Review of “Challenges in Simulating Ozone Depletion Events in the Arctic Boundary Layer: A case study using ECHAM/MESSy for Spring 2019/20” by Falk et al.

This is a modelling study of Arctic surface ozone depletion events due to bromine explosions – an important process that greatly impacts Arctic ozone in the boundary layer in the springtime. The authors use the EMAC model with bromine-ozone chemistry, testing out different model settings for several key parameters, such as the source of sea ice age/thickness, and critical temperature for bromine emissions. The authors determine the sensitivity to these parameters, including regional differences (based on several different sites located across the Arctic) and make recommendations for best model settings for accurate simulations of springtime Arctic tropospheric ozone. They then evaluate against TROPOMI BrO measurements, MOSAIC campaign measurements, and station monitoring measurements of surface O3 and find that the best version of their model simulates ODEs though could use improvement, as there weren’t as many or as prolonged ODEs in their simulations as in the observations.

This is a very good study, and all of my comments below are very minor.

Line by line comments:

Line 67: there is an “e.g.” that isn’t followed by anything.

Line 104: In the supplement files I downloaded, I did not see anything labelled “Supplement B”.

Fig 1 and Section 2.1: Peroxyacetyl nitrate (PAN) is a source of NOx in the Arctic from long-range transport -- Does your model contain PAN and its decomposition to NOx?

Line 239-241 and 243-245: To support the text comparing modelled and satellite-measured BrO, can you please either add the TROPOMI VCD of BrO to Figure 6 so that we can see how it compares side by side? Or at the least, you should reference Fig 3(b) in the text here, so that the reader is pointed to where to look to compare.

Line 261: showed a better what?

Fig 6: text font is too small in the panels.

Sec 4, first paragraph: Section 4.3 is referenced before Section 4.2. Text should be re-ordered to the flow of the paper.

Line 268-269: This sentence seems very similar to that at line 260-261. Does it really need repeating here?

Line 280: Should “In 2000,” be “In 2020,” here?

Line 286: “raging within the Arctic cycle in 2019” should that be Arctic *circle*?

Fig 7 and Fig B1: “The periods March–May are highlighted in linen” ß linen is the light yellow colour? As ‘linen’ isn’t typically a word used for a colour (in my experience), maybe better to say ‘light yellow’ here.

Line 294-297: As you mention 2007 as the transition period, how come you chose 2000 as the cut off for the before and after time periods? Wouldn’t the results after 2000 still have the pre-2007 conditions included in the average?

Line 301: “The tails become larger throughout April and March” – do you mean *April and May* since you already mentioned March in the prior sentence?

Fig 9b: The MOSAIC dots look like they fall into vertical lines (e.g. on May 15 the dots span about 2-10 pptv. Is this because the uncertainty on those measurements is quite high or because it has a large diurnal cycle? What would the error bar be for each dot?

Line 373: might not *be* the solution

Sec 5: The Summary and Conclusion section is currently a little light on what you did in this study, and quite heavy on what others have done in the context of your future work. To shorten the latter and reduce duplication with what was already discussed in Section 4, you could simply list the future work items (snow, iodine chemistry, dry deposition). And for the former, for example, you could summarize the steps you took to improve the simulations of ODEs (e.g. sea ice and critical temperature).

Fig A1: Instead of “middle”, “right” and “left”, can you please include labels for each panel? (e.g. “(i) for upper left, (ii) for upper middle, etc). It is somewhat difficult to determine what is what from the current caption. Please also increase font size of the panel titles.

Fig A3: Similarly, It is not clear here that Br2 is the left and BrCl is the right, since it’s not mentioned except in extremely small font on the colour scale label. Please add additional panel labels, and increase the font size in this figure.

Fig A4: the text in this figure is unreadably small.

[Summary]

This study by Falk et al. reports ECHAM/MESSy model experiments for simulating bromine explosions and ozone depletions in the Arctic boundary layer for the springs of 2019 and 2020. The model was evaluated and trained in its empirical parameter choices by using routine observational data of surface ozone and satellite BrO retrievals from 2019 and was re-evaluated using equivalent data from 2020 and also against shipboard data over sea ice in the central Arctic from the MOSAiC expedition in 2020. It demonstrates the capability and limitation of the simplified treatment of gaseous bromine sources from snow- and ice-covered surfaces proposed originally by Toyota et al. (2011) and adopted later by ECHAM/MESSy and a few other models. The current, adapted version of the Toyota scheme implementation in ECHAM/MESSy shows some overall improvement in the model performance, but it also shows unresolved discrepancies against observed ozone and BrO temporal and spatial variations, suggesting a need for more advanced approaches to capture more precisely the entire processes of bromine sources from polar snow and ice covers, transport and recycling of bromine in the atmosphere and their impact on ozone photochemical destruction. In addition, decadal trends in the occurrence of surface ozone depletion events from coastal station data are analyzed and compared with the model results for 2019 and 2020. This decadal trend information from observations implies that the source strength of bromine in the polar region is not controlled primarily by the presence of first-year sea ice, contrary to what the Toyota scheme, and the science community in general, normally assumes. The spatial distribution patterns in modeled and observed BrO VCDs also indicate that the surface source of bromine on multi-year sea ice is underrepresented in the model. Overall, it is a useful case study to help efforts in the modelling community towards improved capability for simulating polar boundary-layer bromine explosions and surface ozone depletions. As I will explain below (major comment #1), however, I have one reservation about the choice of supposedly best performing model parameters from 2019 to conduct the model run for 2020 (sfa017). I also have a fair number of technical and editorial suggestions for the manuscript revisions.

[Major comments]

1. Since sfa011 (T_crit = -10 deg C) shows the best overall performance from evaluation against the 2019 surface ozone data from four Arctic coastal stations, the authors have decided to use -10 deg C as T_crit for modelling the 2020 case (sfa017). However, model run sfa012 (T_crit = -2.5 deg C) significantly improves the 2019 surface ozone performance at Zeppelin in terms of R² and RMSE (Table A1) as compared to other model runs assuming lower values of T_crit. Since R/V Polarstern (MOSAiC expedition) drifted at locations relatively close to Zeppelin during the spring of 2020, it is probably fair to ask if the model can perform much better using T_crit = -2.5 deg C for capturing observed surface ozone and BrO variabilities at the MOSAiC ship locations in 2020 (section 4.2). For example, had the surface ozone been more rapidly deleted in the model by the time air masses arrive at the MOSAiC ship locations due to stronger bromine sources arising from the use of T_crit = -2.5 deg C, simulated BrO concentrations could have been lower at the ship locations and agreed better with measurements from the ship (because the photochemical steady state between Br and BrO under sunlight calls for the presence of ozone to maintain BrO). Relationships among surface bromine source strengths, surface ozone and BrO concentrations are more complex than linear responses because of feedbacks such as those indicated above (e.g., Zhao et al., 2016, section 3.4.2; Hausman and Platt, 1994). This possibility should be discussed. Please also read my minor comment on Lines 257-259 concerning the likelihood of bromine explosion occurrences at temperatures as high as close to -2.5 deg C. Nevertheless, the authors have raised valid points about the limitation of the Toyota scheme. I do not claim that simplicity in the Toyota scheme is adequate for very precisely capturing the physical and chemical processes pertaining to the behavior of bromine across the air-snowpack-seaice-seawater system.
2. Figure annotations can and should be made more reader friendly. Table 2 lists all the model experiment numbers/IDs with specific differences among them, but figures and their captions sometimes lack information from which model experiment(s) the presented graphics have been generated. For example, I would add the model experiment IDs (sfa010, sfa011 and sfa012?) on the lefthand side of maps in Figure 6 and/or in the figure caption. Similarly, I believe the authors can add model experiment IDs in Figure 5 or its figure caption. I would also add the model experiment IDs on the lefthand side of scatter plots in Figure A4. Fonts may need to be larger, if possible, to make the annotations legible especially when printed on paper. I had difficulty reading some annotations in Figures 5, 6, A1, A3, and A4 (well, this last one was the most problematic for me; I needed to magnify, magnify and magnify the PDF to be able to read the annotations at last).
3. Lines 133-136: This paragraph may want to be expanded and detailed a little more. A sentence that briefly describes the relationship between dry deposition velocity and surface resistance can be helpful. More importantly, there is no explicit description for the value of surface resistance assumed on ice/snow surfaces for ozone in the present model runs, but it is discussed later as an important factor for the simulated behavior of surface ozone by referring to what sounds like a different value (10000 s/m) used/tested in the previous ECHAM/MESSy study (Lines 279-280 and Lines 407-410). Please add a statement in this section on the value of surface resistance for ozone on ice and snow surfaces assumed in the present study.
4. Line 178 and Figure 4: It will be nice if the authors can demonstrate this statement visually by a scatter plot for temperatures vs. their modelled biases with dots colored by BL height.
5. Lines 280-281: Is it worthwhile referring to the possibility of the Arctic ozone hole in 2020 for having created distinct lower tropospheric photochemical activities in March? The Toyota scheme as implemented in several models to date is unlikely to capture the influence of increased UV irradiance on bromine emissions from the ice surface. Also, although only remotely related to the present study, Steinbrecht et al. (2020) noted a minor impact attributable to 2020 stratospheric ozone depletions in the decrease of ozone from the free troposphere in the CAMS reanalysis data (in which the boundary-layer bromine explosion chemistry was likely neglected). For the model performance weakness in May 2020, I wonder again if T_crit = -2.5 deg C (not carried out for the year 2020 model run) could have solved part of the problem.

[Minor comments]

Line 36: Was Coburn et al. (2016) a study on Hg chemistry in the polar boundary layer? If not, I suggest citing other references. Brooks et al. (2006) may be a good fit here.

Line 48: Magnesium, calcium and potassium are not necessarily negligible cations in natural seawater and are potentially more important than sodium for the maintenance of liquid brines on salty ice surfaces at very low temperatures (e.g., Koop et al., 2000).  I suggest the authors revise the statement here from “sodium bromide (NaBr)” to “bromide (Br-)”.

Line 61: Consider citing Adams et al. (2002) and Fickert et al. (1999) in addition to Oldridge and Abbatt (2011). However, the Oldridge and Abbatt study did not really focus on reactions (R6-R8) but examined other physicochemical aspects of bromide oxidation on salty ice surfaces such as the role of acidity and temperature-dependent bromine formation using ozone as an oxidant. In other words, the Oldridge and Abbatt study should be cited more in the context of uncertainties in the process-level understanding. You might want to consider citing Wren et al. (2010) as well along this line.

Lines 61-62: While Sander et al. (2006) used a box model in their study of reactions mentioned here, Toyota et al. (2014) used a one-dimensional model in their study. Aqueous-phase radical reactions rather than Reactions (R6-R8) were suggested to be also important in the deeper snow layers.

Line 72: Can you be more specific about what was indicated as a source of bromide in Ridley et al. (2003).

Line 96: It will be helpful to indicate the height scales in meters represented by 1-3 lowest model levels.

Lines 99-100: “Heterogeneous reactions are mainly restricted to the stratosphere” – does it mean that the heterogeneous reactions in the model are mainly for simulating those taking place in the stratosphere, or in the troposphere?  This is an important point that must be clarified because heterogenous reactions on aerosols are among the key factors that control the activity of reactive bromine chemistry in the troposphere including in the polar boundary layer (Fan and Jacob, 1992).

Line 153: Do you interpolate the model output vertically as well?  It could matter for Mt. Zeppelin (474 m ASL), which perhaps resides above the lowest model level, disconnected often from immediate influences from surface emissions near the station both in the real environment and in the model.

Lines 205-206: The lack of data coverage by ICDC AoSI automatically should result in the assignment of sea ice age to be less than 1 year in the Canadian archipelago region. Can this be a source of artifact?

Lines 209-210: It will be informative if you can explicitly state the range of values given by Regan et al. (2023).

Table 2 and Lines 255-256: The meaning of “frozen lakes masked” in Table 2 is unclear, so is the meaning of “these are excluded” on Lines 255-256. Do they mean that frozen lakes were neglected from air-surface chemical interactions in corresponding model runs?

Line 257-259: Gong et al. (2025) assumed T_crit close to -2.5 deg C in their model (GEM-MACH) at 15-km grid resolution and found a reasonable surface ozone performance during the spring of 2015 at Zeppelin. As such, we may also argue that the improved surface ozone performance by using T_crit = -2.5 deg C in ECHAM/MESSy is not necessarily an artifact from inadequate model resolutions and that bromine explosions near Zeppelin may indeed persist at relatively high temperatures as compared to other ice-covered regions for unidentified reasons.

Line 278 (and Figure 7): I am confused about this statement. In Figure 7c, I do see “… in March … a pronounced dip in ozone at Summit” from model time series for March 2020, not from observed time series. In other words, contrary to the authors’ statement, I see false decrease in simulated surface ozone from March 2020 not observed at Summit. Here is a side note from me out of this matter: the choice of colors for time series from different model runs and observations could be revised in Figure 7 to make the observed time series more distinct from the model time series.

Lines 279-280 and Lines 409-410: Helmig et al. (2007) was the first modelling study in which essentially the same conclusion was reached. Cite this study here.

Line 284 and Line 378: Consider citing Steinbrecht et al. (2020) and Weber et al. (2020) as well here.

Line 285-290: Do oil sector emissions from Prudhoe Bay play a role here as well in the model overprediction of surface ozone in the summer at Utqiagvik?

Lines 329-330: I do not see visual information related to the statement “tropospheric BrO VCD remains lower than observed” at least in Figure 9.  Can this information be added to Figure 9?  

Line 334: How high is the albedo assumed on sea ice for computing J values in the model runs?  If it indeed seems to be too low as an assumption, the authors may want to cite some studies to support their speculation on the snow/ice surface albedo. I also wonder if solar zenith angles as large as 100 degrees (or even greater?) matter for tropospheric twilight chemistry. Or do the authors imply chemistry in the dark? Are there references to back up this part of speculation, too?  The range of solar zenith angles indicated here obviously contains typos (100 cannot be smaller than 80), which obscure what is being speculated here.

Line 353-357: Consider citing Fernandez et al. (2024) here.

Line 385: I find the statement “an inherent feature of chemistry and advection in the troposphere” a little too fuzzy. Do the authors want to say, “a consequence of observing different phases of non-linear temporal evolutions in the O3 and BrO concentrations in air masses being transported under the influence of chemistry that leads to BEs and ODEs in the troposphere (e.g., Hausmann and Platt, 1994)”?     

[Technical suggestions]

Line 30: observed TO DROP below

Line 49: constantly -> continually

Line 79: spring 2019 CONDITIONS

Line 81: climatological ozone depletion TRENDS

Line 88: Delete a hyphen before “chemical”

 Line 95: latitude greater than 68 degrees in both hemispheres

Figure 1: Toyota et al. (2012) -> Toyota et al. (2011)

Line 115: Cl to Br MOLAR ratio

Line 120: Henry constant -> Henry’s law constant

Line 123: Has ODS been defined?

Line 125: HCFCs = hydrochlorofluorocarbons

Line 126: I suppose NMHCs are not part of VOCs in your definition. Do VOCs then represent oxygenated VOCs like acetone?

Line 131: surface concentrations concentrations -> surface concentration climatologies

Line 134: dry deposition resistances -> dry deposition velocities

Line 136: Henry coefficient -> Henry’s law coefficient

Table 1 caption: Delete “Referred”. Data sources could include more information such as websites and the date of data download in the references.

Line 172: was -> were

Line 173: Which data did Mahajan (2022) report?

Line 177: -2 -> 2

Line 178: -10 -> 10

Line 178: This WARM bias increases with decreasing temperatures

Figure 4 caption: Temperature -> surface air temperature; BL -> BL height

Line 191: Any citable document for the release candidate of MESSy v2.56?

Table 2: MESSy model version (2.55.2) and MECCA chemistry (CCMI2-base-01) are all the same among model runs. Do they need to be presented like this in the table?  Also, what is the difference between “diagnostic” and “interactive” approaches for sea salt Br2? Can it be clarified by a more detailed description in the text?  Lastly, a note on sfa002 indicates that this run was used for the CATCH arctic bromine model intercomparison project, which becomes clearer only when we come to read acknowledgements. If the authors wish to indicate this fact about sfa002, it should probably be done more clearly within Table 2 or in Section 3.

Figure 5 caption: a model experiment -> model experiments

Line 215: Is the “potential vorticity tropopause index” a commonly used terminology? If not, I would rephrase it to a more descriptive statement like “we … used a dynamical tropopause metric based on the model potential vorticity (PV) at X (1.0, 1.6, or 2.0?) Potential Vorticity Unit (1 PVU = 10^-6 m² s^-1 K kg^-1) to separate …”

Line 224: temporally in springtime -> in time and space during the spring

Line 233: lower threshold -> higher threshold temperature

Line 237: this does not result in NOTABLY enhanced BrO VCD

Lines 237-239: From this sentence, it is not very clear what was measured in a study by Jalkanen and Manninen (1996). Did they measure chloride and bromide concentrations in particulate matter or more of their total gaseous and particulate concentrations? Referring to their sampling and detection methodologies might clear things up.

Lines 239 and 245: T_crit = -2.5

Line 260: AS STATED ABOVE, tropospheric BrO VCD showed a QUALITATIVEBLY better RESULT with

Figure 6 caption: … flux integral for April 2019 …

Line 273: contained -> confined

Line 276: late April/early May -> late April to early May

Line 277: late March/late April -> late March to late April

Line 283: emissions of these -> their emissions

Line 286: Arctic cycle -> Arctic Circle

Line 287: These -> Such wildfire emissions

Line 295: To test this HYPOTHESIS

Line 300: March -> May (?)

Line 303: following -> owing to

Lines 303-304: Please be more specific about what “this” refers to in this sentence.

Line 309: atmospheric background O3 has increased -> background O3 concentrations apparently have increased

Line 318: an underestimation of the frequency of χ_O3 becoming greater than 37 ppbV

Line 340-343: Consider citing Peterson et al. (2019) and referring to discussions therein in regard to the role of multi-year ice snowpacks as a source of gaseous bromine for the Arctic boundary layer.

Line 348: confirms the result -> strongly suggests

Line 351: as well -> again

Line 363: During these PERIORDS

Line 364: towards -> at

Line 373: might not the solution -> might not be a solution

Line 399: motivated -> oriented

Line 405: The accuracy of surface resistance for dry deposition is essential for modelling the removal of trace gases in the atmospheric boundary layer.

Line 412: an inclusion -> its inclusion

Line 424: deeper -> more precise

Figure A1 caption: Revise the second sentence to “Comparison between T42 and T106 illustrates the potential benefit of a higher model resolution for resolving inhomogeneity in the characteristics of sea ice such as its age.”

Figure A2 caption: Add “monthly” before “time series” if that is the case. I do not quite understand the meaning of the second sentence. Please rephrase. The third sentence may sound clearer if you say “MYSIC total areas estimated the two methods differ by 30%”.

Figure A3 caption: integral FOR April 2019

[References]

Adams, J. W., Holmes, N. S., and Crowley, J. N.: Uptake and reaction of HOBr on frozen and dry NaCl/NaBr surfaces between 253 and 233 K, Atmos. Chem. Phys., 2, 79–91, doi:10.5194/acp-2-79-2002, 2002.

Brooks, S. B., A. Saiz-Lopez, H. Skov, S. E. Lindberg, J. M. C. Plane, and M. E. Goodsite (2006), The mass balance of mercury in the springtime arctic environment, Geophys. Res. Lett., 33, L13812, doi:10.1029/2005GL025525.

Fan, SM., Jacob, D. Surface ozone depletion in Arctic spring sustained by bromine reactions on aerosols. Nature 359, 522–524 (1992). https://doi.org/10.1038/359522a0.

Fernandez, R. P., Berná, L., Tomazzeli, O. G., Mahajan, A. S., Li, Q., Kinnison, D. E., Wang, S., Lamarque, J.-F., Tilmes, S., Skov, H., Cuevas, C. A., and Saiz-Lopez, A.: Arctic halogens reduce ozone in the northern mid-latitudes, P. Natl. Acad. Sci. USA, 121, e2401975121, https://doi.org/10.1073/pnas.2401975121, 2024.

Fickert, S., J. W. Adams, and J. N. Crowley (1999), Activation of Br2 and BrCl via uptake of HOBr onto aqueous salt solutions, J. Geophys. Res., 104(D19), 23719–23727, doi:10.1029/1999JD900359.

Gong, W., Beagley, S. R., Toyota, K., Skov, H., Christensen, J. H., Lupu, A., Pendlebury, D., Zhang, J., Im, U., Kanaya, Y., Saiz-Lopez, A., Sommariva, R., Effertz, P., Halfacre, J. W., Jepsen, N., Kivi, R., Koenig, T. K., Müller, K., Nordstrøm, C., Petropavlovskikh, I., Shepson, P. B., Simpson, W. R., Solberg, S., Staebler, R. M., Tarasick, D. W., Van Malderen, R., and Vestenius, M.: Modelling Arctic lower-tropospheric ozone: processes controlling seasonal variations, Atmos. Chem. Phys., 25, 8355–8405, https://doi.org/10.5194/acp-25-8355-2025, 2025.

Hausmann, M., and U. Platt (1994), Spectroscopic measurement of bromine oxide and ozone in the high Arctic during Polar Sunrise Experiment 1992, J. Geophys. Res., 99(D12), 25399–25413, doi:10.1029/94JD01314.

Helmig, D., Ganzeveld, L., Butler, T., and Oltmans, S. J.: The role of ozone atmosphere-snow gas exchange on polar, boundary-layer tropospheric ozone – a review and sensitivity analysis, Atmos. Chem. Phys., 7, 15–30, https://doi.org/10.5194/acp-7-15-2007, 2007.

Koop, T., A. Kapilashrami, L. T. Molina, and M. J. Molina (2000), Phase transitions of sea-salt/water mixtures at low temperatures: Implications for ozone chemistry in the polar marine boundary layer, J. Geophys. Res., 105(D21), 26393–26402, doi:10.1029/2000JD900413.

Peterson, PK, et al. 2019. Snowpack measurements suggest role for multi-year sea ice regions in Arctic atmospheric bromine and chlorine chemistry. Elem Sci Anth, 7: 14. DOI: https://doi.org/10.1525/elementa.352

Steinbrecht, W., Kubistin, D., Plass-Dülmer, C., Davies, J., Tarasick, D. W., von der Gathen, P., et al. (2021). COVID-19 crisis reduces free tropospheric ozone across the Northern Hemisphere. Geophysical Research Letters, 48, e2020GL091987. https://doi.org/10.1029/2020GL091987

Weber, J., Shin, Y. M., Staunton Sykes, J., Archer-Nicholls, S., Abraham, N. L., & Archibald, A. T. (2020). Minimal climate impacts from short-lived climate forcers following emission reductions related to the COVID-19 pandemic. Geophysical Research Letters, 47, e2020GL090326. https://doi.org/10.1029/2020GL090326

Wren, S. N., T. F. Kahan, K. B. Jumaa, and D. J. Donaldson (2010), Spectroscopic studies of the heterogeneous reaction between O3(g) and halides at the surface of frozen salt solutions, J. Geophys. Res., 115, D16309, doi:10.1029/2010JD013929.

Zhao, X., K. Strong, C. Adams, R. Schofield, X. Yang, A. Richter, U. Friess, A.-M. Blechschmidt, and J.-H. Koo (2015), A case study of a transported bromine explosion event in the Canadian high arctic, J. Geophys. Res. Atmos., 120, doi:10.1002/2015JD023711.