Gao et al present a study of ozone loss near the surface of an outdoor mesocosm sea ice facility in Winnipeg, Canada. The sea ice facility is unique and provides an opportunity to control the formation of the sea ice with complete exposure to the atmosphere, allowing ambient snow accumulation. The comparison of O3 levels during daytime for the UV-transmitting vs UV-blocking tubes is useful. The authors attribute O3 loss at 10 cm above the sea ice surface (observed for the UV-transmitting but not UV-blocking, showing that this is a photochemical process) to reaction with bromine radicals based on their enhancement of seawater Br-, which then migrated into the snow above. My major comment is the lack of discussion of the role of NOx, which is known to be important in O3 and bromine chemistry, and should be particularly important in this urban ambient study (see detailed comments below). This is particularly important because snow photochemical reactions produce NO, which can react with O3, and so disentangling this from reaction with Br is important to consider. Additionally, there is additional published literature, stated below, that is relevant and should be considered in this manuscript.

The role of NOx needs to be discussed in this study, as there is currently no mention of NOx in the paper. NO reacts with O3, and it seems possible that this could be contributing in part to the O3 loss observed over snow (see, for example, Peterson and Honrath 2001, Geophys. Res. Lett., “Observations of rapid photochemical destruction of ozone in snowpack interstitial air”), as NOx is released from the snowpack from snow nitrite and nitrate photolysis. Peterson and Honrath (2001) calculated the fraction of the observed ozone loss rate that could be attributed to NOx (they found it was small for their study and then considered bromine reaction), and this seems important to consider here. Were snow nitrate and nitrite measured? If so, this seems important to report. Further, in the urban environment of this study, NOx is likely elevated due to combustion emissions, especially since the authors discuss O3 formation (involving hydrocarbons and NOx) from vehicle exhaust.

Another aspect for which NOx is important is that BrONO2 production dominates over HOBr at NO2 > ~100 ppt (depending on HO2 as well) (Wang and Pratt 2017, JGR), which is surely the case for this urban study. This should be added to Figure 1 (the role of BrONO2 in molecular halogen production is also discussed by Wang and Pratt (2017)) and included in the introduction at Lines 44-48.

On Lines 114-115, the authors define “boundary layer air” as “the air mass above the sea ice surface inside the tubes, whereas the air outside of the tubes is considered the “ambient air””, and these terms are then used throughout the manuscript, with the comparison between these air samples being critical to the results. While it is helpful that the authors defined this phrasing in the methods section, it was quite confusing and difficult to remember through the Results & Discussion section, as all air within the boundary layer (not just inside the tubes) would be boundary layer air. I suggest that the authors choose different phrasing that is easier to remember – for example “in-tube air” vs “ambient air”.

Some clarifications of the study conditions are needed. Please directly state in Section 2.1 that the tubes are open to the overlying air. Also, it would be useful to directly state that the sea ice exposure to the atmosphere results in the deposition of atmospheric trace gases and particles to the sea ice. From the comparison in Table 1, however, it is clear that the snow composition (for the ions reported) above the sea ice is dominated by ions from the sea ice brine, based on the comparison to nearby ‘land snow’; other ions that would be more impacted by the atmosphere and may be important (e.g. nitrate and nitrite for snow NOx production) should be reported if possible. Also, please clarify in the methods when O3 was measured where (heights and which tubes), as Section 2.2 discusses switching between sampling ports at different heights and locations with 5 min resolution, but then Section 3.2 and Figure 4 seem to show O3 measurements at different heights only occurring on different days. This needs to be very clear if vertical profiles of O3 were not measured on the same day.

It would be useful on Lines 84-86 and in Table 1 to report Br-/Cl- ratios to place this in the context of previous studies of Arctic snow Br-/Cl- and the potential to produce Br2 (Peterson et al 2019, Elementa (Figure 5 and associated text); Pratt et al 2013, Nat. Geosc.). Also, it would appear that the Br-/Na+ and Cl-/Na+ labels are reversed in Table 1; please fix.

There are several additional related manuscripts that the authors should consult and incorporate into their manuscript. Nakayama et al. (2015, Tellus, “Ozone depletion in the interstitial air of the seasonal snowpack in northern Japan”) is a useful paper for the authors to compare their results to, as they also observed photochemical O3 loss outside of the Arctic. Helmig et al (2012, JGR, “Ozone deynamics and snow-atmosphere exchanges during ozone depletion events at Barrow, Alaska”) is also likely useful, particularly to discuss ozone loss near the surface due to deposition (a topic that should be discussed in this manuscript, but so far is not). Additional important laboratory saline ice studies to cite (especially on Line 62 when referring to previous frozen halogen release studies; consider whether these are helpful for interpreting results as well) include: Adams et al (2002, Atmos. Chem. Phys., “Uptake and reaction of HOBr on frozen and dry NaCl/NaBr surfaces between 253 and 233K”), Huff and Abbatt (2000, J. Phys. Chem. A, “Gas-phase Br2 production in heterogeneous reactions of Cl2, HOCl, and BrCl with halide-ice surfaces”), Wren et al (2013, Atmos. Chem. Phys., “Photochemical chlorine and bromine activation from artificial saline snow”), Halfacre et al (2019, Atmos. Chem. Phys., “pH-dependent production of molecular chlorine, bromine, and iodine from frozen saline surfaces”).

Additional Comments:

Lines 44-47: Note that Pratt et al (2013, Nat. Geosc.) showed that the initiation step of condensed-phase snowpack photochemical production does not require HOBr. Rather HOBr and BrONO2 participate in the bromine explosion cycle that propagates the bromine chemistry. This should be clarified here, as R1 does not represent an ‘initiation step’ as stated. Dark reaction of O3 with Br- has also been proposed as an initiation step (Artiglia et al. 2017, Nat. Comm.; Simpson et al. 2018, Geophys. Res. Lett.).

Lines 50-52 and 290-291: Note that Pratt et al (2013, Nat. Geosc.) showed directly that Br2 was not produced from sea ice or brine icicles (frost flower proxies) and showed that Br2 production was related to acidity, as supported by lab studies.

Lines 52-54: Burd et al. (2017, J. Geophys. Res., “Snowmelt onset hinder bromine monoxide heterogeneous recycling in the Arctic”) would be useful to cite here.

Lines 81-84: Did the prepared synthetic seawater contain carbonate/bicarbonate? It appears that it did not. Regardless, this should be stated, as it is important for understanding the pH of the sea ice surface, based on the pH dependence of molecular halogen production and the work of Wren and Donaldson (2012, Atmos. Chem. Phys., “How does deposition of gas phase species affect pH at frozen salty interfaces?”) that showed that the sea ice surface is buffered against pH change.

Lines 192-193 and Table 1: Error should be reported with 1 significant figure throughout the manuscript.

Lines 199-201: Please state the absolute magnitude of this [O3] difference here in the text and compare to that observed for the ambient vs boundary layer air (Lines 191-193 and Figure 3).

Figure 3: Since snow cover is key in this study, can shading be added to this time series to indicate when snow was present?

Line 228: When was the pH of fresh snow measured, and how/where was this snow obtained?

Gao et al. demonstrate that a mesocosm sea-ice facility can be utilized to study bromine explosion events (BEEs) and resultant ozone depletion events (ODEs). The study is to my knowledge an unprecedented demonstration of the utility and potential of mesocosms to study chemistry above sea ice. In particular, the authors are able to compare and contrast bromine and ozone inside two acrylic cylinders one UV-transmitting the other UV-blocking. The authors attribute differences in O₃ at 10 cm above the ice surface to a photochemical process being key. Changes in ice and snow conditions and temperature which are largely naturally driven complement this finding by further demonstrating a likely importance of snow and cold temperatures. Where the work can be most improved is that since the work serves primarily as a demonstration of the mesocosm facility and mesocosm experiments generally, certain limitations of this demonstration are inadequately assessed and discussed. In particular the authors note that mercury depletion events (MDE) which typically accompany BEEs can be studied, but this is not yet demonstrated. Further whether the facility replicates organic chemistry and biology which might be relevant to BEEs is not assessed or discussed.

Given that the authors point specifically at using the facility to investigate MDEs in future but have not done so, it would be useful to have some general outline of how such an experiment could be conducted to demonstrate that the mesocosm facility is useful in this regard. This can be very general as much as indicating whether Hg would be a controlled or free variable in such an experiment with some limited detail. Without this limited detail it is difficult to assess if the facility can be used for this purpose as contended.

There are additional variables which might be relevant to BEE which is it unclear if the facility has replicated or can replicate. One is organic carbon content and composition, especially organic carbon present at the surface. Studies have found that the presence of organics can have complex effects on bromide oxidation (Edebeli et al., 2019). In addition, a dark process leading to significant organic bromine from ice and snow in the Antarctic appears to operate under conditions similar to those investigated (Abrahamsson et al., 2018). Can the authors comment on any constraint on the possible effects of organic carbon and whether these are variables the facility can control?

A further potential contributing factor are biota. In particular these are likely contributors to ice-nucleation (Irish et al., 2017; Ickes et al., 2020) which can be relevant to the freezing of the simulated sea-water and potentially to the formation of ice particles above the simulated sea or sea-ice surface. Given the identified importance of snow the latter would appear to be especially relevant. Can the authors comment on the microbiome of the microcosm and the capability of the facility to simulate the sea in this regard?

Regarding this latter point,

I have the following specific comments:

L14: First encountered here but found throughout. I would not use the term “mesocosm scale” or “mesocosm-scale”. I am not an expert in mesocosm experiments but I understand them to span several decades of scale (perhaps more) and do not associate them with any inherent scale. In most if not all instances the word “mesocosm” as a noun or adjective would communicate the authors’ intent as far as I can tell.

Figure 1: Realizing this is meant to be a simplified schematic, reactions reducing Hg^II are not necessarily limited to photolyses, e.g. modeling indicates reduction of HgBrO by CO may be significant (Khiri et al., 2020). Consider adding dashed lines similar to the current photolyses.

L45-47: It is very likely that Br^- oxidation is the source of volatilized bromine in BEE, however, the central role of HOBr is less certain. In particular, some plausible alternative oxidants include HOCl (Kumar and Margerum, 1987), HOI (Holmes et al., 2001), and H₂O₂ (Bray and Livingston, 1928).

L84-86:  The authors note that they have increased the bromide content of the simulated sea water by a factor of roughly eight to enhance the resulting effects and signals. To paraphrase, one function of the mesocosm is to elucidate fundamental processes. It is not known whether BEE are linear with respect to sea water bromide or have some other relation. As such the ability to vary bromide across different mesocosm experiments or even over the course of one experiment would be valuable. The total mass of bromide required, mixing time of the pool, and/or replacement time for the simulated sea water would be helpful in this regard. Can the authors provide this information or point to a relevant reference?

L119: “at every minute” could the authors clarify what is meant by this? Is O₃ measured as an integration over each minute, or is it a shorter sample taken once each minute.

L171: “isothermal” here should be “isotherm”

L193: This is the first instance but more follow. Significances for the one-way ANOVA should not be reported as 0.00. They should be reported with the determined significance at higher precision or else as p<0.01.

Fig. 4 c: Certain periods on 2/13 and 2/15 appear to show less difference between the acrylic cylinders, from Fig. 3 these appear to have less shortwave radiation as well. This would seem to be a significant supporting argument for the importance of UV which is not commented on. Could the authors comment?

Fig. 5 and discussion: The authors comment primarily on pH, however they should also consider the effects of pH on the availability to relevant oxidants and the redox reaction. For instance the pKa of HOBr is 8.59 and HOBr concentrations would change significantly for the pH conditions in this plot.

Table 1: The columns for Br^-/Na⁺ and Cl^-/Na⁺ appear to be reversed.

L249: The one way ANOVA demonstrates that the UV-transmitting cylinder has lower O₃ on average, but not that it is consistently so. See the comment on Fig. 4 on how the consistency is not clear at certain times.

L261: I believe this refers to Fig. 4c not Fig. 6c

References:

Abrahamsson, K., Granfors, A., Ahnoff, M., Cuevas, C. A. and Saiz-Lopez, A.: Organic bromine compounds produced in sea ice in Antarctic winter, Nat. Commun., 9(1), 1–8, doi:10.1038/s41467-018-07062-8, 2018.

Bray, W. C. and Livingston, R. S.: The rate of oxidation of hydrogen peroxide by bromine and its relation to the catalytic decomposition of hydrogen peroxide in a bromine-bromide solution, J. Am. Chem. Soc., 50(6), 1654–1665, doi:10.1021/ja01393a020, 1928.

Edebeli, J., Ammann, M. and Bartels-Rausch, T.: Microphysics of the aqueous bulk counters the water activity driven rate acceleration of bromide oxidation by ozone from 289-245 K, Environ. Sci. Process. Impacts, 21(1), 63–73, doi:10.1039/c8em00417j, 2019.

Holmes, N. S., Adams, J. W. and Crowley, J. N.: Uptake and reaction of HOI and IONO2 on frozen and dry NaCL/NaBR surfaces and H2SO4, Phys. Chem. Chem. Phys., 3(9), 1679–1687, doi:10.1039/b100247n, 2001.

Ickes, L., Porter, G. C. E., Wagner, R., Adams, M. P., Bierbauer, S., Bertram, A. K., Bilde, M., Christiansen, S., Ekman, A. M. L., Gorokhova, E., Höhler, K., Kiselev, A. A., Leck, C., Möhler, O., Murray, B. J., Schiebel, T., Ullrich, R. and Salter, M. E.: The ice-nucleating activity of Arctic sea surface microlayer samples and marine algal cultures, Atmos. Chem. Phys., 20(18), 11089–11117, doi:10.5194/acp-20-11089-2020, 2020.

Irish, V. E., Elizondo, P., Chen, J., Choul, C., Charette, J., Lizotte, M., Ladino, L. A., Wilson, T. W., Gosselin, M., Murray, B. J., Polishchuk, E., Abbatt, J. P. D., Miller, L. A. and Bertram, A. K.: Ice-nucleating particles in Canadian Arctic sea-surface microlayer and bulk seawater, Atmos. Chem. Phys., 17(17), 10583–10595, doi:10.5194/acp-17-10583-2017, 2017.

Khiri, D., Louis, F., Černušák, I. and Dibble, T. S.: BrHgO• + CO: Analogue of OH + CO and Reduction Path for Hg(II) in the Atmosphere, ACS Earth Sp. Chem., 4(10), 1777–1784, doi:10.1021/acsearthspacechem.0c00171, 2020.

Kumar, K. and Margerum, D. W.: Kinetics and Mechanism of General-Acid-Assisted Oxidation of Bromide by Hypochlorite and Hypochlorous Acid, Inorg. Chem., 26(16), 2706–2711, doi:10.1021/ic00263a030, 1987.