To new heights by flying low: comparison of aircraft vertical NO<sub>2</sub> profiles to model simulations and implications for TROPOMI NO<sub>2</sub> retrievals

Riess, Tobias Christoph Valentin Werner; Boersma, Klaas Folkert; Van Roy, Ward; de Laat, Jos; Dammers, Enrico; van Vliet, Jasper

doi:10.5194/amt-16-5287-2023

Articles | Volume 16, issue 21

https://doi.org/10.5194/amt-16-5287-2023

© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/amt-16-5287-2023

© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 21

Research article

|

08 Nov 2023

Research article |

| 08 Nov 2023

To new heights by flying low: comparison of aircraft vertical NO₂ profiles to model simulations and implications for TROPOMI NO₂ retrievals

Tobias Christoph Valentin Werner Riess, Klaas Folkert Boersma, Ward Van Roy, Jos de Laat, Enrico Dammers, and Jasper van Vliet

Download

Final revised paper (published on 08 Nov 2023)
Supplement to the final revised paper
Preprint (discussion started on 05 Jun 2023)
Supplement to the preprint

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2023-1059', Anonymous Referee #1, 26 Jun 2023

Review of the manuscript: “To new heights by flying low: Comparison of aircraft vertical NO2 profiles to model simulations and implications for TROPOMI NO2 retrievals” by Reiss et al.
The manuscript deals with the assessment of the accuracy of the TROPOMI NO2 retrievals over sea, which is particularly relevant for monitoring ship emissions using satellite observations. The paper is well written and clear. I recommend the publication after addressing the following minor comments:
L47 “(Kurchaba et al. (2021); Finch et al. (2022))” should be “(Kurchaba et al., 2021; Finch et al., 2022)”, I think.
L51 therefor -> therefore (unless you really mean “for that”, better check this with English mother tongue, I am not sure)
L64 Again, I think this should be: (e.g., Visser et al., 2019; Douros et al., 2023). The same happens again later, perhaps the recommended journal format should be checked.
L102 “on the solar zenith angle, the satellite viewing zenith angle, on the scattering properties of the atmosphere and the surface, and on…” maybe here is missing “on” -> on the satellite viewing zenith angle (or a least all the others have it, so perhaps it should consistent)
L254-2555 The observed NO2 layer at 300 m is also present in the co-sampled LOTOS-EUROS profile but not in CAMS -> do you mean the peak at 500 m in LE?
L259 where TROPOMI has a higher sensitivity -> would it be good to say here that you will show the averaging kernels the describe the instrument sensitivity later in the manuscript? Something like: “(see Sect. 4)
L272 LOTOES-EUROS -> please correct the name
L293 It looks to me that there is confusion between 1 (the number one) and l (the letter L in minuscule). As I read it, it looks like you use level one (1) of the measured profile while it should be l (the letter). Or am I missing something? Anyways, please use italics for the letters as in the equation or clarify this. Also note that the line numbering got messed up due to the formula, I think (latex does this I guess).
Sect. 4.1 for this recalculation of the AMF you need to have the vertical profiles sampled at the same levels than the AKs, right? How did you do that: interpolating? Please mention that.
L297 as indicated by the averaging kernel. -> Here again you mention the averaging kernel but the figure appears only later: could you mention here that you will discuss this aspect later?
L310 It should be noted that CAMS shows systematically higher NO2 columns compared to measurements and TM5. - > What you mean here? The measured 3.37 is larger than CAMS value 3.03, maybe I misunderstood something. Please clarify.
Table 4 Could you add the percentage difference to the aircraft measurements as well as new column in the table?
L322 Which area exactly cover the land and sea mean AK?
L344 a periori -> Please correct.
L364 he -> the
L382 there is a parenthesis missing

Citation: https://doi.org/10.5194/egusphere-2023-1059-RC1
- AC1: 'Reply on RC1', Christoph Rieß, 21 Jul 2023
  
  Please see our response to the comments of referee 1 in the attachment.
  
  Citation: https://doi.org/10.5194/egusphere-2023-1059-AC1
RC2:
'Comment on egusphere-2023-1059', Anonymous Referee #2, 10 Jul 2023

Review of “To new heights by flying low: Comparison of aircraft vertical NO2 profiles to model simulations and implications for TROPOMI NO2 retrievals”:

This paper by Riess et al. compares aircraft-based vertical NO2 profiles to model simulations from TM5, CAMS-Europe, and LOTOS-EUROS. It investigates the influence of the a priori vertical NO2 profile on the TROPOMI tropospheric NO2 column retrieval. The comparison between actual aircraft measurements of NO2 vertical profiles and model simulations can provide valuable insights into the accuracy and reliability of model outputs. The analysis focuses on the polluted North Sea and the influence of shipping emissions. It is known that the a priori vertical NO2 profile is a dominating factor in the tropospheric NO2 column retrieval and that good knowledge of the vertical distribution is necessary. Since the TROPOMI NO2 product has started to be extensively used for the analysis of NO2 shipping signals, shipping emissions, and even single ships, analyzing the influence of the NO2 profile, especially over the North Sea, can provide valuable insights and improvements on the NO2 columns product. The study contributes to the field of atmospheric measurements by analyzing the vertical distribution of NO2 and its agreement with models and satellite retrievals. It is improving our understanding of atmospheric composition and the validation of satellite measurements. The results can contribute to improving the TROPOMI NO2 retrieval as well as model simulations. Therefore, I recommend publication in AMT after addressing the following comments.
General comments:

The study is based on a relatively small dataset. Even if ten profiles are included in the analyses, these measurements have been performed on only four days, from which 3 out of 4 days are dominated by land outflow. When results are discussed, this should be mentioned, and representativeness should be included more in the discussions.
The low NO2 concentrations in TM5 close to the surface compared to the measurements are mainly driven by a few profiles (3 & 6), which are too low; most profiles show an excellent agreement in the surface level. CAMS and LOTOS-EUROS, on the other hand, show similar low surface values for the same profiles as TM5, but since they show too high values on the other profiles at the surface level compared to the measurements and TM5, the mean looks better.

How would Fig. 1 (right panel) look like if the data are separated into land-outflow and clean cases, maybe this would be a useful additional plot.

Regarding this comment see also, for example, Page 18, lines 362-368, and further comments in the specific comments section.
The used heights for the analysis or plots are often 100 m or 200 m, for the plume dispersion flights in 40m are used. I don’t see a clear explanation of why these specific heights were used. Can you comment on this? You mentioned that the modeled potential temperature profiles show a strong gradient in the lowest 400 m, so why not higher levels than 100m/200m for the analysis?

Specific comments:
Page 2 Line 43: “as well as urban OH concentrations” I think this is not very relevant to your study, I think this part can be deleted
Page 3 line 67: “of up to 30%” is quite general. I think it would be good to be a bit more precise “typically between 5-30% strongly dependent on location and time”
Page 4 Table 1: Please change the covered periods to:
29 November 2020 – July 2021
July 2021 – 14 November 2021
April 2018 – November 2021
April 2018 – November 2021 (?)
What is meant by adjustment of surface albedo? Is the albedo database changed between the version? If yes, add the database which is used in the individual product.
Page 4 line 104, page 5 line 109: What is the phrase “residual cloud fraction” and “residual cloud scattering” mean? I would delete the word residual in line 104. Regarding line 109: the improvement of the new cloud algorithm is mainly that it provides a more realistic estimate of the cloud pressure for scenes with low cloud fractions.
Page 6 Figure 1: Strange references, they should perhaps instead appear at the first time mentioning the different models, here perhaps even simply omit or provide a reference to the data product itself, for TM5 probably from the TROPOMI product...? Why is there no reference for LOTOS-EUROS?
Page 6 line 146: You provide more detailed information about LOTOS-EUROS here, but please also provide some more information about the used TM5 and CAMS model data you used. Also, change the section title to Model data then.
Page 6 line 157: Isn't wind data also needed for this? If you also use wind data, please mention which dataset is used.
Page 7 Table 2: I think IFS is not a well-known abbreviation. Please add something like ECMWF Integrated Forecasting System (IFS).
Page 8 line 163: I think heterogeneity in general can be driven also by other factors like wind, lifetime, and other emission sources. The here observed heterogeneity might be mainly caused by the given factors. Please rewrite the sentence to avoid misunderstandings.
Page 8 line 178: How have you decided on the 100m height? Please give some additional information.
Page 9 line 1: How well can these horizontal scans represent the dispersion, especially for plumes not covered by these horizontal scans? How many plumes are analyzed in total, and how many have been covered by horizontal scans to account for plume dispersion? These in- and out-of-plume patterns were flown at an altitude of ~40 m (page 5, line 141). How representative is this height for the plume dispersion in 100m?
Page 10 line 226: Not all outflow cases show pollution close to the surface. Profile 9 doesn’t show pronounced pollution close to the surface but is an outflow case profile. Are there any reasons why this profile looks different? Please comment on this.
Page 11 Fig. 4: The figure would benefit from mentioning the wind direction, respectively, the kind of scenario (land outflow or clean). This could be added in the title of the individual plots or at least in the caption.

The plots in Fig. 4 are small and hard to read, especially the legend.
Page 12 line 233: “… boundary layer dynamics are a mix of sea and land characteristics for pixels at the coast”
Page 12 line 233: It is stated that the “TM5 profiles show too little NO2 in the lowest layer” compared to measurements, but here you are discussing the land outflow cases with profiles 1,2,7,8,9,10, and for these profiles, the agreement at the surface level with the measurements is nearly perfect except for profile ten which is also stated in line 235.

See also Page 13 line 262.

See also general comments.
Page 12 line 234: You wrote, “Nonetheless, the coarse TM5 columns show reasonable agreement with TROPOMI retrieved columns with the exception of profile 10.” I don’t get the context. Where are the TM5 and TROPOMI columns compared for the respective profiles? Fig. 5 shows the columns, but only for profile 1, Fig. 4 shows profiles but no columns. Do you refer to a Figure in the supplement?
Page 12 line 256: “profile 2 (which was sampled right before)” right before what, the flight path which is shown in Fig. 6?
Page 14 Fig. 7: Profiles 4,5,6 are taken on the same day. Are they taken in collocation to the same TROPOMI overpass? If yes, adding the locations of the other profile flights in the same figure would be helpful to see the location without looking into the supplement.
Page 14 line 273: You were just talking about profiles, but I think now you mean columns or both? Please be more precise.
Page 14 line 281: “…are spatially diluted over the area of the model grid cell” add ”especially for the coarse TM5 model”
Page 15 line 301: “total tropospheric columns of NO2” is misleading since the total column describes the column to the top of the atmosphere, better write only “We compare tropospheric vertical columns of NO2”
Page 15 line 302: How are the “measured columns” retrieved? Please add more information. From the profile flights up to 1400m added with TM5…
Page 15 line 309: columns determined from aircraft measurements
Page 16 Table 4: Why only a table, why not an additional scatter plot comparing the individual profiles?

Since measurements are only from 4 days, how many satellite pixels go into this comparison (correlation coefficient,…), are the profiles in different pixels when they were taken on the same day or different orbits (on days with two overpasses)?
Page 16 line 326: Measured how at the tower or aircraft measurements at the tower?

Page 17 Fig. 8: Make clear how measurements left/right plot were made, aircraft/tower.
Page 17 line 336: This is a bit misleading since you also showed that land outflow dominates for 6 of the ten measured profiles, and ship emissions and land outflow are difficult to distinguish.
Page 17 line 346: Of course, it is correct that your measurements show no significant pollution above 150m with a very shallow boundary layer height. Nevertheless, these findings are the result of only four measurement days. Please mention here also possible other meteorological conditions, different seasons…
Page 18 line 363: “This is likely an effect of the coarse TM5 resolution of 1◦x1◦ where ship emissions are smeared out over a larger area and time.” See general comments and earlier comments. This sounds like a general statement, but it can only be shown for the “clean” cases.
Page 19 conclusion: The conclusion and discussion on page 18 are very similar, especially lines 399-402 and 407-409 are very repetitive. Maybe the sections can be combined to avoid repetitions in the conclusion.

Technical corrections:
Page 1 line 4: delete “which was already”, since it is not so important here
Page 1 line 10: add “the” to “and the LOTOS-EUROS model”
Page 1 line 11: improves instead of improve
Page 1 line 11: Very long sentence. Maybe split it into: “The higher horizontal resolution in the regional CAMS ensemble mean, and the LOTOS-EUROS model improves the surface level pollution estimates.” and “However, the models still systematically overestimate NO2 levels at higher altitudes, indicating exaggerated vertical mixing in the models over the North Sea.”
Page 2 line 29: change “intensity of ocean going ships” to “the intensity of ocean-going ships”
Page 2 line 39: For a consistent, temporally and spatially complete approach
Page 2 line 42: add “NOx” to “study emissions patterns”
Page 2 line 51: change “therefor” to “therefore”
Page 2 line 51: used for the TROPOMI NO2 column retrieval
Page 3 line 71: In this study, we investigate aircraft-based in-situ measurements of NOx (and more) over the polluted North Sea with major shipping routes and nearby industrial and densely populated centers.
Page 3 line 75: delete “profile”
Page 3 line 80: delete one of the “of” and change “satellite trace gas retrieval” to “satellite trace gas retrievals”
Page 4 line 86: delete “and” and replace with a comma
Page 4 line 89: the abbreviation S5P was already introduced earlier
Page 4 line 90: delete “scientific quality”
Page 4 line 90: retrievals of various trace gases, including NO2, since April 2018
Page 4 lines 91-93: Make two sentences and replace “the instrument” with TROPOMI or S5P. Multiple sounds like a lot, but I think it's a maximum of twice a day. “With a swath width of approximately 2600 km TROPOMI has near daily coverage at the equator. At the latitude of the North Sea (52◦N) TROPOMI (or S5P?) frequently overpasses twice a day.”
Page 4 line 96: provide reference earlier: “with the DOAS-method (Platt and Stutz, 2008) in the visible spectrum (405-465 nm)”
Page 4 line 107: delete “these”
Page 5 line 109: change improved to modified
Page 5 line 114: and vertical profiles within the lower troposphere, from the sea surface (<30 m) to 1500 m.
Page 5 line 117: delete one “of” or replace it with “the”: for the purpose of monitoring the compliance
Page 5 line 126: replace “make sure” with “ensure”
Page 6 Figure 1: for measurements in flight heights below 200 m. Blue circles indicate the location of the spiral flights.
Page 8 line 163: add a reference to Fig. 2: within the spatial extent of a TROPOMI pixel, see Fig. 2.
Page 8 line 165: The text is repetitive: “In general, aircraft spatial sampling characteristics are not uniform across a TROPOMI pixel as evident from Fig. 2.” Is already mentioned two sentences earlier.
Page 8 line 168: Maybe change the section title to: Representativeness of NO2 vertical profile measurements
Page 8 line 170: I am not sure if I understand the meaning of the first sentence.
Page 8 line 176: “Fig 3 (left panel)” instead of “Fig. 4(a)”
Page 9 line 187: add (right panel)
Page 10 line 210: replace “sharp drop above the lowest 100 m” with “strongly decreasing within the lowest 100m”
Page 10 line 211: Split into two sentences, start with a new sentence after (Fig. 1): This is in agreement with…
Page 10 line 227: replace “at 200m and above” with “usually up to 200m and above”
Page 10 line 228: Split into two sentences: …and above. This leads to an overestimation in column NO2.
Page 11 line 232: replace “are originating” with “can originate”
Page 12 Fig.5: Caption is hard to read and sometimes misleading. Please rewrite, maybe something like this: NO2 columns (indicated by the bottom color bar) as seen by TROPOMI and several model products for the time of the first profile measurement. The aircraft measurements are overlayed in grey for flights above 200 m and in colors for flight heights below 200 m, indicated by the color bar on the right. Wind speed and direction from ERA 5 in 10m (?) are indicated by the arrows in the left panel.
Page 12 line 248: Please add the date to highlight that it is a different day than the 2nd June on which profile 2 was taken, which is mentioned later.
Page 13 line 264: The very shallow pollution layer visible in the NO2 measurements is also…
Page 14 line 269: “…as observed by TROPOMI, see Fig. S1 and Fig. 7.
Page 14 line 272: While the profiles were taken right above the shipping lane, marked by the blue circle in Fig. 7, in CAMS and LOTOES-EUROS the shipping pollution can be seen south of the profile location, caused by the northerly winds.
Page 15 line 291: replace “more accurate” with “modified”
Page 15 line 292/294: Move the sentence from line 294 “As the measured NO2 profiles only extend to 1400 m, 295 we use TM5 profiles to fill the gap to the tropopause.” to the beginning of the section line 292 after “…based vertical profiles.”
Page 15 line 293/294: the layer index “l” is sometimes italics, sometimes not
Page 15 line 298: “The shallow boundary layer depth over sea in combination…”
Page 15 line 302: “re-calculation of the AMFs and tropospheric NO2 vertical columns”
Page 15 line 304: Pearson correlation coefficient
Page 15 line 313: RMSE already defined in line 305
Page 15 line 314: Unit is missing
Page 15 line 320: replace contrast with a different word, maybe “differences”
Page 17 line 338: replace “retrieved NO2 columns” with “TROPOMI NO2 column retrieval”
Page 17 line 343: recalculate the AMFs and tropospheric NO2 columns
Page 17 line 348: in contrast to the model profiles on these measurement days
Page 18 line 382: missing bracket

Citation: https://doi.org/10.5194/egusphere-2023-1059-RC2
- AC2: 'Reply on RC2', Christoph Rieß, 21 Jul 2023
  
  Please see our response to the comments of referee 2 in the attachment.
  
  Citation: https://doi.org/10.5194/egusphere-2023-1059-AC2

Peer review completion

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

AR by Christoph Riess on behalf of the Authors (16 Aug 2023) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (27 Aug 2023) by Meng Gao

RR by Anonymous Referee #1 (05 Sep 2023)

RR by Anonymous Referee #2 (16 Sep 2023)

ED: Publish as is (01 Oct 2023) by Meng Gao

AR by Christoph Riess on behalf of the Authors (02 Oct 2023)

Short summary

Satellite retrievals of trace gases require prior knowledge of the vertical distribution of the pollutant, which is usually obtained from models. Using aircraft-measured vertical NO₂ profiles over the North Sea in summer 2021, we evaluate the Transport Model 5 profiles used in the TROPOMI NO₂ retrieval. We conclude that driven by the low horizontal resolution and the overestimated vertical mixing, resulting NO₂ columns are 20 % too low. This has important implications for emission estimates.

To new heights by flying low: comparison of aircraft vertical NO2 profiles to model simulations and implications for TROPOMI NO2 retrievals

Download

Interactive discussion

Peer review completion

To new heights by flying low: comparison of aircraft vertical NO₂ profiles to model simulations and implications for TROPOMI NO₂ retrievals