the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Can TROPOMI-NO2 satellite data be used to track the drop and resurgence of NOx emissions between 2019–2021 using the multi-source plume method (MSPM)?
Janot Tokaya
Christian Mielke
Kevin Hausmann
Debora Griffin
Chris McLinden
Henk Eskes
Renske Timmermans
Abstract. Nitrogen dioxide (NOx) is an important primary air pollutant, dominantly produced by anthropogenic, mostly combustion based, activities from sectors such as industry, traffic and transport. NOx is directly linked to negative health and environmental impacts. Currently, the construction of emission inventories to keep track of NOx emissions is based on official national reported emissions and proxies such as activity data as well as direct measurements. The effort to properly construct an accurate inventory is significant and time consuming which causes a reporting offset between one and five years with respect to the current date. Next to this temporal lag difficulties in composed inventories can arise from legislative and protocol differences between countries and over time in reporting of emissions. Satellite based atmospheric composition measurements provide a unique opportunity to fill this gap and independently estimate emissions on a large scale in a consistent, transparent and comprehensible way. They give the possibility to check for compliance with emission reduction targets in a timely manner as well as to observe rapid emission reductions such as experienced during the COVID-19 lock-downs. In this study we apply a consistent methodology to derive NOx emissions over Germany for the years of 2019–2021. For the years where reporting is available differences between satellite estimates and inventory totals were within 100 kt. The large reduction of NOx emissions related to the COVID-19 lock-downs were observed in both the inventory and satellite derived emissions. The recent projections for the inventory emissions pointed to a recovery of the emissions towards pre-COVID19 levels this increase was not observed. While emissions from the larger power-plants did rebound to earlier levels, others sectors such as road transport and shipping did not and could be linked to a reduction in the number of heavier transport trucks. This again illustrates the value of having a consistent satellite based methodology for faster projections to guide and check the conventional emission inventory reporting. The method described in this manuscript also meet the demand for independent verification of the official emission inventories, which will enable inventory compilers to detect potentially problematic reporting issues. Transparency and comparability, two key values for emission reporting, are thus bolstered by this technique.
- Preprint
(16634 KB) - Metadata XML
- BibTeX
- EndNote
Enrico Dammers et al.
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2022-292', Anonymous Referee #1, 03 Feb 2023
This paper presents an interesting comparison of TROPOMI-NO$_2$ satellite data and inventory data for Germany over 2019--2021. Scientifically the text is well organised, but the text needs to be tightened up and made more readable. The manuscript deserves to be published after attending to the points raised below.
General points:
- The manuscript needs to be thoroughly checked for the logic and for the English grammar. Many statements are unsupported or confusing. Many commas are misplaced, making sentences difficult to read and understand. Some sentences make no sense by themselves, but do make sense if joined to the next sentence. Many citations use parentheses incorrectly. Subscripts are sometimes missed.
- The authors assume an NO2 lifetime of 4 hours when deriving emissions, but shouldn’t this lifetime impact the footprint on the sources when comparing to the satellite data? With 5 m/s winds for example 4 hours of advection transports the NOx ca. 72 km, but the Gaussian smoothing used to produce Fig. 3 uses a σ of 1 grid cell, ca. 10km. Would more smoothing produce better SSIMs and more accurate inversion of the emissions?
- Related to this, the TROPOMI-derived emission uncertainty is stated to be 30–40% (p21, L428), but it isn’t clear where these numbers come from. The text on p21 discusses many sources of uncertainty (including the lifetime issue), but how were these combined to make 30–40% and the error bars (whatever they are) in Fig. 8?
- p2, L39–49, and also p20, L403 onwards. I missed a discussion of the known problems of real-world emissions. For example, as cited in Oikonomakis et al. (2018), several studies showed a significant discrepancy (a factor of 2–4) in the NOx emissions from light-duty diesel vehicles between two driving cycles, and Anenberg et al. (2017) and found similar issues for heavy duty vehicles, indicating inadequacy of the testing procedures to capture real-world emissions. The diesel-gate scandal was also a good example of the limitations of emissions reporting (Jonson et al., 2017).
Other points:
• p1, L12, Explain 100kt as percentage, so that the reader knows if this is a large or small number.
• p1, L14, Add a year here so that the reader knows when this recovery was ‘projected’.
• p1, L17. It sounds odd that satellites help faster ”projections”, since that term is usually reserved for future values. Re-phrase.
• p1, L18. Change meet to meets, or method to methods. Check such things throughout the manuscript.
• p2, L2. Add also a more recent reference than Crutzen 1970.
• (nit-picking I know): Say ‘largely’ primary. Some NOx is produced by lightning, and NO2 is mainly produced from ozone reactions with primary
NO.• p2, L27. The units should be formatted correctly, not in italics. Also, be consistent. The unit on L28 has a space between g and m, whereas on L27
it doesn’t.
• p2, L30. Give reference for the statement about acidification and eutrophication.
• p2, L36. There is no such thing as the ‘Geneva Convention ...’, in this context at least. The authors mean the Air Convention or CLRTAP equivalents, e.g. https://unece.org/environment-policy/air. ‘Nations’ is also not an appropriate reference; maybe ’UN-ECE’ or similar.
• p2, L44. The word projections confused me here and elsewhere. Usually the term is used for future scenarios, e.g. for 2030 or 2050. Here I think the authors just mean emission estimates.
• p3, L60. Give info on this ‘unprecedented horizontal resolution’, or refer to appropriate section for details.
• p3, L65. Refer to appropriate section for details of code and availability.• p3, L71. Tell the reader where this is ‘described further’.
• p3, L75 (also p7). What is Umweltbundesamt (UBA) for those not familiar with German institutions?
• p3, 1st paragraph. This section is somewhat repetitive of Sect. 1, and isn’t really ‘Methodology and Datasets’. Some is also repeated, or better placed, in Sect. 2.1.
• p4, L104. Where does this 300% number come from?
• p4, Sect. 2.1.1. Some of the text here is also more introductory material (e.g. L104 onwards) than technical description of the emission inventory.
• p4, L120. Change ‘mol’ to emissions (mol is a unit, not a quantity).
• .5, L124. NOx is a mixture of NO and NO2 , so one needs to specify the assumed molecular weight associated with your 5 kt NOx per year figure, or state as e.g. kt(N) NOx per year.
• p5, L130 onwards. Same issue with NOx units and emission amounts.
• p5, L151. What is ATBD?
• p6, L166. Say ‘well correlated with ...’; the sentence was difficult to read.
• p6, L165–173. Various statistics are given concerning bias, but are the instruments being compared with (MAX-DOAS, PANDORA) free of bias themselves? Are some of the difference due to problems with these instruments?
• p7, L205. Mangled ‘from in the naive’?
• p10. The ‘Column’ term in Eqns. (2) and (7) looks very ugly. Use a symbol, as is done for all other terms. In any case, shouldn’t this be VCD?
• p10, equations (3)–(6). Give in order of usage, thus σ1 before f (x, y), λ1 before g(y, s).
• p10, L255. Where is ai explained?• p10, L258: ‘Following Beirle et al. (2016) we assume a lifetime of about 4 hours(±25%)’. I can’t find the terms hour or lifetime in Beirle 2016, and that paper deals mainly with the stratosphere. Why didn’t you use estimates of NO2 lifetime from LOTOS-EUROS for Germany? Does the 25% estimate really capture the uncertainty here?
• p11, L287. I didn’t find the factor 1.32 in Beirle et al. (2016) either.
• p11, L292-293: ‘The gridded NFR data .... summed to the ... grid’. Does gridded data need to be gridded? CLRTAP inventories are usually gridded
by NFR categories.
• p12, L307. Why a comma after ‘way’? This is just one example of a common problem.
• p12, L323. Again, somewhat sloppy. Fig. 4 doesn’t say anything about previous sensors, and the text doesn’t explain what the authors are thinking. If you make comparative statements, back them up.
• p12, L330. Here I also wonder about the 4-hour footprint issue mentioned above.
• p14, L345. Remove ‘the before mentioned’.
• p14, L349. How should non-Germans know where the A1 motorway is?
• p15, Fig 5. Explain letters in top-right fig. Also explain whether positive values (red color) means that the satellite has more or less emission than the inventory.
• p16, Fig 6. What are the triangles? What are the varuious letters (NEU, WW, ...)? The latter are explained in the text, but the caption should be informative.
• p16, L372. Why are 65 kt NOx added only at this stage?
• p18, Fig. 8. Again, the caption explains too little. What are the error bars? Be explicit and say slight rise in ‘reported’ emissions.
• p18, L375. Start a new paragraph frpm ‘Emissions sources ̈, so that the reader knows the subject has changed.
• p18, L380. Fig. A3 doesn’t support that the Agri emissions are spread out across the country, at least not if the text is about the >50% region.• p18, L382. Why are ‘non-agricultural sources’ mentioned here? Only 3 sources are addressed, so many sources are excluded.
• p20, L395. When starting the discussions, be explicit that the reported emissions are for Germany.
• p20, L402. Why does Fig. 5 ‘hint’ at a small and widespread source? The values seem close to zero in most areas.
• p20, L403. Again start a new paragraph when the subject is changing.
• p21 L432. ‘improved’ - from what?
• p21, L436. ‘approach u’ typo
• p21, L437. CAMS-Europe - reference?
• p21, L440. ‘detect’ should be ‘detection’
• p21, L445. Mangled sentence.
• p21, L445. How do you know that the 4h timescale is correct for Germany as a whole in 2019?
• p21, L448. Lifetimes are not only location dependent; they depend on complex interactions between meteorology, chemistry and vegetation state (via deposition).
• p21, L453. What is meant by several %? That number seems low.
• p22, L472. What is ‘(8)’?
• p22, L484. The web-tool is mentoned, but I would hardly say it was ‘presented’.
• p23, L496. What is ‘ads’? Provide a web address.
• p23, L498. No need to use words like ‘truly’ in a scientific statement.
• p23, L502. It was discussed earlier that the resolution is not really 3.5x5.5km2, and the derived-emissions resolution are certainly not at that level.
• p22-24. Again, one sinlge paragraph over more than a whole page! Break up the text into separate topics.• p23, L508. Reference the Green Deal.
• p23, L513. The sentence starting ‘While’ ends abruptly, making no sense. Also, rephrase ‘whole there here developed methodology’.
• p23, L518. ‘tooling’ should be ‘tools’.
• p23, L52. What is ‘link website, mode fields’?
• Fig A1. Be explicit: NOx emissions.
• p28, Fig A4. Again (as with Fig.5), explain what positive values mean.
• p30, L562. Mangled NO 2.References
Anenberg, S.C. et al., Impacts and mitigation of excess diesel-related NOx emissions in 11 major vehicle markets, Nature, advance online publication, http://dx.doi.org/10.1038/nature22086, 2017.
Jonson, J.E. et al., Impact of excess NOx emissions from diesel cars on air quality, public health and eutrophication in Europe, Environ. Res. Lett., 12, 094 017, http://stacks.iop.org/1748-9326/12/i=9/a=094017, 2017.
Oikonomakis, E. et al., Low modeled ozone production suggests underestimation of precursor emissions (especially NOx) in Europe, Atmos. Chem. Physics, 18, 2175–2198, https://doi.org/10.5194/acp-18-2175-2018, 2018.Citation: https://doi.org/10.5194/gmd-2022-292-RC1 - AC1: 'Reply on RC1', Enrico Dammers, 13 Jul 2023
-
RC2: 'Comment on gmd-2022-292', Anonymous Referee #2, 03 May 2023
This study compares the emissions of NOx estimated from NO2 satellite data with inventory data from 2019 to 2021, which provides a useful method to evaluate and improve NOx emission inventory. The manuscript is overall well organized, but scientifically needs clarification and more discussion regarding the uncertainties.
Major comments:
- Satellite images only contain information of NO2, but NOx from the emission inventories includes NO and NO2. In emission inventory, the ratio of NOx to NO2 is different from the ratio in ambient concentration. The conversion of NO to NO2 changes the ratio. The study uses a factor of 1.32 (based on ambient concentrations) for all sources definitely leading to uncertainties.
- Photochemical reactions are different among seasons and day-night. The life time of 4 hours for NO2 uniformly seems unreasonable for all days during 2019-2021. Radiation could be a good indicator for the lifetime.
- The comparison of inventory and Tropomi in Figure 3 is not clearly, showing the difference is better for readers.
- The abstract needs to be modified as the currently version is not clear about the method and the key results.
- The discussion of uncertainties is qualitative rather than quantitative. A ranking of uncertainties from different assumptions is helpful for assessing the results when this method is used for other cases.
Citation: https://doi.org/10.5194/gmd-2022-292-RC2 - AC2: 'Reply on RC2', Enrico Dammers, 13 Jul 2023
Enrico Dammers et al.
Enrico Dammers et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 663 | 195 | 16 | 874 | 7 | 11 |
- HTML: 663
- PDF: 195
- XML: 16
- Total: 874
- BibTeX: 7
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
