the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Jan 2025
Modelling Arctic Lower Tropospheric Ozone: processes controlling seasonal variations
Abstract. Previous assessments on modelling Arctic tropospheric ozone (O3) have shown that most atmospheric models continue to experience difficulties in simulating tropospheric O3 in the Arctic, particularly in capturing the seasonal variations at coastal sites, primarily attributed to the lack of representation of surface bromine chemistry in the Arctic. In this study, two independent chemical transport models (CTMs), DEHM (Danish Eulerian Hemispheric Model) and GEM-MACH (Global Environmental Multi-scale – Modelling Air quality and Chemistry), were used to simulate Arctic lower tropospheric O3 for the year 2015 at considerably higher horizontal resolutions (25-km and 15-km, respectively) than the large-scale models in the previous assessments. Both models include bromine chemistry and a representation of snow-sourced bromine mechanism: a blowing-snow bromine source mechanism in DEHM and a snowpack bromine source mechanism in GEM-MACH. Model results were compared with a suite of observations in the Arctic, including hourly observations from surface sites and mobile platforms (buoys and ship) and ozonesonde profiles, to evaluate models’ ability to simulate Arctic lower tropospheric O3, particularly in capturing the seasonal variations and the key processes controlling these variations.
The study found that both models behave quite similarly outside the spring period and are able to capture the observed overall surface O3 seasonal cycle and synoptic scale variabilities, as well as the O3 vertical profiles in the Arctic. GEM-MACH (with the snowpack bromine source mechanism) was able to simulate most of the observed springtime Ozone Depletion Events (ODEs) at the coastal and buoy sites well, while DEHM (with the blowing-snow bromine source mechanism) simulated much fewer ODEs. The study showed that the springtime O3 depletion process plays a central role in driving the surface O3 seasonal cycle in Central Arctic, and that the bromine-mediated ODEs, while occurring most notably within the lowest few hundred metres of air above the Arctic Ocean, can induce a 5–7 % of loss in the total pan-Arctic tropospheric O3 burden during springtime. The model simulations also showed an overall enhancement in the pan-Arctic O3 concentration due to northern boreal wildfire emissions in summer 2015; the enhancement is more significant at higher altitudes. Higher O3 excess ratios (ΔO3/ΔCO) found aloft compared to near the surface indicate greater photochemical O3 production efficiency at higher altitudes in fire-impacted air masses. The model simulations further indicated an enhancement in NOy in the Arctic due to wildfires; a large portion of NOy produced from the wildfire emissions is found in the form of PAN that is transported to the Arctic, particularly at higher altitudes, potentially contributing to O3 production there.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Atmospheric Chemistry and Physics. The authors have no other competing interests to declare.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(8066 KB) - Metadata XML
-
Supplement
(4271 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-3750', Anonymous Referee #1, 13 Feb 2025
Review of “Modelling Arctic Lower Tropospheric Ozone: processes controlling seasonal variations” by W. Gong et al.
This manuscript describes a detailed evaluation of Arctic tropospheric ozone in two regional chemical transport models using a variety of surface and vertical profile measurements for the year 2015. The study undertakes a thorough observational comparison, followed by a detailed investigation of the role of halogen chemistry in controlling model ozone, an analysis of wildfire emission contributions, and finally a regional Arctic tropospheric ozone budget analysis from one of the two models.
Among other aspects, the detailed comparison of high resolution ozone simulations with hourly data is an important advance on many previous studies that have evaluated coarse global-scale models with monthly mean observations. The presentation of comparison of model simulations that include detailed halogen chemistry against simulations with this removed is also very informative, as are the investigations of sensitivity to assumptions in the model Br mechanisms. In addition, evaluation of the structure of modelled and observed ozone vertical profiles using aircraft observations during springtime is of great benefit. These results are of high value to the Arctic atmospheric composition and modelling communities.
Generally, I do not have any major concerns or reservations with the paper. The analysis presented is very thorough, and the manuscript is well written. I recommend that the manuscript is suitable for publication in ACP once the following comments have been addressed.
General comments
The paper is very long, however I recognise that the analysis presented is very thorough. I have made one suggestion of where text could be shortened by avoiding separation of parts of the results that could be better linked (see below).
In a couple of places more could be done to compare the performance of the models presented here with previous model assessments of ozone using the same datasets. I have highlighted a couple of examples below in my specific comments.
Specific comments
Line 61: Archibald et al., (2020) is not a primary reference for the impacts of ozone on health and ecosystems. Can the authors provide alternative references for these two aspects of ozone impact.?
Line 65: “changes [..].. in the transport pattern from lower latitudes” This needs to be more explicit to provide context. i.e. “changes in the patterns of transport of ozone and precursors from lower latitudes ”.
Line 88: “The ability of current models to simulate Arctic tropospheric O3 has been evaluated in several studies (e.g., Monks et al., 2015b; Shindell et al., 2008; Whaley et al., 2023)” What is meant by “current” in this context (given that citations from 2015 are relevant here)? I agree with the need to cite some of these older studies, since it is not clear that the models have improved substantially in this time. Maybe omit “current” and rephrase as “… evaluated in several previous and recent studies…”
Line 190-193: Does this imply that within the European domain ECLIPSE emissions are not used (replaced by EMEP)? What is the motivation for this? Is it simply more information from higher resolution? How different are the emissions?
Table 1: It would be useful to add to this table the temporal resolution of the data measured and/or used in the study. Could this perhaps be added into column 3? Similarly for sondes, what is the approximate vertical resolution of the data?
Section 3.1 and Fig. 2 discussion. There is no mention of the low ozone simulated in both models over the northern Eurasian region during winter. This is also evident in Figure 3, which highlights the winter months as being the time of the minimum at the surface. Is this the impact of ozone titration by Eurasian NO emissions in winter?
Figure 4 - Would it be possible to add a legend to the figure labelling the coloured lines used?
Line 557: This text describes the statistical evaluations for the comparisons shown in Fig. 4. I am not sure this needs to be separated from the presentation of performance of the models in the previous paragraph. The text could be combined to reference the statistics as part of the discussion of model performance. This would also help qualify several subjective terms such as “compare well” (e.g. line 551).
Lone 575: The authors make the statement that the comparisons shown demonstrate improved model performance compared with similar evaluations using global models. Would it be possible to be more quantitative, given that previous studies have used the same surface sites for evaluation and will have quoted e.g. mean bias values (notwithstanding the use of different tome resolution data)?
Section 3.1: The Whaley et al., (2023) study presented evaluation of a set of global models against ozone sonde data (Figure 8 in their paper). It would be informative to make some sort of reference / comparison to this in putting the results presented by the authors into context.
Figure 10 - A minor point, but maybe it is worth spelling out “interquartile range” (IQR) in the legend or caption.
Figure 11 - It might help in comparison of the different sensitivity simulations to provide some quantitative metrics for the comparisons with observations (i.e. mean bias / r2 values).
Line 955 - The Arnold et al., (2015) evaluation of fire-impacted O3/CO enhancement ratios are also based on monthly mean large-scale Arctic enhancements, so these could be more directly compared with results presented here (i.e. they are also not plume specific enhancements).
Page 47: Discussion of PAN/CO enhancement ratios. In the Arnold et al., (2015) study, a difference in PAN/CO enhancement values was identified between models forced using different reanalyses products (models forced using GEOS-5 data displayed lower enhancements compared with models forced by ERA-Interim data). It would interesting to know how the models presented compare and if they are consistent with the Arnold et al., (2015) values according to the meteorological dataset used (for DEHM using ERA-5 for example).
Editorial / typographical corrections
Line 79: “variations in the Arctic tropospheric O3” Omit “the”.
Line 458: Better as “…varying degrees of complexity..”
References
Arnold, S. R., et al., Biomass burning influence on high-latitude tropospheric ozone and reactive nitrogen in summer 2008: a multi-model analysis based on POLMIP simulations, Atmos Chem Phys, 15, 6047–6068, https://doi.org/10.5194/acp-15-6047-2015, 2015.
Whaley, C. H., et al., (2023), Arctic tropospheric ozone: assessment of current knowledge and model performance, Atmos. Chem. Phys., 23, 637-661, https://doi.org/10.5194/acp-23-637-2023.
Citation: https://doi.org/10.5194/egusphere-2024-3750-RC1 -
RC2: 'Comment on egusphere-2024-3750', Anonymous Referee #2, 10 Mar 2025
This manuscript describes a modelling study of Arctic tropospheric ozone using two chemical transport models (DEHM and GEM-MACH), each parameterized with a different bromine source mechanism (blowing snow and snowpack, respectively). Notably the regional models were run at high spatial resolution (~20 km). Sensitivity tests were performed for bromine chemistry and boreal wildfire emissions. This is an important modelling study that includes comparisons to a large number of observational datasets - not only ground-based monitoring sites, but also buoys, mobile platforms, and vertical profiles. I expect this is likely the most comprehensive high resolution model-measurement intercomparison thus far. The manuscript is comprehensive and well-written, with clear figures that enhance understanding. The authors do an excellent job of comprehensively citing literature and discussing their results throughout the paper. Specific comments below mainly focus on improving clarity.
DEHM Methods: How was blowing snow itself parameterized? I did not see that described in the methods. Chen et al. 2022 (ACP, https://doi.org/10.5194/acp-22-15263-2022) showed that the commonly used parameterization based on wind speed and temperature overpredicts blowing snow conditions, likely due to the lack of inclusion of a snow age term.
Given the stated goal to investigate key processes driving surface O3 seasonal cycles, combustion NOx emissions have recently been shown to have regional impacts on bromine recycling and ozone, as described by Peterson et al. 2025 (Faraday Discussions, https://doi.org/10.1039/D4FD00166D) and Widmaier et al. 2025 (Faraday Discussions, https://doi.org/10.1039/D4FD00166D). Since this wasn’t investigated in the current study, the authors are encouraged to at least add a sentence in the conclusions pointing to this as a suggested future direction.
Additional comments:
L99-100: Another pertinent paper to cite is Peterson et al. 2017 (ACP, https://doi.org/10.5194/acp-17-7567-2017).
L109 & L568: An earlier important paper to also cite is Raso et al. 2017 (PNAS, https://www.pnas.org/doi/10.1073/pnas.1702803114).
L323 & Appendix 1 table: Add Jeong et al. 2022 (ACS Earth Space Chem, https://doi.org/10.1021/acsearthspacechem.2c00189), in addition to Burd et al.
Table 1: I encourage writing “(Utqiagvik)” after “BARC” (under MAX-DOAS section) for improved clarity.
Pages 20-21: There is a very long paragraph that extends ~1.5 pages. I encourage breaking it up to make it easier to read.
Figure 4: While the colors are described in the caption, I encourage adding a legend as well for the three lines.
Figure 5: This figure was confusing at first, as I initially couldn’t figure out why the gray traces were different between the plots. To improve clarity, I suggesting removing the header “O-Buoy & Mirai” and replacing with “Observations”, “GEM-MACH”, and “DEHM” above each plot, to emphasize that this is the comparison being shown.
Figure 6: It would be helpful if the gray background behind the entirety of the plots could be removed. It would also be helpful if font sizes could be increased. Both of these edits should improve readability. Define the month abbreviations in the caption.
L694: Moore et al. 2014 (Nature, https://doi.org/10.1038/nature12924) is an important paper to cite here for convection-based springtime ozone recovery.
L696-699: This speculation can be supported by prior observations of reactive bromine in Utqiagvik in February by Custard et al. 2017 (ACS Earth Space, https://doi.org/10.1021/acsearthspacechem.7b00014) and Simpson et al. 2018 (GRL, https://doi.org/10.1029/2018GL079444). For context, polar sunrise occurs at Utqiagvik in late January.
L715: It would seem that the improved simulation of the ODEs only at the Zeppelin site is associated with the site being above the inversion layer, as opposed to the other sites near sea level. This seems worth noting.
Figure 8: It would be helpful to increase the font sizes, especially the y axis labels, to make them more readable. Also, the wind vector symbols are very small and difficult to discern.
L753: I suggest citing Oltmans et al 2012 (JGR, https://doi.org/10.1029/2011JD016889) and Seabrook et al 2011 (JGR, https://doi.org/10.1029/2011JD016335) as excellent examples of vertical ozone profiles over extended field campaign periods.
Figure 9: Please increase the date font size in the center map figure, as well as the outer legend font size.
Figures 12, 15, 16, 18: The upper left label, above the colorscale, on each plot is not readable. Please reformat to increase font size, or remove if the label is the same as the top header on the figure. For Figure 18, the larger font (right corner labels) currently do not include units.
Figure 17: It would be helpful to increase linewidths of the font throughout (or make them bold) to make the text easier to read.
L1105: Simpson et al 2017 (ACP, https://doi.org/10.5194/acp-17-9291-2017) is an important paper to cite here, as it points to the importance of aerosols and needed chemical composition measurements, since the presence of aerosols alone did not the reactive bromine.
Citation: https://doi.org/10.5194/egusphere-2024-3750-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 158 | 34 | 9 | 201 | 27 | 6 | 6 |
- HTML: 158
- PDF: 34
- XML: 9
- Total: 201
- Supplement: 27
- BibTeX: 6
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 77 | 37 |
| Canada | 2 | 32 | 15 |
| United Kingdom | 3 | 12 | 5 |
| Germany | 4 | 10 | 4 |
| France | 5 | 10 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 77