the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Two-years of stratospheric chemistry perturbations from the 2019/2020 Australian wildfire smoke
Abstract. The very large pyrocumulonimbus events that occurred during the Australian summer of 2019/2020 caused extremely unusual partitioning of stratospheric chlorine in Southern Hemisphere midlatitudes and Antarctic regions. This was likely caused by enhanced HCl solubility in organic species that increased heterogeneous chemistry. Here, we show that observed HCl and ClONO2 values remain outside the pre-wildfire satellite range since 2005 in both the Southern Hemisphere midlatitude and Antarctic regions in 2021. Through model simulations, we replicate this multi-year prolonged chemical perturbation, in good agreement with observations. This was achieved by calculating HCl solubility in mixed wildfire and sulfate aerosols consistent with assumptions of 1) liquid-liquid phase separation and 2) linear dependence on organic and sulfate composition. The model simulations also suggest that the Australian pyrocumulonimbus organic aerosols contributed to low midlatitude ozone values in 2021. A marked photochemically controlled seasonality of the chemical perturbations and ozone depletion is also observed and simulated, and its underlying chemical drivers are identified. This work highlights that lower concentrations of smoke still had profound effects on stratospheric heterogeneous chemistry more than a year after the 2019/2020 wildfire event.
- Preprint
(1521 KB) - Metadata XML
-
Supplement
(487 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2948, missing info', Anonymous Referee #1, 12 Nov 2024
General Comments
The paper presents interesting and new model studies on heterogeneous chlorine activation on wildfire smoke particles. This includes different sensitivity
studies on microphysical interaction with background stratospheric organic aerosol. Unfortunately some figures can convey the message that the effects on ozone are minor in the model and somehow not consistent to the observations. Several times important information is missing.Specific Comments
Line 12: Insert 'in 2020 and 2021'
Line 103: Please mention the selected retrieval method for OMPS-LP (NASA, USask...) here since this can cause large differences (see also line 226).
Line 108: Â Is an existing transient simulation (with or without nudging) used for initialization on 29 December? This is critical for the correct distribution of halocarbons and total inorganic chlorine and bromine. Which halocarbons are included (including lumped ones)?
Line 111: Self-lofting can occur also with weak nudging but your option might be the cleaner approach.
Line 124 or earlier: Which emission inventory is used for background organics? Why is there a problem in Southern midlatitudes?
Line 165: Organics are a mixture of species with different molecular weights. How is mass here defined? Variable in space and time? Or assumptions (e.g. molecular weight of hexanoic acid)? More details please since heterogeneous chemistry depends on molecular weight.
Line 177: Anomalies here still undefined. Provide selected time reference periods for observations (MLS 2004-2019?) and model (best it should be the same number of years, e.g. 2004-2019 transient, to avoid artifacts which show up in case of a much smaller number of years).Â
Line 180: I don't understand this unexpected behaviour in 2021. It can be also due to an ill definition of anomaly.
Line 211: This depends critically on initialization. Don't speculate here.
Line 228: Isn't it dangerous to compare different time periods for estimation of anomalies of different quantities? This can introduce additional uncertainties.
Line 274: Why is there no effect on ozone in 2021 in contrast to observations?
Line 276: Expand caption: What is shown as greyshading? Variability of observations? Timeframe? I suppose from text that ACE is monthly and the other curves daily. Please mention for clarity.
Line 320 or earlier: Did the model reproduce the self-lofting smoke filled and ozone poor anticyclonic vortices mentioned in the introduction?
Line 328: 'and absence of sunlight?'
Figure 4: Years without ozone hole dominate the variability, that is somewhat distracting. I don't understand why there is no ozone depletion except for the small response in June in contrast to MLS. Dynamics cannot explain that, especially not for the nudged simulations. Please elaborate. Is there some artifact due the used anomaly method? From this figure you get the impression that heterogeneous chemistry on organics does not matter for Antarctic ozone in contrast to other studies and the conclusions. Maybe an additional conventional time series plot in the supplement with model (mostly nudged) and observations can help here.
Supplement: Fig S1: What is 'volcanic background'? Is the value for January 1 already perturbed?ÂTechnical Corrections
Fig S5: The legends for MLS and ACE have the wrong time.ÂCitation: https://doi.org/10.5194/egusphere-2024-2948-RC1 -
RC2: 'Comment on egusphere-2024-2948', Anonymous Referee #2, 22 Nov 2024
The manuscript be Stone et al. is a follow-up study based on the paper by Solomon et al. (2022) which demonstrates the heterogeneous
chemistry on organic aerosols is important for explaining the observations of chlorine compounds in the air contaminated by bushfire
exhaust of the Australian New Year (ANY) wildfires in late December 2019 and early January 2020.In the current manuscript three different model setups for the handling of HCl solubility in the aerosols are discussed.
- homogeneous mixture of background sulfate aerosol and ANY wildfire organic aerosol
- separate treatment of background sulfate aerosol and ANY wildfire organic aerosol
- liquid-liquid-phase separation (LLPS) only in the ANY aerosols
It seems that the last setup reproduces the observations best. Â This simulations involved a liquid-liquid phase separation only in the air
influenced by the ANY wildfire exhaust.The authors show model results and comparison with observations for aerosol extinction, ozone and chlorine compounds HCl, ClONO2, and
ClO. The results are only shown as anomalies and not comparison of the absolute model quantities with observations. This at least leaves some
suspicion that an absolute comparison of the shown model parameters with the observations does not look well. Therefore this absolute
comparison should be shown to (hopefully) gain confidence in the suggested parametrisations.I recommend recommend this paper for publication after this point has been clarified and also the following issues have been addressed.
Major IssuesIt seems that the third set of assumptions is the best and is discussed in the main part of the paper, while the results from the other
assumptoions are only shown only in the supplementary material. Â It was not evident to me how the presented three assumptions
with respect to HCl solubility relate to the original simulation in Solomon et al. (2022). Â Is one of them identical or are they all different?The advantage of a hybrid model setup with 2 months free running and specified dynamics (SD) afterwards is not clear. Â Was it proven, that
in SD run with nudged winds the self-lofting of the smoke plume is not present? Â On the other hand, can you show, that in the presented
simulation, this effect is well simulated? Â Lestrelin et al. (ACP, 2023) and Selitto et al. (ACP, 2023) showed the dynamics of developing
vortices from the ANY fires. This should be mentioned in the introduction and it would also be nice to see, how well the model
describes these vortices. Furthermore it would be nice to see, if the observed plume structures are reproduced by the model.
line 94ff and figs 2e,f and 4e,f: In fact, it is not recommended to use daytime minus nighttime MLS ClO measurements for polar latitudes.
ClO has a significant diurnal cycle with maxima near the local noon and typically near zero values during the night. Further, MLS has
a coverage that typically observes at very similar local times for a given latitude. Therefore an average of MLS at a certain latitude
would give the mean value for the two corresponding characteristic local times, which is different than a diurnal average calculated
by the model. Therefore, for a meaningful comparison, one should calculate model output for the given MLS observation locations
and times.Besides the sampling effect discussed above, it is not at all clear, if the very small anomaly number is realistic, given the precision and
accuracy of the MLS measurements. To me, it seems meaningless to show this ClO comparison and I would suggest leaving it out.line 226ff (and elsewhere): What do you mean exactly by "anomaly"? The difference between the values and the long running mean value?
Or its difference with the mean annual cycle?  It sounds like the latter, but please clarify that.
Minor issuesl. 93: Please explain what you mean by PressureZM
l. 94: It is not true that the use of daytime minus nighttime ClO is recommendedfor observations in high polar latitudes
l. 98: Please use ACE-FTS instead of ACE here and throughout the paper as the ACE satellite does carry other experiments as well
l. 101: explain OMPS-LP
l. 116: What are primary organic aerosols in primary organic sections? Please explain better, such that a reader may understand
the very basic principle without having to read the aerosol model description papers.l. 120: What do you mean by 1e-6Â ? Likely 10-6Â , or really e-6 ?
l. 128-131, 305-307: Please use proper arrows in chemical reactions.
l. 134: change to "HOBr"
Â
Citation: https://doi.org/10.5194/egusphere-2024-2948-RC2
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
250 | 54 | 38 | 342 | 30 | 2 | 6 |
- HTML: 250
- PDF: 54
- XML: 38
- Total: 342
- Supplement: 30
- BibTeX: 2
- EndNote: 6
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1