the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Source attribution of near-surface ozone trends in the United States during 1995–2019
Pengwei Li
Hailong Wang
Su Li
Pinya Wang
Baojie Li
Hong Liao
Abstract. Emissions of ozone (O3) precursors in the United States have decreased in recent decades, and near-surface O3 concentrations showed a significant decrease in summer but an increase in winter. In this study, an O3 source tagging technique is utilized in a chemistry-climate model to investigate the source contributions to O3 concentrations in the U.S. from various emitting sectors and regions of nitrogen oxides (NOx) and reactive carbon species during 1995–2019. We show that domestic emission reductions from energy and surface transportation are primarily responsible for the decrease in summertime O3 during 1995–2019. However, in winter the emission control also weakens the NOx titration process, resulting in considerable increases in O3 levels from natural sources. Additionally, increases in aviation and shipping activities and transpacific transport of O3 from Asia largely contribute to the winter O3 increase. Changes in large-scale circulation also explain 15 % of the O3 increasing trend.
- Preprint
(2198 KB) -
Supplement
(398 KB) - BibTeX
- EndNote
Pengwei Li et al.
Status: final response (author comments only)
-
RC1: 'Comment on acp-2022-678', Anonymous Referee #1, 13 Jan 2023
Title: Review of acp-2022-678
# Summary
This manuscript details a multi-year ozone tagged contribution analysis. The specific value is the attempt to explain observed trends with trends of model contributions. I think this manuscript has high value, but needs some additional analysis to be published. Below are sections that summarize comments related to model performance, methods, and editorial notes. Last, is a line-by-line section that has more specific feedback.
# Model Performance Evaluation
Generally, this manuscript is missing basic model evaluation in the body and supplement. The current manuscript jumps into trends of contributions and only mentions evaluation in the conclusions. In particular, only evaluation in China is ever discussed. The manuscript focuses on trends associated with titration without ever demonstrating the model reasonably captures the phenomenon. The representation of titration in the Eastern US by the model is important given that it drives the trends. Given that coarse models (2x2.5 degree) are often extremely biased at nighttime, the authors should provide some evidence that nighttime titration is reasonably simulated and/or describe how model artifacts may play a role in the trend. Overall, it seems odd that performance over China is used to suggest underestimation of long-range transport while the performance and possible errors of local contributions are omitted.
# Methods Feedback
In the methods sections, more detail is needed on several fronts. The emissions are currently under-described even though they are well referenced. Recommendations are made in the line-by-line section. Similarly, the model-observation pairing is mostly left to the reader to infer. Again, recommendations are made in the line-by-line.
# Editorial Feedback
Especially in the conclusions, there are several statements where increase/decrease seem to be used incorrectly. These incorrect directional statements should be corrected. See line-by-line section for specific recommendations.
The section describing the overall trends would benefit from a table. The descriptions and parenthetical references make it somewhat difficult to easily compare. See line-by-line section for specific recommendations.
Lastly, recent application of a very similar system was made by Butler et al. 2020. That publication is referenced, but more should be done to compare these methods and results to those. From a methods standpoint, it would be nice to provide a short summary that explains how these experiments are different. From a results standpoint, you should compare the overlapping 2010 year for comparable. Are the results comparable for overlapping (2010) or proximate years? If not, do methodological differences explain discrepancies?
# Line-by-Line
* 117, please describe the depth of the first layer and the number of layers in near the surface (e.g., under 2km). This helps contextualize the model representation of titration later.
* 121, please describe how the stratospheric values are set. Are they based on climatological values? Are they scaled based on something?
* 146, are XTR tags really neither NOx nor VOC? Are they included in both?
* 159, It says CO and CH4 are not tagged by individual sources? Does that mean just by regions? Or, all CO is lumped? The wording is currently unclear. Particularly interested for CO.
* 164, as you note, this limitation of CO seems odd.
* 168, It would be useful to note here (in addition to later) that TgN and TgC are shown in the appendix.
* 173, This seems like an important methodological shift. Can the authors highlight whether conclusions are robust to analysis from 1995-2015 or 1995-2019?
* 175, Can you clarify what "present-day" means here? Is this a climatology based on a range of "present-day" years or a specific year?
* 177, Please elaborate on Price parameterization. I think you are saying online parameterization based on simulated cloud top heights. There are also climatologies based on Price, so it is good to be clear.
* 180-185, Please clarify whether the simulation is being sampled only at observation site-days or averaged seasonally and then sampled at sites.
* 200, The results should start with some estimate of model performance over the target analysis areas. At least, 1) a map of the model with obs scattered on it for an early year and a late year and 2) a description of how basic performance stats change over time. Because this paper focuses on the JJA and DJF, I would expect the model performance to have a similar separation. This will help the readers contextualize results.
* 205, Please add lightning NOx in the supplemental figures.
* 214, related to 180-185, are these trends based on the model only at observation sites or based on averages of the regional "box"
* 214: Looking at Figure 4b, there is a lot of heterogeneity in the western summer trends. The Western cities are fairly isolated leading to misrepresentation by coarse global models. Can you discuss what would happen if you only looked at CASTNet or rural monitors? Or just the IPCC sites?
* 216-223, I found it difficult to keep the text organized in my mind. I recommend adding a table here.Table X: Trends by season (DJF, JJA) for observations and the model sampled at observation sites (right?).
Season,Source,East,West
DJF,Obs,2.1±0.29,2.2±0.23
DJF,Mod,6.1±0.40,3.2±0.28
JJA,Obs,-3.0±0.41,-0.5±0.42
JJA,Mod,-3.0±0.29,-2.3±0.20It would also be good to add some clarify on what "well produce" means Based on a 95% certainty, the CI are not overlapping for Eastern winter or Western summer. The CIs for Western winter are barely overlapping and only after rounding. The model seems to clearly reproduce the trend only for Eastern summer.
For me, the titration performance in the East raises questions about the West. The model seems to dramatically overestimate the reduced titration in the East. Given the population density of the East, the titration is likely more widely spread. Due to the population sparsity of the West, the overestimated titration is likely diluted. How does this impact the conclusion about well representing winter in the West?
* 236, I am surprised to see STR (stratosphere) in both NOx and VOC. Is that via XTR?
* 241, can you add error bars to the figure?
* 243-247, I think this is a very interesting finding! If the atmosphere is increasingly NOx sensitive, that should have important implications for VOC tagging in later years. Can you discuss that a bit more?
* 247-251, What role does the location of monitors play in the conclusion here? Is there a strong spatial gradient to the SHP contribution? This is important because the populations tend to be skewed toward near the ocean. In an ideal world, it would be interesting to see a few maps (1995 and 2019) of contributions trends that have a strong spatial gradient.
* 259, I find this to be a particularly interesting finding that has implications for the estimation of climate/air quality co-benefit assessments. I wish it was expanded a bit in the conclusions.
* 272, I find this confusing. Most of this sentence makes perfect sense to me. It is introduced, however, in the context of reduced VOCs. At aircraft heights, you say that only NOx increase. Does that mean that there are no aircraft VOCs? If so, are you suggesting that VOCs at 6-10km were meaningfully reduced and that contributed to the large aircraft trends?
* 280-282, Similar comment to earlier. The spatial nature of this enhancement is important. It'd be great to see a map of the contribution and trends.
* 283, I don't think this is strictly speaking true. Your lightning emissions are parameterized based on simulated clouds. Can you clarify that this is only true for VOC?
* 289, Butler et al compared January and July in Figure 5. It isn't clear to me that the contribution maximized in DJF vs MAM.
* 295, Specify anthropogenic and/or that you are excluding soil NOx. Soil NOx in summer has a large anthropogenic component and the contribution from soil is likely "domestic" (e.g. Lapina et al. 2014).
* 316, This is definitely interesting... I'm struck however by the dramatic overestimate in the trend, which might be related to the models representation of vertical mixing in winter.
* 326-327, The idea that South Asia, and Southeast Asia East Asia "equally contribute" is a somewhat surprising finding. Many previous refereed articles show a decreased transport efficiency from India (S. Asia) to the US compared to East Asia. Similarly, Butler et al 2020 showed significantly larger East Asia contribution than South Asia. Can you highlight why your results would be so different?
* 335-350, This discussion really highlights the oversimplicity of linear trends. The authors do a good job noting this is likely associated with transport. It would be good if they connected it to known meteorological cycles. A quick look shows that ENSO cycles are likely accounted for by the time averaging, but the Pacific Decadal Oscillation (PDO) is not. This highlights why the trend is likely not significant. It is likely made up of ups and downs seen in the PDO. A 5-year average of the NCEI PDO index shows that the winters of these two periods are of opposite signs (despite inter annual variability). This is in part because in mid 1998, the PDO index shifted. This leads to a smaller difference between summers than winters. You could also reference the Lin et al. paper about the position of the jet stream.
* 342, I think it is a mistake to call the comparison of two five year periods "anomalous".
* 355-357, I think you got the signs of change wrong here. You showed decreasing in the summer (controls) and increasing in the winter (lessening titration).
* 355-357, You showed that it could only replicate the decreasing trend in the eastern summer and the increasing trend in the western winter. You showed that the trends for Eastern winter and western summer were *not* well captured. The trends were significantly different. So, it is wrong to say that it did well in the conclusion.
* 359-361, You need to be clear when you are talking about the model regions and when you are talking about the observed sites as sampled by the model. Are these trends at select sites? Are these trends representative of the larger region? Or the population weighted concentrations?
* 364, This is a little less clear to me. The VOCs were also reduced. How do you distinguish between reduced VOC trends and reduced NOx OPE impacts on VOC trends.
* 391, This was a 3.7 ppb/decade decrease (not increase.
* 392-393, The authors are offering only one of several equally permissible explanations. As you note, one is that the Asian trends are underestimated. Another is that the coarse model overestimates the decrease associated with domestic reductions. Another is that stratospheric variability is underestimated. Another is that the trend in ships contributions are underestimated. What makes the authors confident that this is the only hypothesis to highlight?
* 393-396, 1) The authors should show this model performance if their conclusions rely upon it. 2) The local peaks in China will depend on near surface vertical structure while the continental scale outflow may not. So, you could only says that it "consistent with the low contribution from Asian sources" since you cannot say that it definitively explains anything.
* 396-397, It is unreasonable to think that model evaluation of China is worth discussing, while model evaluation over the US is not.Citation: https://doi.org/10.5194/acp-2022-678-RC1 -
AC1: 'Reply on RC1', Pengwei Li, 08 Mar 2023
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2022-678/acp-2022-678-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Pengwei Li, 08 Mar 2023
-
RC2: 'Comment on acp-2022-678', Anonymous Referee #2, 17 Jan 2023
The manuscript of Li et al. investigates trends of ground-level over the US for 1995 – 2019. They use a source attribution method (tagging) to attribute the ozone trends to trends of emissions of VOCs and NOx from different sectors and regions.
The topic of the manuscript is very interesting and it fits into the scope of ACP. However, the manuscript needs major revisions before it can be reconsidered for ACP.
General remarks:
1) The authors report that during winter time ozone increases due to NOx titration. They use, however, a rather coarse resolved global model. It is well known that at this resolution NOx titration is usually underestimated by the model. Therefore, I agree with referee #1 that a more detailed model evaluation is needed. The authors should also discuss how the ability of the model to capture the observed trends only partly (e.g. increasing trend in EUS in winter is strongly overestimated) influence the conclusions. Further, it would be interesting to analyse if the model is able to capture the chemical regimes in EUS and WUS correctly.
2) Even though the CEDS emissions are well documented by Hosely et al., 2018 the authors should discuss these emissions in more detail as the results of the study heavily depend on the emission inventory.
How do the trends in the emissions of Hosley et al., 2018 for example compare to the trends in McDuffie et al. 2020, the CAMS or the EDGAR emissions? What is the influence of the inconsistency in the aviation emissions in CEDS (Thor et al., 2022) on the results?
3) The labels and fonts in many figures are too small. All of the labels/fonts (and also the station symbols in Fig 2.) needs to be enlarged.
4) The model description misses a lot of basic information. Even though the CAM4-chem model is well known, most important information should be given in the manuscript. As example detailed information about the chemical mechanism is missing. How is dry- and wet deposition represented? What variables are nudged? Further I am missing information about the emission totals for natural emissions (lightning NOx, biogenic VOCs and soil-NOx) as well as the global emissions. All of these information are important to compare results from different studies with each other. Of course not all of these things need to be discussed in detail. Some are also fine in the supplement (for example detailed information about the emissions).
5 ) The authors need to clarify the units they use. They use ppb as unit which sounds like (volume/mass?) mixing ratios but use the term concentration throughout the manuscript. Please clarify if you consider mixing ratios or concentrations. Also for the emissions totals the authors should specify what NOx and NMVOCs are (see below for more details).
6) In the model simulations CH4 mixing ratios are kept fix at 1750 ppb. This represents ~ 1990 levels (https://gml.noaa.gov/ccgg/trends_ch4/). Until 2019 CH4 levels have been increased to ~ 1880 ppb which is a increase of ~7--8 %. This increase influences ozone production and Butler et al., 2020 found very inhomogeneous changes of the contributions by CH4 increases. Therefore, I suggest to perform an additional simulation with an CH4 increase.
7) The authors should reconsider the choice of tagging labels. In my opinion the region North America should have been spitted into US and Canada. Further, important information are lost because the shipping emissions have the ROW tag in the “region tagging” runs. Why are they not tagged as shipping/oceanic in the “regional tagging” runs? Further, I wonder why the results of many “unchanged” sectors changed between the “sector” and “region” tagging runs. Shouldn’t the results for the tags “STR”, “LGT”, “AIR”, “SOIL” be identical in Fig 5. and 7? If only anthropogenic sources get either sector or region tags the results of the natural sources should not change? In my view this is a very critical inconsistency which needs to be clarified (maybe I also don’t understand the approach correctly).
In addition, the authors should motivate the special category for CO in more detail. Many information are lost by lumping all CO emissions in one category.
8) The manuscripts lacks a detailed discussion of the results in comparison to other global source attribution studies. Are the results in accordance with other studies? For example are results of O3 from STR, lightning or biogenic sources comparable with other studies? Much more comparison with previous work is needed (see below for some references; in addition also Guo et al. 2017 .and all the studies from the HTAP framework (https://htap.org/) can be of interest here). In my opinion also the introduction needs to be improved (see detailed comments below).
9) I am wondering about the trend of O3 due to aircraft emissions. Usually most aircraft emissions take place in the (upper) troposphere and not near ground-level. Therefore I wonder if there is a trend of O3 from aviation (check values in the upper troposphere) or if there I an increase in downward transport. If so, it would be interesting to separate effects due to increased emissions and due to changes in dynamics.
Detailed comments:
p4l70-p4l87: This section needs some corrections. The perturbation approach and labeling techniques are two different methods answering different scientific questions. The perturbation approach provides (potential) impacts. Tagging provides contributions. There is many literature discussing this issue which can be checked for more details– some are: Grewe et al. 2010, Emmons et al. 2012, Clappier et al. 2017 and Tunis et al., 2020.
p4l91: This is not correct. There are approaches applied on the regional scale which use chemical indicators (Dunker et al., 2012, Kwok et al., 2015). However, there are also approaches on the regional scale which do not use chemical indicators (e.g. LupaÅcu and Butler, 2019; Mertens et al. 2020)
P5l97: I think this heavily depends on how the boundary conditions are implemented (see literature above).
P5l100ff: There are also global approaches which use a sector wise attribution or a combination of sector wise and regional attribution (Emmons et al. 2012, Grewe et al. 2017, Butler et al. 2018). Butler et al. 2018 includes a comprehensive overview of different approaches which the authors could check.
P6L120: Is O3 is nudged at the tropopause towards ‘stratospheric’ values? What happens with the stratosphere tagged. tracer?
P7l152ff: See also general comments above. Why not tagging shipping emissions separately as Ocean (see Butler et al., 2020). What about aviation in this list. If I understand the analysis correctly aviation has been tagged as sector in the regional runs?
P7l164: See also general remarks. Is this the explanation why CO is lumped?
p7l168: What about emissions of SO2 and NH3?
P8l177: Please specify the lightning NOx total emissions?
P9l208-p9l225: See also general remarks. For summer EUS many stations show an increase of O3. It seems that this is not captured by the model? For winter EUS stations some stations show no or even a decreasing trend. These trends are not captured by the model. Please comment.
Fig 3 : Why do VOC emissions of the ENE sector increase while NOx emissions decrease? Please comment and compare the trend of ENE with other emission inventories. Please specify if emissions are Tg(N), TG(NO) etc. (also for VOCs)
P10l235: Fig 4 and 5 should be reordered; same for 6 and 7.
P10l246: A more detailed analysis of the change of the O3 production efficency would be very valuable here.
P10l247: See also general remarks: Information about the trends of global emissions (e.g. shipping etc.) would be very valuable here.
P10L257ff: Could you please explain the argumentation here in more detailed? I think additional analysis would help here to make this point more clear.
P11l278ff: This sentence is very long. I suggest to split it up.
P12l295: See general remarks. Why did you applied one combined tag for North America and not separate tags for US and Canada?
P13l318: I don’t understand the sentence. Please explain. Thanks!
P13L334: The figure suggest that none of the results are significant (by the way; with which method did you check for significance. Please explain in detail were appropriate).
Section 3.4: See general comments above. More details/analysis are missing here. In my opinion especially an analysis of the contributions would be very valuable here. How large is for example the trend of ozone from the stratosphere due to changes in the dynamics.
P14l355: I don’t agree with the statement that the model captures the trends ‘well’. Please rephrase.
P15l374ff: What is the reason for the the increase of shipping and aviation – emissions or dynamics? Please analyse in more detail.
References:
Butler, T., Lupascu, A., Coates, J., and Zhu, S.: TOAST 1.0: Tropospheric Ozone Attribution of Sources with Tagging for CESM 1.2.2, Geosci. Model Dev., 11, 2825–2840, https://doi.org/10.5194/gmd-11-2825-2018, 2018.
Butler, T., Lupascu, A., and Nalam, A.: Attribution of ground-level ozone to anthropogenic and natural sources of nitrogen oxides and reactive carbon in a global chemical transport model, Atmos. Chem. Phys., 20, 10707–10731, https://doi.org/10.5194/acp-20-10707-2020, 2020.
Clappier, A., Belis, C. A., Pernigotti, D., and Thunis, P.: Source apportionment and sensitivity analysis: two methodologies with two different purposes, Geosci. Model Dev., 10, 4245–4256, https://doi.org/10.5194/gmd-10-4245-2017, 2017.
Emmons, L. K., Hess, P. G., Lamarque, J.-F., and Pfister, G. G.: Tagged ozone mechanism for MOZART-4, CAM-chem and other chemical transport models, Geosci. Model Dev., 5, 1531–1542, https://doi.org/10.5194/gmd-5-1531-2012, 2012.
Grewe, V., Tsati, E., and Hoor, P.: On the attribution of contributions of atmospheric trace gases to emissions in atmospheric model applications, Geosci. Model Dev., 3, 487–499, https://doi.org/10.5194/gmd-3-487-2010, 2010.
Grewe, V., Tsati, E., Mertens, M., Frömming, C., and Jöckel, P.: Contribution of emissions to concentrations: the TAGGING 1.0 submodel based on the Modular Earth Submodel System (MESSy 2.52), Geosci. Model Dev., 10, 2615–2633, https://doi.org/10.5194/gmd-10-2615-2017, 2017.
Guo, Y., Liu, J., Mauzerall, D. L., Li, X., Horowitz, L. W., Tao, W., and Tao, S.: Long-Lived Species Enhance Summertime Attribution of North American Ozone to Upwind Sources, Environ. Sci. Technol., 51, 5017–5025, https://doi.org/10.1021/acs.est.6b05664, 2017.
LupaÅcu, A. and Butler, T.: Source attribution of European surface O3 using a tagged O3 mechanism, Atmos. Chem. Phys., 19, 14535–14558, https://doi.org/10.5194/acp-19-14535-2019, 2019.
McDuffie, E. E., Smith, S. J., O'Rourke, P., Tibrewal, K., Venkataraman, C., Marais, E. A., Zheng, B., Crippa, M., Brauer, M., and Martin, R. V.: A global anthropogenic emission inventory of atmospheric pollutants from sector- and fuel-specific sources (1970–2017): an application of the Community Emissions Data System (CEDS), Earth Syst. Sci. Data, 12, 3413–3442, https://doi.org/10.5194/essd-12-3413-2020, 2020.
Mertens, M., Kerkweg, A., Grewe, V., Jöckel, P., and Sausen, R.: Attributing ozone and its precursors to land transport emissions in Europe and Germany, Atmos. Chem. Phys., 20, 7843–7873, https://doi.org/10.5194/acp-20-7843-2020, 2020.
Thor, R. N., Mertens, M., Matthes, S., Righi, M., Hendricks, J., Brinkop, S., Graf, P., Grewe, V., Jöckel, P., and Smith, S.: An inconsistency in aviation emissions between CMIP5 and CMIP6 and the implications for short-lived species and their radiative forcing, Geosci. Model Dev. Discuss. [preprint], https://doi.org/10.5194/gmd-2022-250, in review, 2022.
Thunis, P., Clappier, A., Pirovano, G.: Source apportionment to support air quality management practices, A fitness-for-purpose guide (V 3.1), EUR30263, Publications Office of the European Union, 2020, ISBN 978-92-76-19744-7, doi:10.2760/47145, JRC120764
Citation: https://doi.org/10.5194/acp-2022-678-RC2 -
AC2: 'Reply on RC2', Pengwei Li, 08 Mar 2023
The comment was uploaded in the form of a supplement: https://acp.copernicus.org/preprints/acp-2022-678/acp-2022-678-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Pengwei Li, 08 Mar 2023
Pengwei Li et al.
Pengwei Li et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 417 | 121 | 10 | 548 | 36 | 3 | 8 |
- HTML: 417
- PDF: 121
- XML: 10
- Total: 548
- Supplement: 36
- BibTeX: 3
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1