the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2024
Identifying Missing Sources and Reducing NOx Emissions Uncertainty over China using Daily Satellite Data and a Mass-Conserving Method
Abstract. This study applies a mass-conserving model-free analytical approach to daily observations on a grid-by-grid basis of NO2 from TROPOMI, to rapidly and flexibly quantify changing and emerging sources of NOx emissions at high spatial and daily temporal resolution. The inverted NOx emissions and optimized underlying ranges include quantification of the underlying atmospheric in-situ processing, transport and physics. The results are presented over three changing regions in China, including Shandong and Hubei which are rapidly urbanizing and not frequently addressed in the global literature. The day-to-day and grid-by-grid emissions are found to be 1.96±0.27 µg/m2/s on pixels with available priori values (1.94 µg/m2/s), while 1.22±0.63 µg/m2/s extra emissions are found on pixels in which the a priori inventory is lower than 0.3 µg/m2/s. Source attribution based on thermodynamics of combustion temperature, atmospheric transport, and in-situ atmospheric processing successfully identify 5 different industrial source types. Emissions from these industrial sites adjacent to the Yangtze River are found to be 160.5±68.9 Kton/yr (163 % higher than the a priori) consistent with missing light and medium industry located along the river, contradicting previous studies attributing the water as the source of NOx emissions. Finally, the results demonstrate those pixels with an uncertainty larger than day-to-day variability, providing quantitative information for placement of future monitoring stations. It is hoped that these findings will drive a new approach to top-down emissions estimates, in which emissions are quantified and updated continuously based consistently on remotely sensed measurements and associated uncertainties that actively reflect land-use changes and quantify misidentified emissions, while quantifying new datasets to inform the bottom-up emissions community.
- Preprint
(2190 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1903', Anonymous Referee #1, 09 Aug 2024
This study uses an emerging mass-balance based analytical framework to estimate spatially resolved (daily) NOx emission fluxes and (monthly) associated properties of NOx sources and processes from high-resolution satellite (TROPOMI) retrievals in eastern China. The results are then interpreted to derive 1) insightful discussions regarding these properties, especially the NOx/NO2 emission ratio above different types of sources; 2) significant overestimation of NOx emissions above strong NOx sources by the a priori emissions inventory (MEIC), and underestimation in the downwind areas; and 3) strong industrial and shipping NOx above and adjacent to rivers in China. Most of the results and insights over the highly emitted study region are unprecedented in existing literature. The paper is overall well-written with sound methods and results. At the same time, some critical details are missing and need clarification. I support the publication of this manuscript, provided that the following comments can be addressed.
Main comments:
1) In the current description of the approach and Figure 3, it is unclear to me how the NOx emissions are assumed in the first fitting step to derive three monthly coefficients. Of course, they cannot vary everyday as in the second step (otherwise the equations cannot be solved). Did you directly use MEIC emissions? Or did you assume monthly invariant emissions? Please clarify this important issue, and verify that your approach is robust against your assumption (the fitted alpha1-alpha3 changes little with imperfect assumptions on NOx emissions in the first step).
2) It is unclear to me the locations of different sources used in Figures 4 and 5. An idea is to present maps of the derived monthly alpha1 (e.g., median values in representative months) overlaid by the different types of sources. Such maps will present an overview of combustion efficiency/types across the domain, and their association with the labeled emitters, greatly aiding the interpretation of Figures 4 and 5. Please consider adding such information in the revision or as an appendix.
3) In Line 236, it is mentioned that emissions are monitored by the Continuous Emission Monitoring Systems (CEMS) network. However, no validation attempt is made to compare their derived NOx emissions against these ground truths. Please consider adding this important validation or clarify why it was not conducted.
4) Section 3.4 appears abruptly in the current manuscript. I suggest some motivation words regarding NOx emissions above water should be presented in the Introduction Section. In addition, although it is obvious, please quantitatively discuss how the estimated NOx emissions cannot be explained by river/lake emissions (e.g., comparing the magnitudes vs. typical emission numbers in the literature).
Specific comments:
1) Please indicate a typical unit conversion factor from the estimated emissions at μg/m2/s to some more widely used unit (e.g., T/year) so that they are broadly understood and comparable.
2) In the revision, please use hyphen (-) in the words separated by two lines.
3) Line 45: change "This approach also includes..." to "This approach can also be applied to biomass burning emissions by including…".
4) Line 55-58: Indeed, these existing approaches focus on monthly emissions. But evidences of information content (e.g., degrees of freedom of the framework) and validation (e.g, vs. daily CEMS emissions?) should be presented to justify the applicability of the proposed approach to estimate daily NOx emissions.
5) Line 62 criticized the "inability to scale zero emissions" as a drawback of previous methods. However, Line 156-157 directly omit any locations with small MEIC emissions in the fitting. Is there a better way to resolve emission locations that are misrepresented by the a priori in your method?
6) Line 78-79: "The results are checked against independent measurements of NOx emissions flux". Where in the paper?
7) Figure 1: While Regions 2 and 3 seem representative to me, please explain why some large and industrialized cities in the nearby Henan and Shanxi provinces are excluded in Region 1?
8) Line 145: please clarify what weights (surface area?) are used in the re-gridding.
9) Line 149: Higher resolution (e.g., 1-km) emissions inventory is available in China (e.g., https://doi.org/10.1016/j.scib.2020.12.008).
10) Line 160-165: the ERA5 wind is at 0.25 degree resolution. How would the averaging-out of the wind fields affect the divergence calculation at 0.05 degree?
11) Line 215: please explain why only alpha-1 is bootstrapped? Is it considered as the dominant contributor to the uncertainty of the system?
12) Line 251-252: this observation is consistent with the very dense distribution of strong NOx emission sources in China (whereas sources are more distant in the US and Europe), as will be presented in Figures 6 and 7. It is suggested to synthesize these findings in the discussion.
13) Line 263: It seems to me that distributed residential heating might have relatively lower NOx/NO2 because of the relatively lower temperature and exposure to ambient ozone, than large point sources like power plants. Is the heating in North China somewhat different?
14) Line 270-273: Overall there is very weak seasonal variation in the derived lifetimes. This is in contrast to other studies where the NOx lifetime in summer and winter can differ by a factor of 4 (e.g., https://doi.org/10.5194/acp-20-1483-2020). How should we interpret these differences?
15) Figure 4d: Please adjust the labels in the x-axis to align correctly with each bar. For example, the distances between Wuxi and Changzhou (or between Beihai and Hongkong) are somewhat too large?
16) Section 3.1: Besides the 5 sources discussed here, what is the typical range of NOx/NO2 for vehicle emissions? Can this information help explain the low values of cities like Macao (should be dominated by vehicle emissions)?
17) Line 311: Why does this NOx/NO2 ratio of power plant sources increase with more power plant emissions (activities)? Is it because your derived values are still contaminated by other emission sources for these grids (thus these July months are more dominated by power plants)?
18) Table 2: Very interesting classification method. Can this be used to categorize sources of grid cells where the dominant source is unknown (e.g., on the maps of alpha1)?
19) Line 348-357: It is suggested to make a table summarizing these numbers, and only highlights several of them in the main text.
20) Line 366-373: Do you mean "urbanization" when referring to "land surface changes"? I do not foresee a clear connection between rapid "land changes" and "emission changes".
21) Figure 7 and associated discussion: It looks to me that Figure 7b mostly highlights locations with strong NOx emissions (e.g., urban centers), and Figure 7a contains their downwind locations. Can part of this be explained by the transport (smearing) effects that are not fully accounted for by the inversion?
22) Line 461: Please clarify why UV radiation is related with alpha1 (NOx/NO2).
Citation: https://doi.org/10.5194/egusphere-2024-1903-RC1 -
AC2: 'Reply on RC1', Jason Cohen, 17 Oct 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1903/egusphere-2024-1903-AC2-supplement.pdf
-
AC2: 'Reply on RC1', Jason Cohen, 17 Oct 2024
-
RC2: 'Comment on egusphere-2024-1903', Anonymous Referee #2, 28 Aug 2024
Review of the paper entitled “Identifying Missing Sources and Reducing NOx Emissions Uncertainty over China using Daily Satellite Data and a Mass-Conserving Method” by Lingxiao Lu et al.
This study introduces a novel mass-balance based analytical framework designed to rapidly and flexibly quantify NOx emission fluxes with high spatial and temporal resolution, using daily observations from TROPOMI across three rapidly changing regions in eastern China. The study effectively quantifies source attribution by revealing unprecedented insights into NOx/NO2 emission ratios across five industrial sources. The results also indicate significant discrepancies between observed emissions and those predicted by the MEIC a priori inventory. Furthermore, the paper emphasizes the substantial NOx emissions linked to small and medium industrial and residential activities in regions adjacent to rivers. This paper stands as a well-executed and valuable piece of research which is overall well-written. The methodology is emerging and innovative, and the results contribute valuable insights into the emissions landscape, there are certain critical details that require clarification. In my opinion, this manuscript is worthy of publication after minor revision.
Specific comments:
The paper presents fitted alpha1-alpha3 values, but it is important to assess how sensitive these values are to variations in the initial assumptions about NOx emissions. Did the authors conduct a sensitivity analysis to determine if the alpha values remain stable under different scenarios? This information would provide insight into the reliability of the emissions estimates and the overall methodology.
In Section 2.1, line 110, why did you choose to exclude column NO2 has a climatology smaller than 1.43x1015 molec/cm2. How was this number determined?
In Section 3.1, Figure 5 plots the distribution of monthly NOx/NO2 over grids from different sources, how many facilities were counted for each emission source?
In Section 3.2, a key aspect to consider is the performance of the fitting process and the resulting errors in each parameter, as this is essential for assessing the reliability of the MCMFE framework. Given that the fitting incorporates all observations, does it remain unbiased across different months and grid cells?
In Section 3.4, Figure 8 plots have the unit of (Kton yr-1 cell-1), however, the size of a cell is not mentioned in the context.
Did the authors use MEIC emissions data or assume monthly invariant emissions?
Minor comments:
Line 215 and line 338, “densitiy” should be “density”.
Line 127, “boarder” should be “border”.
Line 232, typo error “heat productin and supportion”. Line 234, “form” should be “from”.
Line 262, typo error “reated”.
Line 283, typo error “Qingdai”.
Line 363, typo error “differnces”.
Line 289 and 424: However -> Moreover
Line 360, 380 and 388: differences -> the differences
Line 18: Kton/year -> Kt/year. Please ensure that units are consistently presented and aligns with standard publication practices.
L106: “… areas …”
L153: “the 0.05°x0.05° grid”
L234: “… obtained from the …”
L240: “… the coefficient results …”
L391: “… Xingtai …”.
L446: “… Province …”
Citation: https://doi.org/10.5194/egusphere-2024-1903-RC2 -
AC1: 'Reply on RC2', Jason Cohen, 14 Oct 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1903/egusphere-2024-1903-AC1-supplement.pdf
-
AC1: 'Reply on RC2', Jason Cohen, 14 Oct 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 295 | 70 | 22 | 387 | 11 | 17 |
- HTML: 295
- PDF: 70
- XML: 22
- Total: 387
- BibTeX: 11
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1