the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 01 Apr 2026
Sea Salt Aerosols from Blowing Snow: Contributions to Radiative Forcing
Abstract. The Arctic and Antarctic regions experience significant climate impacts from aerosol-cloud-radiation interactions, yet the role of sea salt aerosols (SSA) emitted through blowing snow remains poorly quantified. This study implements a parameterization of the SSA production of blowing snow in both the TM5 global chemical transport model and the EC-Earth3 global climate model, for AMIP-type as well as transient (SSP3-7.0 for 2015–2051) experiments, assessing the contributions of the blowing snow process to aerosol mass, number, cloud condensation nuclei (CCN) and radiative forcing in both polar regions. Model results are evaluated against observations from the MOSAiC campaign and coastal stations (Villum, Zeppelin, Alert). EC-Earth3 Simulations show that blowing snow substantially increases SSA concentrations during polar winter and spring, especially in the Antarctic where enhancements can exceed 100 % increase in particle numbers, leading to improved agreement with surface and in situ observations. Regionally, TM5 reveals an increase in accumulation mode aerosol and CCN. The resulting surface radiative forcing is globally negative due to increased scattering of shortwave radiation, while enhanced CCN increases longwave cloud effects in the polar lower troposphere. Overall, this work demonstrates that including blowing snow SSA emissions is essential for realistically representing polar aerosol burdens, seasonal cycles, and climate feedbacks in global models.
- Preprint
(2937 KB) - Metadata XML
-
Supplement
(6591 KB) - BibTeX
- EndNote
Status: closed
- RC1: 'Comment on egusphere-2026-1575', Anonymous Referee #1, 20 Apr 2026
-
RC2: 'Comment on egusphere-2026-1575', Anonymous Referee #2, 12 May 2026
This manuscript presents modelling work (using TM5 and EC-Earth3) on the impacts of SSA sourced from blowing snow on polar particles, CCN, and radiative forcing in both the Arctic and Antarctica. The models were evaluated against observations from the MOSAiC campaign and coastal stations. The authors demonstrate that blowing snow can substantially increase polar wintertime SSA concentrations, e.g. by up to 100% in Antarctica, thereby improving agreement with surface and in situ observations. In addition, they show that including blowing-snow SSA is essential for realistically representing polar aerosol burdens, seasonal cycles, and climate feedbacks in global models. Moreover, they compare their model results with those from the GOES-Chem model, highlighting the complexities of climate feedbacks in response to SSA sourced from blowing snow and indicating the importance of cloud microphysical processes in determining cloud-radiation interactions. This study is both scientifically important and timely. Therefore, I support its publication in this journal after careful consideration of my comments, as shown below.
Major comments:
1. Section 3.1 Comparison to MOSAiC observations
The authors compared their model results with the MOSAiC cruise aerosol number densities (and with GOES-Chem results) to validate the model. My major concern is the absence of Aitken-mode aerosol in your model outputs. As in Gong et al.’s paper, their kappa analysis strongly indicated that a large proportion of particles fall within this size mode; therefore, your model likely “underestimates” Aitken-mode SSAs, though this is not due to the emission from BS. As you later (in Section 5) attributed this absence to enhanced transfer from coagulation and condensation effects. In my view, this should be clearly discussed here and highlighted as a model limitation in reproducing Aitken-mode SSAs to avoid misleading. Since these particles significantly affect clouds and radiation, this “limitation” needs to be addressed when interpreting the model’s results and its climate feedbacks.
In addition, you later said, “The correlation between observation and TM5 baseline is 0.37, while with blowing snow it is 0.39; however, the difference between the correlations is not statistically significant using Williams’ test.” If I understood correctly, this comparison covers the full year (results in Figure 2). If so, you may need to state this clearly, as there is no blowing-snow effect in summer and early autumn. If you wanted to highlight the BS effect, it would be better to focus on the winter period, e.g., Nov/2019-Apr/2020, rather than the full-year results. The same applies to your CCN comparison.
2. Section 2.3 on Blowing Snow
The description of blowing snow and SSA production is too simple and incomplete. For example, you stated that “snow salinity (which is fixed per size bin)”. What does this mean exactly? Did you apply a fixed salinity for all size bins, or different salinities to different bins? Please provide more information. I feel that a simplified scheme, rather than a full scheme, was used, though you stated “… the same sea-salt aerosol (SSA) production from blowing snow parameterisation as used in the TOMCAT chemical transport model...”
In the same section, you said, “The blowing-snow parameterisation is most sensitive to how the size distribution is described. Among the input meteorological variables, the blowing-snow parameterisation is most sensitive to changes in relative humidity (Ranjithkumar et al., 2025).” I cannot follow what you mean in the first sentence by “how the size distribution is described”. The statement in the second sentence is incorrect. The impact of RH on SSA production is mediated through the BS bulk sublimation calculation. Although many factors may affect SSA production, wind speed remains the key factor governing SA production, as addressed in Ranjithkumar et al.’s work.
Minor comments:
Some paragraphs were indented at the beginning, but many others (e.g. on pages 4 and 6) were not. Tidy the text up to use the same format.
Line 7: change capital “S” in the word “Simulation”.
Lines 26-27: add references to the statement that “Model estimates show that SSA account for more than a quarter of the Arctic aerosol number concentration”. Note that the GOES-Chem results (Gong et al., 2023) showed a much larger contribution, e.g., up to 45% in the central Arctic.
Line 30: Note that wave breaking (de Leeuw et al. 2011) is an important pathway that also produces bubbles and generates sea spray directly from the open-ocean surface. Therefore, both bubble bursting and wave breaking should be considered direct sources of open-ocean SSA.
de Leeuw, G., E. L. Andreas, M. D. Anguelova, C. W. Fairall, E. R. Lewis, C. O’Dowd, M. Schulz, and S. E. Schwartz, 2011: Production flux of sea spray aerosol. Rev. Geophys., 49, RG2001, https://doi.org/10.1029/2010RG000349.
Lines 47-48: Rephrase or delete the sentence “Yang et al. (2019), Gong et al. (2023) and Ranjithkumar et al. (2025) implemented blowing snow SSA production in the p-TOMCAT chemical transport model.” As Gong et al.’s work used GOES-Chem rather than p-TOMCAT results, and the other two works focused on aerosols rather than climate impacts.
Lines 79, 83: change TOMCAT to p-TOMCAT.
Line 124-125: “In the Arctic, most of the sea salt emissions due to blowing snow occur during Boreal wintertime”, but I see an increase in March and April in Fig 2, they are spring, not winter. What is your definition of “wintertime”?
Line 212: Rephrase “The particle size distribution of snow particles”, I can not follow what you meant.
Line 222: delete the unnecessary dot.
Page 15: is “GOES” the right name of GEOS-Chem?
Figure 1: add locations for stations (Villum, Zeppelin, Alert) in Figure 1a.
Figure 4: add sub-ticks to the X-axis.
Figures 6 & 7: It is unusual to use the model “lev” when presenting the model’s latitude-height cross-section results. Use either the actual height (above the surface) in metres or kilometres, or the pressure on your Y-axis.
Citation: https://doi.org/10.5194/egusphere-2026-1575-RC2 - RC3: 'Comment on egusphere-2026-1575', Anonymous Referee #3, 14 May 2026
- AC1: 'Comment on egusphere-2026-1575', Kalle Nordling, 01 Sep 2026
Status: closed
-
RC1: 'Comment on egusphere-2026-1575', Anonymous Referee #1, 20 Apr 2026
This manuscript presents blowing snow’s contribution to aerosol, cloud condensation nuclei, and radiative forcing in polar regions based on a modelling study. The paper addresses an important gap in polar aerosol-climate feedback. However, the contribution to the scientific community is limited by the lack of comparison with existing studies and a failure to discuss the broader implications of the findings. I suggest a major revision before publishing on ACP.
Major Comments
- The blowing snow mechanism is not fully clear yet. For example, the salinity, how many ice crystals are produced for each snowflake, sublimation rate under sub-/super-saturations are not clear yet. This paper lacks such information for the model simulation. More importantly, it would be even better to run the sensitivity test for each parameter.
- The current results section is primarily descriptive (i.e., simply stating what is in the figures) rather than an in-depth analysis. The authors fail to place their findings in the context of existing literature. For instance, the simulated SSA AOD and its contribution to the total aerosol load should be compared with previous studies or satellite/AERONET observations. Without this, it is difficult to gauge the reliability of the parameterization.
- Long-term blowing snow simulations lead to cooling at both poles, yet the net flux values provided are positive (Lines 203-204). This paradox requires a detailed explanation. Is this driven by the strong negative forcing during the summer compared to the winter (Fig. S2 & S3)? The efficiency of blowing-snow-SSA production in summer versus winter needs to be explicitly shown to support these conclusions. Furthermore, the lack of radiation simulation results for the Antarctic—where the impact is claimed to be more significant—is a major omission.
- The figures require significant aesthetic and functional improvements. Some panels are cluttered (see Specific Comment Fig. 2), making them nearly unreadable. Data should be presented in a more intuitive manner (e.g., using dual axes or improved colorbar position) to help the reader distinguish between different variables and modes.
- The manuscript concludes abruptly without discussing the broader implications of the work. How do these findings affect our understanding of polar climate sensitivity? What are the limitations of the current TM5 implementation, and what should be the focus of future research?
Specific Comments
Line 38: When mentioning specific CMIP6 models, please provide appropriate citations for the models discussed.
Line 53: Please clarify if the “2019-2020” period applies to both study locations or if there was a time distinction between the Arctic and Antarctic datasets.
Line 131: The correlation improvement here is relatively low. Please elaborate on the physical or numerical reasons behind this discrepancy, or can we only show the correlation during the blowing snow events?
Line 143 & Figure 4: The description and the figure show a different conclusion: as the figure shows, TM5 exhibits a slight decrease in Aitken mode particles when considering blowing snow instead of increasing. The following statement of results also indicates a decrease.
Figure 1:
The unit in the legend should use proper superscripts.
Add the specific locations of the three in-situ observation sites and the corresponding TM5 grid cells in Fig.1 Panel (a).
Standardize panel labels description as (a), (b), (c).
The colorbar is currently too large.
Figures 2 & 3: The readability is low.
Please specify the time resolution of the data.
Suggestion: Use a dual x-axis, moving the “added particle” data to the top x-axis to reduce overlap.
Figure 7: It should be “0.3% SS” instead of “0.3 supersaturation”. Additionally, the colorbar scale should be adjusted to clearly highlight the differences in CCN enhancement capabilities between the North and South Poles.
-
RC2: 'Comment on egusphere-2026-1575', Anonymous Referee #2, 12 May 2026
This manuscript presents modelling work (using TM5 and EC-Earth3) on the impacts of SSA sourced from blowing snow on polar particles, CCN, and radiative forcing in both the Arctic and Antarctica. The models were evaluated against observations from the MOSAiC campaign and coastal stations. The authors demonstrate that blowing snow can substantially increase polar wintertime SSA concentrations, e.g. by up to 100% in Antarctica, thereby improving agreement with surface and in situ observations. In addition, they show that including blowing-snow SSA is essential for realistically representing polar aerosol burdens, seasonal cycles, and climate feedbacks in global models. Moreover, they compare their model results with those from the GOES-Chem model, highlighting the complexities of climate feedbacks in response to SSA sourced from blowing snow and indicating the importance of cloud microphysical processes in determining cloud-radiation interactions. This study is both scientifically important and timely. Therefore, I support its publication in this journal after careful consideration of my comments, as shown below.
Major comments:
1. Section 3.1 Comparison to MOSAiC observations
The authors compared their model results with the MOSAiC cruise aerosol number densities (and with GOES-Chem results) to validate the model. My major concern is the absence of Aitken-mode aerosol in your model outputs. As in Gong et al.’s paper, their kappa analysis strongly indicated that a large proportion of particles fall within this size mode; therefore, your model likely “underestimates” Aitken-mode SSAs, though this is not due to the emission from BS. As you later (in Section 5) attributed this absence to enhanced transfer from coagulation and condensation effects. In my view, this should be clearly discussed here and highlighted as a model limitation in reproducing Aitken-mode SSAs to avoid misleading. Since these particles significantly affect clouds and radiation, this “limitation” needs to be addressed when interpreting the model’s results and its climate feedbacks.
In addition, you later said, “The correlation between observation and TM5 baseline is 0.37, while with blowing snow it is 0.39; however, the difference between the correlations is not statistically significant using Williams’ test.” If I understood correctly, this comparison covers the full year (results in Figure 2). If so, you may need to state this clearly, as there is no blowing-snow effect in summer and early autumn. If you wanted to highlight the BS effect, it would be better to focus on the winter period, e.g., Nov/2019-Apr/2020, rather than the full-year results. The same applies to your CCN comparison.
2. Section 2.3 on Blowing Snow
The description of blowing snow and SSA production is too simple and incomplete. For example, you stated that “snow salinity (which is fixed per size bin)”. What does this mean exactly? Did you apply a fixed salinity for all size bins, or different salinities to different bins? Please provide more information. I feel that a simplified scheme, rather than a full scheme, was used, though you stated “… the same sea-salt aerosol (SSA) production from blowing snow parameterisation as used in the TOMCAT chemical transport model...”
In the same section, you said, “The blowing-snow parameterisation is most sensitive to how the size distribution is described. Among the input meteorological variables, the blowing-snow parameterisation is most sensitive to changes in relative humidity (Ranjithkumar et al., 2025).” I cannot follow what you mean in the first sentence by “how the size distribution is described”. The statement in the second sentence is incorrect. The impact of RH on SSA production is mediated through the BS bulk sublimation calculation. Although many factors may affect SSA production, wind speed remains the key factor governing SA production, as addressed in Ranjithkumar et al.’s work.
Minor comments:
Some paragraphs were indented at the beginning, but many others (e.g. on pages 4 and 6) were not. Tidy the text up to use the same format.
Line 7: change capital “S” in the word “Simulation”.
Lines 26-27: add references to the statement that “Model estimates show that SSA account for more than a quarter of the Arctic aerosol number concentration”. Note that the GOES-Chem results (Gong et al., 2023) showed a much larger contribution, e.g., up to 45% in the central Arctic.
Line 30: Note that wave breaking (de Leeuw et al. 2011) is an important pathway that also produces bubbles and generates sea spray directly from the open-ocean surface. Therefore, both bubble bursting and wave breaking should be considered direct sources of open-ocean SSA.
de Leeuw, G., E. L. Andreas, M. D. Anguelova, C. W. Fairall, E. R. Lewis, C. O’Dowd, M. Schulz, and S. E. Schwartz, 2011: Production flux of sea spray aerosol. Rev. Geophys., 49, RG2001, https://doi.org/10.1029/2010RG000349.
Lines 47-48: Rephrase or delete the sentence “Yang et al. (2019), Gong et al. (2023) and Ranjithkumar et al. (2025) implemented blowing snow SSA production in the p-TOMCAT chemical transport model.” As Gong et al.’s work used GOES-Chem rather than p-TOMCAT results, and the other two works focused on aerosols rather than climate impacts.
Lines 79, 83: change TOMCAT to p-TOMCAT.
Line 124-125: “In the Arctic, most of the sea salt emissions due to blowing snow occur during Boreal wintertime”, but I see an increase in March and April in Fig 2, they are spring, not winter. What is your definition of “wintertime”?
Line 212: Rephrase “The particle size distribution of snow particles”, I can not follow what you meant.
Line 222: delete the unnecessary dot.
Page 15: is “GOES” the right name of GEOS-Chem?
Figure 1: add locations for stations (Villum, Zeppelin, Alert) in Figure 1a.
Figure 4: add sub-ticks to the X-axis.
Figures 6 & 7: It is unusual to use the model “lev” when presenting the model’s latitude-height cross-section results. Use either the actual height (above the surface) in metres or kilometres, or the pressure on your Y-axis.
Citation: https://doi.org/10.5194/egusphere-2026-1575-RC2 - RC3: 'Comment on egusphere-2026-1575', Anonymous Referee #3, 14 May 2026
- AC1: 'Comment on egusphere-2026-1575', Kalle Nordling, 01 Sep 2026
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 465 | 219 | 60 | 744 | 82 | 43 | 50 |
- HTML: 465
- PDF: 219
- XML: 60
- Total: 744
- Supplement: 82
- BibTeX: 43
- EndNote: 50
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents blowing snow’s contribution to aerosol, cloud condensation nuclei, and radiative forcing in polar regions based on a modelling study. The paper addresses an important gap in polar aerosol-climate feedback. However, the contribution to the scientific community is limited by the lack of comparison with existing studies and a failure to discuss the broader implications of the findings. I suggest a major revision before publishing on ACP.
Major Comments
Specific Comments
Line 38: When mentioning specific CMIP6 models, please provide appropriate citations for the models discussed.
Line 53: Please clarify if the “2019-2020” period applies to both study locations or if there was a time distinction between the Arctic and Antarctic datasets.
Line 131: The correlation improvement here is relatively low. Please elaborate on the physical or numerical reasons behind this discrepancy, or can we only show the correlation during the blowing snow events?
Line 143 & Figure 4: The description and the figure show a different conclusion: as the figure shows, TM5 exhibits a slight decrease in Aitken mode particles when considering blowing snow instead of increasing. The following statement of results also indicates a decrease.
Figure 1:
The unit in the legend should use proper superscripts.
Add the specific locations of the three in-situ observation sites and the corresponding TM5 grid cells in Fig.1 Panel (a).
Standardize panel labels description as (a), (b), (c).
The colorbar is currently too large.
Figures 2 & 3: The readability is low.
Please specify the time resolution of the data.
Suggestion: Use a dual x-axis, moving the “added particle” data to the top x-axis to reduce overlap.
Figure 7: It should be “0.3% SS” instead of “0.3 supersaturation”. Additionally, the colorbar scale should be adjusted to clearly highlight the differences in CCN enhancement capabilities between the North and South Poles.