the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Extratropical cyclones drive the spatial heterogeneity distribution of Sea Salt Aerosol (SSA) and vertical transport in the Southern Ocean
Abstract. Sea salt aerosols (SSA) were emitted via bubble bursting during wave breaking, exhibiting a strong positive correlation with wind speed. However, the generation and emission of SSA driven by cyclones was still lack of knowledge. In this study, we combine cruise observations and GEOS-Chem simulations to investigate the contribution of extratropical cyclones to SSA dynamics. During the R/V Xuelong cruise, observed SSA concentrations were consistently lower in cyclonic periods compared to non-cyclonic periods, a pattern probably linked to updraft transport within cyclone systems. Model results revealed that SSA concentrates predominantly north of cyclone centers. As altitude increases, these high-concentration zones gradually shifted northwestward. Cyclone-associated high-wind regions accounted for 63 % of total SSA emissions across the Southern Ocean. The maximum upward SSA transport flux occurred at 450 m altitude within Warm Conveyor Belt regions, with stronger and longer-lasting cyclones generating greater transport intensities. Our results demonstrate that cyclones modulate SSA distribution primarily through turbulent mixing, with synergistic effects from wet deposition and advective transport. This study highlights the spatial heterogeneity of SSA distribution during cyclones and elucidates how combined physicochemical processes regulate SSA transport under cyclonic conditions.
- Preprint
(1104 KB) - Metadata XML
-
Supplement
(774 KB) - BibTeX
- EndNote
Status: open (until 21 Oct 2025)
- RC1: 'Comment on egusphere-2025-3987', Anonymous Referee #1, 23 Sep 2025 reply
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
1,270 | 16 | 8 | 1,294 | 8 | 7 | 4 |
- HTML: 1,270
- PDF: 16
- XML: 8
- Total: 1,294
- Supplement: 8
- BibTeX: 7
- EndNote: 4
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
Review of “Extratropical cyclones drive the spatial heterogeneity distribution of Sea Salt Aerosol (SSA) and vertical transport in the Southern Ocean” by Xiaoke Zhang et al. (egusphere-2025-3987)
This work investigates the role of cyclones in generating and regulating sea salt aerosol (SSA) in the Southern Ocean. First, measurements of SSA and meteorological conditions from a cruise are presented. Then, simulations with GEOS-Chem and composite analysis with a catalog of cyclones are used to understand the dynamics and drivers of SSA during cyclonic periods. A quantification of how much each underlying process contributes to the SSA budget over the SO is then derived. SSA is a key species for climate and undergoing strong changes. At the same time, the dynamics of cyclones are also evolving over the Southern Ocean. The topic of this paper is therefore of interest to the atmospheric community.
General comments
As stated in the specific comments below, some of the arguments that are made are not well presented and/or relatively weak, and many details about the figures are missing to actually understand what they represent. Additionally, except for a comparison of SSA concentrations no link is made between model and observation, which would largely benefit the quality of the manuscript. At this stage, I think the paper would reach the same conclusions without including the measurements. Furthermore, there are times where too much detail is provided (e.g. the authors explain what AOD is, the description of the instrument measuring [SSA] is 12 lines long, part of the model analysis concludes that higher winds generate more SSA in the model while it is clearly expected since it is how it is parameterized…), while at the same time the reasoning leading to key results is not clearly explained which make them sound more like extrapolations sometimes. Overall, a major work of clarification needs to be done to make the arguments clearer, stronger and improve the quality of this paper for it to be fit for publication in ACP.
Specific comments
L11: “strong positive correlation with wind speed” - you say line 147 that the correlation is 0.33, lower than in other studies. I would not call this a strong correlation.
L80: I would not say “majority” of the SO. Looking at figure S1 you actually cover the SO between 170E and 109W, which is a significant portion but clearly not the majority.
L106: Is a 2°x2.5° horizontal resolution okay for appropriately reproducing cyclones and/or observations from the ship cruise?
L119: “The data” - which data? The cyclone tracks? The GEOS-Chem fields? Both?
L125: “4,000 km × 4,000 km grid points were generated at the identical locations of the original cyclones for all time steps” - I do not understand this sentence and thus the methodology. Please rephrase/clarify.
L147: “lower than values reported in previous studies” - which studies? missing reference.
L147–148: “At the same time, less than the 980 hPa pressure level accounted for 46.11 % of observations” - okay but according to Fig S1 as soon as the ship was south of 60°S the surface pressure was almost exclusively below 980hPa, which is related to higher latitudes having lower pressures on average, not necessarily to cyclone occurence. Low pressure anomalies would be a better indicator of the potential role of cyclones, not absolute low pressures.
L151: “The research vessel was considered influenced by a cyclone when the distance to its center was less than the cyclone's radius” - as you go on to show in the next parts of the manuscript, depending on where you are with respect to the cyclone center (e.g. north vs south), the effects of an ETC on [SSA] differ. In this part you are inferring the influence of cyclones on observed [SSA] on the ship only by looking at whether the ship is within reach of the cyclone so this is incomplete information. I think you would gain a lot in clarity and robustness by merging sections 3.1 and 3.2, analyzing Figure 2 first and then look at which quadrant of each cyclone the ship was located in for Figure 1. Otherwise the link/motivation from section 3.1 to section 3.2 seems a bit artificial.
L158–159: Figure 1 indicates that not all the cyclones you consider are indeed associated with higher wind speeds. Some do, but some do not (e.g. 3, 4 and 7 have average to low wind speed). Therefore your argument line 159—161 on the updraft explaining the lower [Na+] is not very strong. Furthermore, there is intense precipitation at the ship location either during the cyclones (2,6,8 as you mention but also 5 to some extent) or right before (1 and 7, which you do not mention). Precipitation before the event can contribute to cleaning the air, and therefore concentrations take time to recover which can explain why they are lower. The only two remaining events are 3 and 4. Cyclone 4 shows well below average wind speeds, and it seems that (3) is probably misplaced by ERA5 since the ship data shows a dip in pressure about a day earlier than the gray shade in Figure 1. At the same time as the actual dip before (3) there is precipitation, so wet scavenging. Overall then it seems to me that in Figure 1 precipitation and wind speeds are sufficient to explain why [Na+] is lower during cyclones without needing the argument about updraft. This reinforces my previous point that the structure of the manuscript should be changed.
L159—16: “This result indicates that during cyclonic events, SSA undergoes vertical transport to higher atmospheric levels via converging updrafts in the lower-level cyclone centers” - I do not think you have clearly showed this. At this point this is merely a hypothesis.
L173—175: correlation only is not enough to characterize model skill, at least bias and RMSE should also be looked at. All of these numbers should ideally be compared to those from other GEOS-Chem studies on sea salt aerosol to actually conclude that the model performs well. In other words, is 0.46 actually a good enough correlation? Could you validate other variables from GEOS-Chem using the cruise measurements, e.g. wind speed and precipitation to see if the differences in Fig S3 could be partly explained by biases in the modeled meteorology?
L175—178: Visually it seems that GEOS-Chem struggles to reproduce observed [Na+] even more during cyclonic periods. I think this can be a major limitation to this study since you want to understand what happens to sea-salt during these specific periods. Furthermore, some analysis of whether GEOS-Chem/MERRA-2 reproduces the same cyclones as ERA5 is needed to support the validity of the results as you cross two different data sources.
Fig S3: what is the meaning of the blue shade? Could you overlay the cyclone periods on this plot?
Figure 2: the compositing technique needs to be further explained. The way I understood it, since it is cyclone-centered, these maps do not correspond to an actual geographical region, rather a “cyclone-space”. But then you overlay the vessel track which makes me really confused as to what is represented in this figure. Also, there is no mention of how many cyclones are part of this composite. This is essential information for the reader to assess the statistical significance of this composite analysis. Overall, there is more explaining to be done here.
L199–202: it seems that you use only model data to reach this conclusion, so it is expected from the start since the modelled SSA emission is parameterized based on wind speed. This is not a result.
L203: “observational data shows that the SSA emission distribution”. There is no mention of observed SSA emission in this paper so I do not know what this is referring to.
L207–210: How frequent are ETCs in the study area at that period, i.e. how many grid cells are in a cyclone at each given time step? You say that ETCs contribute to 63% of SSA emissions, but if the ETCs cover 63% of the grid cells then you cannot say that they enhance emissions, maybe the emission per grid cell is the same with or without cyclone but cyclones happen to cover 63% of the area. On the other hand if the ETCs cover say 10% of the space-time domain but account for 63% of the emitted mass, then you can say that they strongly enhance emissions. This frequency/coverage argument is really missing here to be able to conclude on emission enhancement.
L218: “The average sea level pressure of the 922 cyclone points reaches the minimum of 977.5 hPa” - in your measurements, SLP at the ship location is consistently below 975 hPa during cyclones. What do you make of this modeled minimum value above 975hPa? Is the model adequately reproducing these cyclones?
Figure 4c: what is the meaning of a WCB flux in the “background” case. And what are these background cases? How is this WCB flux calculated?
L271—272: I do not see this on the figure. For example I can see blue dots above yellow and red dots in the 40—60hr duration window.
L290: “This process drives the upward transport of 15,060 kg s-1 of SSA” - it is unclear how you calculated this number. For reproducibility of your results you should state more clearly what this number (and the others) corresponds to and how it was obtained, ideally including a formula for the calculation.
L296—298: what about the southern quadrants? The net dissipative flux accounts only for northern quadrants, why is that?
Figure 5a: I like the idea of this figure, but it is misleading since it only represents results/numbers for a 6-week-long simulation. This should be clearly stated somewhere, otherwise the reader might think this is an annual mean characteristic, whereas it is only valid for a particular season. If you really wanted to make such a figure, I think you should use sea salt concentration fields from e.g. MERRA-2 directly and combined with the ETC catalog for a whole year, which then would give you more general numbers.
Figures 4, S4, S5: the different scales of colorbars make it hard to compare the panels. You should harmonize the colorbars to have the same min and max values.
Please harmonize the tenses used throughout the paper. Several times the past tense is used where it should not, which can make the sentence confusing.