Decadal changes in global surface NO<sub><i>x</i></sub> emissions from multi-constituent satellite data assimilation

Miyazaki, Kazuyuki; Eskes, Henk; Sudo, Kengo; Boersma, K. Folkert; Bowman, Kevin; Kanaya, Yugo

doi:https://doi.org/10.5194/acp-17-807-2017

Articles | Volume 17, issue 2

https://doi.org/10.5194/acp-17-807-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Special issue:

Ten years of Ozone Monitoring Instrument (OMI) observations...

https://doi.org/10.5194/acp-17-807-2017

© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 17, issue 2

Research article

| Highlight paper

|

27 Jan 2017

Research article | Highlight paper |

| 27 Jan 2017

Decadal changes in global surface NO_x emissions from multi-constituent satellite data assimilation

Kazuyuki Miyazaki, Henk Eskes, Kengo Sudo, K. Folkert Boersma, Kevin Bowman, and Yugo Kanaya

Download

Final revised paper (published on 27 Jan 2017)
Supplement to the final revised paper
Preprint (discussion started on 27 Jun 2016)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Review to the manuscript “Decadal changes in global surface NOx emissions from multi-constituent satellite data assimilation” by Miyazaki et al.', Anonymous Referee #1, 30 Jun 2016
- AC1: 'Response to referee #1', Kazuyuki Miyazaki, 23 Sep 2016
RC2: 'Review of Miyazaki et al.', Anonymous Referee #2, 23 Aug 2016
- AC2: 'Response to referee #2', Kazuyuki Miyazaki, 23 Sep 2016

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Kazuyuki Miyazaki on behalf of the Authors (23 Sep 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (03 Oct 2016) by Gerrit de Leeuw

RR by Anonymous Referee #2 (13 Oct 2016)

Suggestions for revision or reasons for rejection

The authors did some effort to address my main concerns. In particular the possible various limitations associated with the method are better explained in the manuscript. A section on validation using independent NO2 data has been added, although I found it mostly unconvincing due to several issues (see further below). As in my first review, my primary concern is the fact that the authors fail to explain the role of the different datasets in the assimilation. It is said repeatedly that non-NO2 observations have a large impact on the optimization of the emissions. And this is indeed shown in the new Table 4. But what is, more precisely, the role of each dataset? We are left almost clueless on that matter. It is not enough to claim that the lifetime of NOx is better represented by the model when those observations are used. I would like a discussion explaining, qualitatively and quantitatively how the measurements of ozone and CO influence the assimilation. It is therefore necessary to, first, present the model biases for CO and O3, and secondly, discuss how the correction of those biases by the assimilation impacts the optimization of NOx emissions.

Regarding the diurnal cycle (my second major comment in my previous review), the authors have not addressed the main issue, which is that the modification to the diurnal cycle of emissions deduced from satellite measurements is not credible as it implies much stronger rush hour emission peaks even in regions where mobile emissions (cars) are not the main source category. The most negative values of the Etc parameter are found (not in Mongolia but) in Inner Mongolia, i.e. in Northern China, around 110 W, 41 N, in a region with very strong emission trends (see Figure 12) due to anthropogenic emissions, i.e. power plants and industries (not cars). This should be mentioned and shortly discussed.

Comparisons with airborne and lidar NO2 measurements (from ARCTAS, INTEX-B and DANDELIONS) have been added (Figure 4). But the ARCTAS profile is almost useless as the assimilation does not change the NO2 tropospheric profile, except in the upper troposphere (UT). The mechanism by which NO2 is increased by the assimilation in the UT is not explained, except for the fact that it is related to HNO3 MLS observations. The text mentions the effect of inter-species correlations. But it is the first (and only) time that such correlations are mentioned. More details would be needed to explain how such correlations are set up in the system. Furthermore, the ARCTAS NO2 measured during ARCTAS in the UT is known to be too high. Browne et al. (ACP 11, 4209-4219, 2011) showed that at low temperature, a large fraction of the measured XNO2 is due to dissociation of CH3O2NO2 and HO2NO2 in the inlet prior to detection, leading to a large overestimation of NO2
in the UT. Also for DANDELIONS, only morning measurements are used for validation, whereas afternoon observations are rejected without a good reason. Only two INTEX-B flights are used, and I find the presentation awkward. For the March 9 flight, I find very weird that the observed values are so low in the layer closer to the surface (most measurements are well below 100 ppt). I checked the INTEX-B files (from www-air.larc.nasa.gov/missions/merges/) and I estimate that the average NO2 in that layer for that day was 228
ppt, even when excluding the Mexico and Houston areas. Please verify the data selection for that flight. Furthermore, the authors claim to see an improvement in the slopes of the linear regressions, but what meaning is there in such slopes when the correlations are so low? It would be useful to show vertical profiles for INTEX-B (as for ARCTAS) and possibly for another campaign like INTEX-A or the more recent ones (SEAC4RS).

Minor comments:

The authors response suggested that the HNO3-forming channel of the NO+HO2 reaction is taken into account in the model. I'm very surprised by that. Some words should be provided in the model decription, including the
references for the rates (is the effect of water vapor also considered?)

Also in their response, the authors added a sentence "The summertime peak enhancement is obvious over remote regions such as high-temerature agricultural land over the South Atlantic (...)". Agricultural land over the
Atlantic? Please rephrase.

Regarding the improved ozone due to higher NOx emissions: that this increase would actually deteriorate ozone
in other models (e.g. Geos-Chem, cf. Travis et al. ACPD 2016) calls into question the reality of the NOx
emission increase. This should be mentioned and possibly discussed in the manuscript.

I find very weird that the authors cannot provide any indication regarding the lifetime of methane in their model. This is an essential and very standard metric of any global atmospheric model. It would be also very useful for the discussion of the assimilation results.

The discussion on the changes of the NOx lifetime states that "both the concentration assimilation (mainly TES O3 and MOPITT CO measurements) (...) lead to an increased in the OH concentrations". This is probably true but needs to be demonstrated. Also, rephrase, e.g. "both the assimilation of non-NO2 compounds (mainly TES O3 and MOPITT CO measurements) (...)".
In the next sentence, it is stated that HO2+NO is enhanced, and that the NOx lifetime is decreased due to
higher OH in the multiple-species assimilation (compared to the model simulation). This is very probably correct,
but I don't see any proof that the non-NO2 observations are essential here for that respect. Therefore, the last
sentence "demonstrate the utility of multiple-species assimilation..." is unsubstantiated.

The authors also did not answer my questions on the trends of NO2 concentrations and NOx emissions over Europe. The manuscript suggests that NO2 has become more long-lived. Surely you can check in your model outputs whether e.g. OH concentrations show a trend. The other explanation "a shift in NO2:NOx emission ratios related to the increasing share of European diesel cars could have occurred" is very strange, it is like if the authors cannot verify what they have in their model.

Other corrections

Page 13 line 7 "the possibility FOR improving"
Page 13 line 14 delete "and" before "used in data assimilation"
Page 15 line 12-13 explain why representativeness errors would be smaller in the morning compared to the afternoon.
Page 15 line 24 replace "corrected" by "sampled"
Page 16 line 5 replace "at polar region" by "in polar regions"
Page 16 line 28 replace "principle" by "main"
Page 20 line 13 insert "only" before "a small effect"
Page 20 line 26 what is meant by "commonly"? Rephrase.
Page 21 1st full paragraph: the difficulty to represent the measurements would disappear when using a large number of measurements, because the errors on the averages will cancel out. The solution is therefore to use larger datasets than used here.
Page 25, line 25 The temporal shift is actually larger than 1 month.
Page 33 line 21 Replace "adjusted" by "modified"
Page 34 line 4: "The monthly total global emissions decrease by up to 6 TgN": that value (6TgN) is impossibly high, please check. Remember that the total global NOx source if of the order of 40-50 TgN per year.
Page 35 line 1: "influences" --> "influence"
Page 36 line 15 "It was confirmed" --> "It is found" (??)
Page 38 line 4 "misleading" --> "inappropriate"
Page 46 line 7 Add "line" after "black"
Page 46 line 9 Delete "six"

Hide

ED: Reconsider after major revisions (23 Oct 2016) by Gerrit de Leeuw

AR by Kazuyuki Miyazaki on behalf of the Authors (21 Nov 2016) Author's response Manuscript

ED: Reconsider after minor revisions (Editor review) (08 Dec 2016) by Gerrit de Leeuw

AR by Kazuyuki Miyazaki on behalf of the Authors (08 Dec 2016) Author's response Manuscript

ED: Publish as is (09 Dec 2016) by Gerrit de Leeuw

Download

Article (11352 KB)
Full-text XML

Short summary

Global surface emissions of nitrogen oxides (NO_x) over a 10-year period (2005–2014) are estimated from assimilation of multiple satellite datasets. We present detailed distributions of the estimated NO_x emission distributions for all major regions, the diurnal, seasonal, and decadal variability. The estimated emissions show a positive trend over India, China, and the Middle East, and a negative trend over the United States, southern Africa, and western Europe.

Decadal changes in global surface NOx emissions from multi-constituent satellite data assimilation

Download

Interactive discussion

Peer-review completion

Decadal changes in global surface NO_x emissions from multi-constituent satellite data assimilation