "The movement of warm and humid air masses..." How does this fit to the rest of the paragraph?
l.60: "Hence, ..." what does the "hence" refer to? To the movement of the warm and humid air masses?
l.66: Add "To" before "characterize"
l.66: "monitoring" --> "monitored"
l.66: "ground level" --> "surface"?
l.70 add "to" between "used" and "investigate"
l.71: what is "technological support"?
l.74: Are those references the best fitting references here?
l.75/75: "The chemical transport model" --> "A chemical ..."
l.77: What are "hypothetical conditions"?
l.80: What are "technical methods"?
l.80: Maybe remove Wang from the reference list here or make clear why it is important to add Wang here
l.92: This has been used by many many many groups, please add more representative references
l.96: What do you mean by "hyperspectral stereoscopic"?
l.98: "technical support"?
l.101: "...impacts on and between regions"?
l.104f: I think this belongs to conclusions.
l.107: Is it not more the other way around: Since you mainly analyzed data from NCP and YRD, you concentrate on transport phenomena between those regions?
l.112: This is unclear: Is the BTH region also called JJJ or what?
l.113: What characterizes the continental monson climate?
l.113: "The regional transport of pollutants is prevalent within the JJJ region" seems to refer to a specific one, "The". Which?
Figure 1: Add a scale, enlarge the color bar, black on dark green/ blue is not well visible. Blue on dark green/blue is not well visible. magenta and read on such a "colored" plot are not very well distinguishable. I suggest to make the underlying map semi-transparent (the orography color scaling only, not the region contours) and to additionally use a different line style to indicate the regions.Â
l.124: This is a skyspec 1D? Please specify
end of page 6: There is a loose Table caption here.Â
Table 1: For the fitting interval for HCHO, are the two NO2 and O3 cross sections orthogonalized?Â
l. 139: "measured DSCDs") maybe retrieved since you do not measure them directly?
l.140: Please check your statement about the ring spectrum. It seems confusing.
l.142ff: regarding the choice of retrieval windows, do you base this on some reference? If so, please add.Â
l.147: Which fraction of data (approximately) does not pass your RMS criterion?
l.148: Can you specify "slowly"? How do you implement this statement "we only use data with slowly varying..."
l.152: I think "maximum" should not directly be used as adjective to "posteriori state vector" but possibly to something like likelihood or so.
l.154: add surface albedo and aerosol properties here in brackets or change i.e. to e.g.
l.155: You come back to how you construct Sa later. However, instead of having this (theoretical concept, practical implementation and construction) at two different places, I would put all of this together here. The same for the a priori profile. Additionally, you do not include information about the trace gas a priori profile (or I missed it). For the aerosol profile, it is unclear (l.164) whether the value you state for the surface, is really the surface or whether it is the value used in your lowest layer and hence at 50 m. This is likely not very different, but I think you should be specific. It might also make sense to actually include which AOD this corresponds to since often the integrated value is stated for the choice of a priori instead of the surface value.
l.157: "our first" what?
l.158: Which previous study? (move reference form l 160 to l-158) Also, if you use this here, maybe highlight the difference to the "usual way".Â
l.159: what is meant by "semi-quantify"?
l.163: add "the" between "For" and "aerosol".
l.163: add "an" between "selected" and "exponentially"
l.163: add  "profile" after a priori (or reformulate as : decreasing profile .... as a priori). Also everywhere: put a priori italic.
l.166: what do you use to convert asymmetry factor to phase function moments?Â
Set. 2.4: Please add the used formulae.Â
l.175: Smoothing error is related to AK and hence on Sa. How does this refer then directly to the DOAS fit?
l.177ff: You start arguing about local minima ("imperfect minima") but then continue talking about which elevation angles hold more info on which profile height (that should be characterized by your gain matrix, right?), however, I don't see the direct connection here. Please elaborate.
As a more common reference for O4, I know Wagner et al. 2009 who quote 10% accuracy which is substantially higher than the quoted 4%, however the quoted reference is also more recent (2013), maybe double check?
l.191: Please check the sentence, something goes wrong here
l.192: What do you mean by "temperature gap"?
l.194: I very strongly recommend, as already mentioned in the general comments, to do such investigations. Given the importance of the trace gas profile shape in this study, it is important to have a really good understanding of the effect of imperfect aerosol retrieval on the trace gas retrieval.
Eq.2: Or just sqrt(0.5)(va + ua)
Eq 3: The dimension of flux should be "quantity over area over time". If you use, as indicated by this equation, in fact the concentration (in e.g. "molec/cm3"), then the dimension is correct, since you get: Â molec/cm3 * m/s --> Â 100 molec/cm2/s. However, this seems not to be what you actually did, considering Fig. 3. Given that the same concentration in terms of molec/cm3 corresponds to a very different mixing ratio in terms of ppb at different heights, I would think that the former would be the better quantity to use (also, it is more in line with the classic definition of a flux as quantity per area per time).Â
Eq.4: In order not to introduce "per unit width" which is somewhat confusing, I would recommend to divide by sum/H_i, that way you keep the correct dimension of flux.
l.221: high --> highest
l.222: Eq. 4 is simply the definition, it does not demonstrate anything
Sect. 2.5: You need to include a discussion here about the model error on velocity and the effect on the total flux error. You do not put any error on your flux, c.f. comment in general comments.
Sect. 2.6: Please make clear how and for what you use this ancillary data. Would it not be better to use reanalysis data for the wind? Why did you perform your own simulations? What is the time resolution of the data from CNEMC?
l.228: add "See" before "Supplement"
l.229: add "of" between "details" and "the model"
l.228ff: This refers to horizontal?
l.231: What is Himawari-8?
l.237: A correlation of ~0.6 -- 0.7 is not really good. Can you relate the differences you find to the errors you quote in Sect. 2.4?Â
l.238: You do not just exclude "some" stations, you exclude exactly half of them. While I agree on your criterion, saying "some" is not ok if it's actually 50%. While 10 km seems like "arbitrary", your group of 8 stations has actually a clear division in terms of "closeness" to a monitoring station: 4 of them are closer than 5 km, the other 4 are further away than 15 km. Maybe there is a way to make it more clear that 10 km is actually a good choice, and choosing any number between 5 and 15 would not have changed anything.
l.242: influenced --> influences
Figure 2: Can you comment on the vast difference in valid data (factor 3)? Please make those correlation plots use equal aspect ratio. Consider using a different estimate such as Theil Sen, by "eye", your linear fit looks like a bad fit. Have you considered including also AOD comparison to aeronet stations? Do you get vastly different values considering the 4 stations separately? Regarding the in-situ data: I assume that those are available on a very fine time resolution (it is not described in sect. 2.6). It is not clear to me whether you use the closest in time or some time average. Please state. If you use the closest in time: Is this a good choice? How does this actually compare with the time resolution of a scan? Similarly: Do you consider a single station or do you maybe have several stations located in the line of sight from your corresponding instruments? How does the distance relate to the area you probe with your instrument?
l.249 ff: I wonder if this paragraph should not better go to the introduction.Â
l.252: "simulation" --> "simulations"
l.257: "According to the TROPOMI results" is a weird formulation. Please reformulate. Also, you could rever your statements (the ones connected with "whereas") and it would still be true. I do not see the difference here which you clearly want to highlight using "whereas".Â
Figure S2: This is double, and Figure S1 is missing in the supplement.Â
l.260: For me it looks much more W --> E than SW --> NE. That is the impression I get from Fig. S3.
Fig. S3: Please enhance the arrow thickness. This plot is not yet too bad. I can see the direction of arrows, but I cannot see the orientation. I do get an impression of the wind speed here, for later wind-field plots this is not the case. This one is still ok in this respect.
l.266: "also": Where did you additionally identify this?
l.267: You just refer to Fig. 3, and I do not see the previous statement in Fig. 3. It is maybe intended to say that you will further explain this in the following paragraph, but this is not clear.
l.268: How do you deal with the different layer distribution between retrieval grid and model layers?
l.266 -- 303: This description jumps all the time between Figures 3 and Figures 4. It is impossible to follow and I cannot see the statements made in this paragraph (it is also far too long) confirmed in the data, or at least, in the way the data is presented. As already suggested in my pre-review: try explaining your points to a not-involved co-worker and take notes of additional information/ plots that were needed to present in order to convince your co-worker that the statements you make are supported by your data.
l.271: The wind field you provide is from 13:30 and it shows W --> E at SJZ not SW --> NE.Â
l.272: high AEC over 300 m at WD around 11:00: I do not see this in the plot. Before 11, it seems higher at the surface. What is the distance here? Does it at all fit together with the wind speed and the time distance here?
l.273: I can maybe see an increase at 11, but at 12? Also again at 13 h but then to > 0.3 / km. Please check.
l.274: "at 200--800 m at CAMS at 12:15", for me it looks more as this is a raising aerosol layer than a transported one. How do you exclude this?
l.275: "After 16:00", is this really 4pm? For me it looks as 5pm? (for CAMS, but it isn't even clear with respect to which station you make this comment).Â
l.276 "from to 300 --1000 m"?
l.277: "similar": similar in which way?Â
l.280: Surely the fact that the emissions are from traffic cannot be seen in Fig.4?Â
l.281: "because of the heavy traffic flow in Beijing" how do you know this from Figure 3?
l.283: What is meant by "variation in high-concentration"?
l.283/284: How does it agree "well with the shift in the corresponding MTLSs (Fig.4)"? I don't even know what I should be searching for in Fig.4.
l.284: When and where do you show this, you haven't yet talked about HCHO or have you?
l.285: Can you actually exclude that this is a retrieval artifact?
l.285: But it is actually intensifying, should it not be diluting if it is transported?
l.287: It is unclear to me how the satellite can capture a transport and not simply a presence.
l.288 "much more likely" sounds like you made some statistical analysis here. Where is it? What is this statement based on?
l.289: What is "strong"? By which measures is it "strong"?
l.290: "Likewise" likewise what? "increased to" increased from what or from which value?
l.291ff: I would start the whole paragraph with this sentence here: "High HCHO concentrations tend to appear at higher altitudes than those of NO2 and aerosols. ...". This is one of the few things in the description which is clearly visible and which has a whole set of sound explanations following the statement.Â
l.297: Can you exclude a retrieval artifact here?
l.298: If it is at the same time, how can it "follow"?
l.299: If I am not mistaken, then you argue on line 274 that the high AEC at CAMS is due to transport? Now you attribute it to formation due to NO2?
l.300: Or you can use the argument from line 272 and argue it is transported from SJZ. How do you decide when you attribute it to transport and when you attribute it to secondary aerosols? This is not clear.
Figure 4: This relates to my comment to Eq.2: I think you should really use concentrations and not mixing ratios. Apart from really relating to the physical quantity (molecules instead of molecules fraction) it has the dimensions of a classical flux. I am not sure how I should treat a flux with dimensions m/s. It is not clear to me here how you treated missing data. Compared to Fig.3, there seem to be more data points "filled". E.g.: No2 at SJZ shows many gaps inn Fig. 3 but not in Fig.4.
l.313ff: Could this also be a bias introduced by having data at different times of the day? E.g.: CAMS has no NO2 data from 8 --1, a time where values for e.g. WD are especially low.
l.318ff: This should be moved to conclusions. However, since it is unclear to me how you discriminate btw transport, local primary production and rising, and secondary production, I do not see your statement supported.Â
l.323: which satellite? what are the satellite results that reveal things?
Fig. S5: I am very lost here. Can you please indicate the regions introduced in Fig.1? The legend is not readable. What is the source here? A description is missing. Also: why is the source of the dust important? In which way is this relevant for the rest of the analysis? What was the main point of doing this back trajectory analysis? This is not clear to me.
Figure 5: Why such an inconsistent choice of time? (minimize missing data I presume. But since you have many more months of data, don't you have better data than this?) Maybe use a different color in the label for the dusty day? Maybe use a separate x-axis scale for the different days (although I also do of course see the point of using the same and if you had used the same, I would maybe say use the same... always difficult. It is just that you comment on some details of aerosol profiles on 2021.03.06 that are hardly visible in the plot. Maybe you could add an insert?
l.334: I agree on the AEC part, but for HCHO, data is mainly missing except for DY?Â
l.334: "while" I do not see any discrepancy here with the statement before, so why "while"?
l.335/336: Fig. 5 does not show a classification, it just shows the profiles but no classification.Â
l.337: It's hard to see this because the curves are on top of each other: see comment Figure 5. Maybe make an insert?
l.339: Do you really consider these profiles very different? If you consider the AKs, aren't they rather almost the same profile? In fact, we cannot know, you don't show a single AK.
l.342: This was not shown. But maybe you mean that you are going to show this in this paragraph? Not clear...
l.345: "On dusty day", maybe on "the" dusty day?
l.350: what's the dof? How does it compare to the a priori?
l.352: really? Seems rather flat if not surface peaked? It seems as if the surface concentration doubled?
l.355f: I don't follow your argumentation.
l.358: I think this should refer to Fig. S8. at NC, there is hardly any valid data. Why not directly refer to Fig. 5= NO2 (the few valid data that is there) at the same times look rather similar to the same times at the 22nd, so the big difference is not between the dusty day and the clean days, but between the clean day and the dusty day together with the clean day after.
l.358: This is not true any longer if you take March 22 as the benchmark. In a later Sect. you make a more detailed comparison, comparing separately to both "benchmark" days. I'd rather keep the more detailed one in the later section. And completely skip this description here.
l.360: XH seems to actually show higher SC in the evening so I'm not sure I agree with your statement here.
l.361: what is meant by "optical variation"?
Figure 6: I would show this together with the histogram from the supplements. In the box plots, the division line you chose seems fairly arbitrary, especially because you would also need to explain why you consider the mean and not the median. I find the distinction more clear directly from the histograms.Â
l.368: It is not clear what the "optical signal intensity" is. The total integrated spectrum? Please specify.
Fig. S10: It is not clear to me at which days this is. All days?
l.372ff: I do not follow this definition (later it makes total sense, but the description is weird)
l.377: Where are the "proofs" of this statement?Â
Fig. S11: Are these means? medians? How do you take into account that the distribution of time during day (in terms of valid measurements) is different? Put labels: BG and DG (also, I think the text is already very acronym heavy, I would just use BRIGHT and DARK).
l.386: "with" --> "to"
l.386: I thought it was just argued that there was a considerable difference also for AEC?
l.386ff/ Fig.7: I do not see the added value of Fig.7 over Fig.S8. Figure S8 also clearly shows that there is a lack of data to compare at NC and that, using the same times than XH should show actually higher SC. (my previous argument: can you really compare the different stations if you compare at vastly different times of day?)
l.386ff: I like this paragraph (in fact both paragraphs) in general and I find it easier to follow than all the rest of the manuscript (with the exception of line 291ff which is also very sound). Given that large parts of HCHO are not at the surface, how good is it to use tracers from surface measurements? Of course I am aware that you have to use what you have. But maybe you can discuss a bit more the effect on e.g. the correlation, what part of the "unexplained" HCHO could be attributed to a location mismatch of tracers and HCHO? How does the fit translate to errors in the division between primary and secondary? I would maybe also build up the paragraph differently: Dust and aerosols can have different effects on the trace gas concentration: On the one hand, it limits the received radiation and hence prevents NO2 destruction. In the same direction (concentration increase) acts the effect of reducing turbulence and hence the diminishing of mixing. However, the aerosols and dust particles act as surface for heterogeneous NO2 destruction processes and this leads to a diminishing of NO2. The received total light intensity anti-correlates with the NO2 concentration and hence.Â
Supplement Sect. 4: This refers to Fig. 10 a which is the histogram? This should probably refer to Figure S13a? Please explain the color coding in the Figure caption of Figure S13a.Â
l.402: Why is this owning to the worsening meteorological conditions?Â
l.405: How did you "note" it?Â
l. 407: Maybe quickly say that you used CO as tracer for primary HCHO and Ox [I am not a chemist, but I have never seen this notation as Ox=O3+NO2, I have only seen NOx.] for photo chemical production (So you use O3?).
l.424: "in" --> "of"
l.424f: You mean this is what you are going to show in the following paragraph?
l.425: "overpassing" --> "exceeding"
l.425ff: according to map 1, only parts of NCP. I cannot see the southward moving very well. The east part of YRD is much more covered in the satellite images already at the start? For me it looks more as if the AE moves west, also it looks more "growing".Â
l.427: How does this south-to-north transport fit together with the southward mentioned in line 425?
l.426: I do not see this in the data
l.429: I just cannot see this in the plots, I cannot see wind directions at all or wind speeds in this plot. Please work with colors for the speeds and reduce the number of arrows but make them way thicker and longer (i.e. the 8 ms arrow should also grow of course).
l.436: "wind gathered towards"??
l.437: Do you base this comment for the whole region on just the four stations? Why don't you use the network of large number of GB stations? (In fact you, but later. I think I would present it earlier)
l.438: I can clearly see a region exceeding 120? "average" means time average? spatial average? Not clear.Â
Fig. S20: add date headers here.
l.439: How do you define and how do you identify a transport belt?
l.440: How do values of 6.1 fit to the color scale (which ends at 2)?
l.440: The value of 1.41 quoted seems to be rather the value in the lowest layer and not in the 100--700 m layer?
l.441: Where do you show this?Â
l.442: I do not see this.
l.446: But the surface AEC seem to increase tremendously?
l.446 vs. line 451: This seems to be a contradiction?
l. 449: XH decreased after the 19th? (c.f. FIg. 9)
l.461 : 4.42 in Fig. 9 with color scale ending at 2
l.466: The station in both regions should be HNU. But HNU drops earlier than DY, not later?
Fig.9: As mentioned above, maybe color scale maximum of 2 is not a good choice if you have values of >4 and > 6. If you fear losing detail at lower aerosol depths, maybe consider a logarithmic scale?
Sect. 4: I do not see much of the conclusions really supported by the data. And exception are lines 493 --lines 502 which I do see supported. The paragraph following that, (lines 503 -- 510) I cannot say whether I agree or not, because I cannot see anything in the provided wind field plots. I could give you the benefit of doubt here, but I would prefer to see this more clearly in the plots. I cannot agree with your summary (line 511 to 515) because I don't think that the data was clearly supporting your statements.Â
l.481-482: "attributed" is repeated
l.486: remove "the"
l.490ff: Is it maybe that clean and dust days are mixed here? You showed this for the dusty days, not for the clean days.
l.491: when is it maintaining a Gaussian shape?
Citation: https://doi.org/10.5194/egusphere-2022-653-RC2 -
AC2: 'Reply on RC2', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Chengzhi Xing, 13 Nov 2022
-
RC3: 'Comment on egusphere-2022-653', Anonymous Referee #3, 06 Oct 2022
General Comments
The manuscript entitled "Evaluation of Transport Processes over North China Plain and Yangtze River Delta using MAX-DOAS Observations" by Song et al investigates the transport patterns and vertical distributions of NO2, HCHO and aerosols using a number of instruments from a MAX-DOAS network in China. The temporal variation of the air mass composition has been investigated using modelled wind fields, which allow to identify air masses moving from the region of one instrument to the other. This is a very useful approach that allows to investigate the dynamical and chemical processing of individual air masses.Â
A main problem of the manuscript is that there are many occasions where no clear distinction has been made between (1) conclusions evidently inferred from the measurements, (2) findings from other studies that support the measurements and (3) hypotheses based on the observations. Many conclusions drawn from measurements are highly speculative, for example that the presence of NO2 and HCHO enhance the AEC, and that secondary aerosols are present (Section 3.1). Another example is the statement "We discovered that secondary aerosol generation always accompanied the regional transport process" (L302), for which there is neither direct evidence from the measurements nor any other study mentioned that would support this finding.
I suggest to remove the discussion on dust properties inferred from the measured intensity. MAX-DOAS instruments are usually not radiometrically calibrated. Even if the spectrometers are of the same type, the signals from different instruments cannot be directly compared to each other since they depend on many parameters, such as the gain of the amplifier, as well as on the adjustment of the telescope optics, the length of the fibre bundle, etc. I therefore suggest to remove the corresponding paragraphs (Fig. 6 and L368-374), which anyway do not provide much extra information compared to the retrieved extinction profiles.Â
Finally, the manuscript appears to require substantial revision regarding of the usage of the English language. I have mentioned only a few in the technical corrections below.
Â
Specific Comments
L46: How does the transport of pollutants lead to the production and emission of pollutants? Please explain.
L73: This sentence is not only too general, but also incorrect. Satellite remote sensing data is certainly extremely useful to monitor variations in the atmospheric composition (although with no or only limited vertical resolution in the troposphere).
L76: I guess the statement "Large uncertainties remain in pollutant distribution estimation" only refers to model simulations. Please clarify.
L77: What do you mean with "hypothetical conditions"?
L86: Describing DOAS as a "a cutting-edge and promising method" seems inappropriate. DOAS is a well established and well validated technique that has been applied for the measurement of atmospheric trace gases since decades.
L93: I think this statement is not correct, since LIDAR has a much better vertical resolution than MAX-DOAS (at least for aerosols).
L96: MAX-DOAS is a not a hyperspectral method since spectral information is only obtained from a single viewing direction at one time. It is also not clear why MAX-DOAS should be a stereoscopic technique.
Section 2.3: Please discuss the fit errors and detection limits for the retrieved species. The optical density of HCHO and HONO shown in figure S2 are very weak. Can these trace gases be detected reliably, and is the signal-to-noise ratio sufficient for a useful retrieval of the vertical distribution of HONO and HCHO?Â
L140: Here it is not clear what you mean with "we calculated the ring spectrum as the measured spectrum, considering the contribution of the stratosphere to the DSCDs".
Section 2.4: Given that the signal-to-noise ratio apparent in Figure S2 seems to be very low, I am surprised that the smoothing and noise error components of the HCHO profiles are similar to those of the NO2 profile. It is not clear what you mean with "algorighmic error". Is this error due to inaccurate model parameters b or due to general incapabilities of the forward model to realistically represent the underlying physics?
L208: Explain why the wind-speed in north-easterly direction (and not in any other direction) is of relevance here.
Figure 2: Converting the in situ NO2 from µ g/m3 to ppb would allow for a much better quantitative assessment of the agreement between both datasets.Â
Figure 3: Why are there so many missing profiles? Is this due to outages of the instrument or has the profile retrieval failed in these cases?
L270ff: I feel that the description of the temporal and vertical distribution of aerosols at the different stations is not representing the overall picture appropriately. For example, it is stated that there is a "subtle increase" in aerosols above NC around 12:00, but it is not mentioned that this is just the onset of the presence of a strong aerosol layer throughout the afternoon. The finding that a persistent and elevated aerosol layer is first present at SJZ, and later at NC and CAMS, is not explicitly discussed. Are the times at which the aerosol layer reaches the different locations in agreement with the transport times from station to station as estimated from the wind speed? This would give further evidence that long-range transport has indeed occurred. What could be the reason for the much lower AECs at WD than at the other stations?
L297: How exactly do NO2 and HCHO enhance the AEC?Â
L300ff: How do you know that secondary aerosols were generated? To my knowledge, this cannot be inferred from MAX-DOAS measurements. I cannot find any evidence for your statement that secondary aerosol generation is always accompanied the regional transport process. Does that come from model calculations or other measurements? If so, please explain in detail. Is NO2 really the main precursor for organic aerosols? Please cite relevant publications that support this statement.
L277: It appears from Fig. 3 that the decrease in MTL of aerosols at NC already occured at 14:00, not 16:00. I do not think that the decrease in the aerosol layer height is related to the formation of a nocturnal boundary layer, which is formed much later right before sunset, and is initially very shallow. Aerosols present aloft would reside in the residual layer above the nocturnal surface layer. It seems much more likely that the increase in aerosol layer width is instead caused by increased vertical mixing due to a heat up of the surface in the course of the day.
L285ff: According to Eq. 3, shouldn't the unit for trace gas flux be ppb·m·s-1, and for aerosol extinction km-1·m·s-1?
L323: What kind of satellite results are your referring to? This should be explained in the main text, but is not even clear from the caption of Fig. S5.
Fig. S6: It is not clear what is shown here. Are these trajectories at different times or at different heights? The trajectories should be colour-coded for different heights/times.
Technical Corrections
L50: "driven by the southwest wind" -> "driven by south-westerly winds"
L66: "pollutant concentrations monitoring" -> "pollutant concentrations monitored"
L67: "Characterize" -> "Characterizing"
L74: "The chemical transport model" -> "Chemical transport models"
L75: "pollutant distribution" -> "pollutant distributions"; "is" -> "are"
Section 2: The title "Method and methodology" is a tautology. Use either "Methods" or "Methodology"
L124: "We operated a commercial MAX-DOAS instrument" -> "We operated seven (?) commercial MAX-DOAS instruments"
L140: Please explain what SCDs are (integrated concentrations along the light path).
Section 2.6: According to the ACP guidelines, all variables should be named according to the IUPAC conventions, with all variables being named using only a single lower-case letter. For example, in Equation (2) the expression va would be by convention interpreted as v·a, which is not what you mean here. I would suggest to use u and v for the meridional and zonal wind, respectively, and to replace Fluxc with Fc, and WSi with wi.Â
L221: Do you mean "layer with highest transport"?Â
L223: "discrepancy" -> "differences"
L257: Two times "that".
L258: "continuously" -> "homogeneously"
Citation: https://doi.org/10.5194/egusphere-2022-653-RC3 -
AC3: 'Reply on RC3', Chengzhi Xing, 13 Nov 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-653/egusphere-2022-653-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Chengzhi Xing, 13 Nov 2022
Peer review completion
Journal article(s) based on this preprint
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
345 | 106 | 16 | 467 | 35 | 1 | 3 |
- HTML: 345
- PDF: 106
- XML: 16
- Total: 467
- Supplement: 35
- BibTeX: 1
- EndNote: 3
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
Yuhang Song
Chengzhi Xing
Jinan Lin
Hongyu Wu
Ting Liu
Hua Lin
Chengxin Zhang
Wei Tan
Xiangguang Ji
Haoran Liu
Qihua Li
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3363 KB) - Metadata XML
-
Supplement
(9868 KB) - BibTeX
- EndNote
- Final revised paper