the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Mar 2024
Regional and sectoral contributions of NOx and reactive carbon emission sources to global trends in tropospheric ozone during the 2000–2018 period
Abstract. Over the past few decades, the tropospheric ozone precursor anthropogenic emissions: nitrogen oxides (NOx) and reactive carbon (RC) from mid-latitude regions have been decreasing, and those from Asia and tropical regions have been increasing, leading to an equatorward emission redistribution. In this study, we quantify the contributions of various sources of NOx and RC emissions to tropospheric ozone using a source attribution technique during the 2000–2018 period in a global chemistry transport model: CAM4-Chem. We tag the ozone molecules with the source of their NOx or RC precursor emission in two separate simulations, one for each of NOx and RC. These tags include various natural (biogenic, biomass burning, lightning and methane), and regional anthropogenic (North American, European, East Asian, South Asian etc.) precursor emission sources. We simulate ~336 Tg O3 with an increasing trend of 0.91 Tg O3/yr (0.28 %/yr), largely contributed (and trend driven) by anthropogenic NOx emissions and methane. The ozone production efficiency of regional anthropogenic NOx emissions increases significantly when emissions decrease (Europe, North American and Russia-Belarus-Ukraine region’s emissions) and decreases significantly when emissions increase (South Asian, Middle Eastern, International Shipping etc.). Tropical regions, despite smaller emissions, contribute more to tropospheric ozone burden compared to emissions from higher latitudes, consistent with previous work, due to large convection at the tropics thereby lifting O3 and its precursor NOx molecules into the free troposphere where ozone’s lifetime is longer. We contrast the contribution to tropospheric ozone burden with that of the contribution to the global surface ozone. We simulate a smaller relative contribution from tropical regions to the global mean surface ozone compared to their contribution to the tropospheric ozone burden. The global population-weighted mean ozone (related to ozone exposure) is much larger compared to surface mean, mainly due to large anthropogenic emissions from densely populated regions: East Asia, South Asia, and other tropical regions, and a substantial contribution from international ship NOx emissions. The increasing trends in anthropogenic emissions from these regions are the main drivers of increasing global population-weighted mean ozone.
- Preprint
(2318 KB) - Metadata XML
-
Supplement
(797 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-432', Anonymous Referee #1, 08 Apr 2024
Comments:
This manuscript details an application of tagging technique to the attribution of tropospheric ozone changes, to identify the contributions of regional and sectoral NOx/RC emissions. The authors present a detailed calculation to demonstrate the influence of the equatorward shift of surface anthropogenic NOx emissions, and how the ozone burden and its trend are contributed by emissions from different regions and sources. The manuscript is well-organized and written. Overall, I think this is a neat study and fits the scope of ACP. However, certain aspects deserve further discussion before publication.
Major comments:
1) The equatorward shift of precursor emissions is a key point to make in this study but is mainly supported by references and inference in this study, for example, line 475. It is important to give more direct evidence that more precursors or O3 itself have been lifted into the free troposphere over past years (e.g. trends of tagged surface anthropogenic NOX in the free troposphere), and it is worth further discussing whether the increase of tropospheric ozone burden (largely in the free troposphere) is mainly contributed by the lifted precursors or the lifted O3 based on your tagged simulations, if possible.
2) For East Asia, when discussing the trend, it makes no sense to discuss the whole period, and the two periods (prior-2011 versus post-2011) should be separated.
3) It is confusing to me how the stratospheric influx is tagged that it could be attributed to NOx and RC. In line 255, the tropospheric ozone is attributed to oxidation of N2O, but if NOx-tagged means tagging the NOx emissions, how could it tag the NOx produced from N2O? It would be clearer if the authors clarify the tagging technique in more detail. Also, as changes of stratosphere-to-troposphere exchange are also important for the trends of tropospheric ozone burden (especially the free troposphere), please include the discussion of STE in the summary and conclusion, and also abstract.
4) Line 231. The methane lifetime trend is inconsistent with some other studies that suggested a longer lifetime of methane due to the reduction of OH. The simulated OH trend should also be shown in Fig.4 and be compared to other model studies to give a better sense of the accuracy of the model mechanism in this study and give a better reasoning for the methane lifetime decrease.
Minor comments:
1)Ozone enhancements in the free troposphere are linked to the monsoon-induced transport. The changes in monsoon and other meteorological factors (i.e. climate noise) may also affect the amount of lifted surface ozone/precursors, and the they may show a trend under global warming. This should be mentioned that the simulated trends are not only contributed by changes in precursor emissions.
2) The authors showed the emissions from different regions and pointed out an equatorward shift. It would be better to give a quantified number of how the total emissions in Tropics versus Midlatitudes change over time, to quantify that the decrease of emissions in East Asia after 2011 won't alter this phenomenon.
3) Line 40. “or” is a typo.
4) Line 172. The comparison of simulations to ozonesonde uses different periods. The authors should also show the results of the overlapping period (2000-2010) to give a better sense of how the model accuracy is.
5) Line 224, is not a good reasoning to me. What is the “inter-annual variability”?
6) Line 370. Isn’t it actually a reasonable OPE result in East Asia? Seeing the emissions experienced an increase and then decreased after 2011.
7) All four-subplot figures (3, 5-8) need to add y-axis labels and add titles for the first row (not just the second row).
8) Format inconsistency in Tables 2-5.
Citation: https://doi.org/10.5194/egusphere-2024-432-RC1 -
RC2: 'Comment on egusphere-2024-432', Anonymous Referee #2, 12 Apr 2024
A review of egusphere-2024-432 by Aditya Nalam et al.
General Comment:
As a follow-up study to Butler et al., (2020), this study aims to provide new knowledge such as long-term trends and contributions to population weighted O3 from each precursor source, and it certainly contains new findings. However, these new findings are not properly presented, and the results of the study are unsatisfactory in terms of creating useful new knowledge. The text is not well organized and is redundant in places. There are some unnecessary repetitions of the similar phrases (e.g. quotes from Edwards and Evans., 2017). English must be proofread. There are some grammar mistakes. It's better to reconsider the structure of the paper to get it published. Below are my main comments followed by specific comments, which I would like the author to use as a reference for revisions.
Major Comment:
Considering the theme of this paper, the flow of the paper would be better to first point out the main driver of long-term changes in TOB and surface O3 from the contributions from each precursor source, and then to interpret the changes in these sources from the changes in precursor emissions and/or OPE. However, the current paper presents the analysis of emissions, TOB, OPE, and surface ozone rather independently. The TOB sections (3.2.1 and 3.2.2) are focusing on the description of averaged quantities and clearly lacking detailed trend analysis. On the other hand, in the OPE section (3.3), there is an analysis of OPE trends, but it is not well discussed in relation to the trends in TOB and/or surface ozone. As a result, the new findings are presented here and there in fragments and do not form a coherent story. This is my main concern. In some places the numbers are not accurate. Authors should take care to use a consistent number of significant digits throughout the text. This seriously impairs the readability of the paper.
Specific Comments:
- Mixing the use of "RC-tagged" and "VOC-tagged" should be eliminated , if there is no particular intension.
- L11: Mid-latitude regions include Asia, so this is not an accurate statement.
- L16: Why is “methane” included in natural emission sources?
-L33: “CAM-Chem” is too technical term to be used in Short Summary.
- L40: or -> of
- L43-44: The definition of NOx and RC should be stated here.
- L93: CAM4-Chem should be spelled out as here is its first mention.
- L116 : What surface fluxes?
- Figure2: Why don't you show the comparison of ozone climatology? I don't think Taylor's diagram is the best choice.
- L191: “being” -> typo?
- L200: Table 3 -> Figure 3?
- L200: remove duplicate “period”
- L216: I don’t think the contribution of Biomass burning (12%) is “very small”
- L231-233: Are there any influences of fixed surface methane concentration for the decrease of CH4 lifetime?
- L237: Only NMVOC emission? RC emission here should be the sum of NMVOC and CO.
- L240: NOx emission -> RC emission
- L244-245: I’m not sure if this definition of TOB is correct. In this definition, if surface O3 exceeds 150 ppb, will that grid not be included in TOB calculations?
- L247-249: Are the years of these TOB estimates the same as the years covered by this study? If not, the time periods covered in the different estimates should be provided here.
- Sections 3.2.1 and 3.2.2 : As stated in Major Comments, several values in these sections are not consistent with the corresponding values in Table 2 to 5. If the values in the tables are cited in the text, they should be quoted accurately.
- L298: Fig4 -> Fig3
- L298-299: The smaller trend of TOB in % than that of NOx emission is apparent for regional contribution, but not in the case for the other contributions.
- L319 and the other places such as Table 4 and 5: In this study methane “emission” is not explicitly considered, so the description about methane “emission” in the text (and also in the Table) should be more scientifically accurate.
- L355-356: These OPE values are inconsistent with Fig 6c. Is the value of East Asia correct here?
- L379: NOy should be defined as here is its first mention.
- L446-448, L488: Why are you quoting relative trends (in %/yr) here? How can you state that these contributions are the main contributors? Are there any qualitative criteria to pick up main contributors? (and those criteria is for relative trend?)
- L454-456, L459-460: Why is the contribution of ship emissions to population-weighted surface O3 large? Could you spend more words for this?
Citation: https://doi.org/10.5194/egusphere-2024-432-RC2 -
RC3: 'Comment on egusphere-2024-432', Michael Prather, 18 Apr 2024
This manuscript is the second within a month that I have been asked to review by ACP editors, which is centered on the use of tagged O3 tracers as way of attributing the impact of different emission sources of NOx and VOCs (called reactive carbon, RC, here). For the first ms, see (1). As for (1), I find that the concept of tagged O3 tracers is inherently incorrect and leads to false attributions. Given my questioning the fundamental validity of this work, it is necessary to sign this review as I did for (1). If wiser heads believe my interpretation unfounded then let the Editor decide.
Tagged tracers for key species with major chemical feedbacks give false results. I hope we can agree that using tagged tracers for CH4 sources fails to include the well established feedbacks and thus underestimates the attributable CH4 perturbation by 40%. This feedback has been standard since the 1995 IPCC SAR and a later heuristic figure is given in (2). For O3 we have lived since Roelofs & Lelieveld with the use of O3S as a tagged tracer of stratospheric ozone that has been implemented in all the MIPs and continues to show that O3S/O3 = 30%-40%, thus attributing 30%+ of tropospheric O3 to the stratosphere.
We have long known that O3 also has feedbacks on it chemical production and loss, but have not pursued it. Last year, (3) used chemical modeling of the ATom profiling of the Atlantic and Pacific Ocean basins to demonstrate that positive O3 perturbations reduced the production of odd-oxygen and thus accelerated its own destruction. That feedback, i.e., dln(P-Ox)/dlnO3 = -0.3 to -0.4, for the ATM data underestimates the sensitivity because it was based on a 24-hour calculation. Over the lifetime of the O3 perturbation (>10 days) the added O3 will accelerate loss of NOx to HNO3 (increased NO2:NO, more OH) and reduce P-Ox further. This year, (4) pursued this effect with direct additions of ‘stratospheric' O3 and quantified the perturbation lifetime (~25 days) and watched the decay (also ~25 days). This implies that the impact of stratospheric O3 on tropospheric is about 8%, not 30%. The tagged tracer fails to alter the baseline O3 as it should. Simply, tropospheric O3 is highly buffered.
I believe that we need to recognize that major gases, like CH4, O3, and N2O, have chemical feedbacks that alter their perturbation lifetimes (change in burden / change in emissions); and we should stop pursuing simple linear models to attribute their impacts.
Michael Prather
UC Irvine
In addition to the problem of tagged O3 tracers, there are other issues that make this manuscript not ready for publication. Some specific detailed comments.
L11: I think this is only northern mid-latitudes?
L22: 'consistent with previous work' Yes, indeed, this was shown 20+ years ago in (5), it is old-hat
L23: this throwaway causal statement about convection and lifetime is probably incorrect. more likely it is just sunlight (lower latitude)
L56: ibid, I doubt that convection in the subtropics is that much more than mid-latitudes over the continents in summer. I really feel that you have no justification for these attributive statements.
L70: again, see (5).
L77: here we greatly disagree about what tagging methods do.
L96: I do not see how this is quantitative.
L112: the extension tagging to surface O3 is even more questionable
L118ff: The model resolution is a bit coarse, especially for modeling pollution NOx, but for that even 1x1 is not enough. What is worrisome is the nudging since that changes the residual circulation and the convection. What is done for convection, since the authors invoke that often in explaining low-latitude O3 production.
L125: Aviation NOx has only 3 altitude? I would think it important to include the full spread of cruise altitudes.
L129: you say "interpolate" but I hope you mean "integrate" as you would want to ensure that the totals are conserved.
L134: This is truly odd, why not just use the observed zonal mean CH4 surface data from NOAA? The idea of using a flux inversion is potentially dangerous and certainly at odds with typical AerChemMIP protocols. (Thus, comparisons become more difficult).
L154: why are you truncating the computer output to get ~1% errors? Is that necessary?
L184: why exclude aviation NOx from surface ozone?
L231: the methane lifetime is very ‘hot’ here, about 6.5 yr, at the bootm end of any of the CMIP models. It is not clear whether this includes stratospheric loss and soil loss. Given the extreme value here, it would be useful to explain how you calculate it. Also, Figure 4 is hard to understand, did you give the correct legend here?
L255: Here is where the tagging shows a clear difference from O3 perturbation experiments.
L330ff: The Ozone Production Efficiency is calculated incorrectly and obviously has the wrong units. As the authors note, most researchers have move to calculating the net production of O3 (Tg-O3/y) per emissions of NOx (Tg-N/y), for a units of Tg-O3/Tg-N, or better in moles. The units given in line 340 are missing the 'time' unit.
L356: The original reference here for more tropically oriented production per NOx emission is Wild 2001 (5). And the reason is not due to the smaller "availability" of NOx, but simply to sunlight, temepratures, and water vapor.
References
(1) https://egusphere.copernicus.org/preprints/2024/egusphere-2024-324/#discussion
(2) Prather, M.J. (2007) Lifetimes and time-scales in atmospheric chemistry, Phil. Trans. R. Soc., A 365: 1705–1726, https://doi.org/10.1098/rsta.2007.2040. FIGURE 2
(3) Prather, M. J., Guo, Hao and Zhu, Xin (2023) Deconstruction of tropospheric chemical reactivity using aircraft measurements: the ATom data, Earth Syst. Sci. Data, 15, 3299–3349, doi: 10.5194/essd-15-3299-2023. TABLE 2 & SECTION 6.1
(4) Prather, M.J. and Xin Zhu (2024) Lifetimes and timescales of tropospheric ozone, Elementa: Science of the Anthropocene, 12 (1): 00112, doi: 10.1525/elementa.2023.00112
(5) Wild, O., M. J. Prather, H. Akimoto (2001) Indirect long-term global cooling from NOx emissions, Geophys. Res. Lett., 28, 1719-1722, doi: 10.1029/2000GL012573. FIGURE 3
Citation: https://doi.org/10.5194/egusphere-2024-432-RC3 - CC1: 'Comment on egusphere-2024-432', Owen Cooper, 23 Apr 2024
- AC1: 'Response to reviewer comments on egusphere-2024-432', Aditya Nalam, 19 Oct 2024
-
AC2: 'Correction related to the calculation of CH4 lifetime in egusphere-2024-432', Aditya Nalam, 28 Oct 2024
We now recognize that our method of calculating the tropospheric CH4 lifetime was wrong.
We therefore made corrections to our manuscript text and figures corresponding to the discussion related to the tropospheric CH4 lifetime.
Please take a look at the attached document for the corrected method and the modified text and figure.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 650 | 158 | 35 | 843 | 36 | 19 | 25 |
- HTML: 650
- PDF: 158
- XML: 35
- Total: 843
- Supplement: 36
- BibTeX: 19
- EndNote: 25
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1