the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
A Spray-Mediated Surface Turbulent Heat Flux Dataset for the Southern Ocean
Abstract. Sea spray generated by wave breaking under high wind speeds strongly influences air-sea turbulent heat exchange. Nowhere is this process more critical than the Southern Ocean where the most extreme wave climate on Earth is located. However, current existing gridded air-sea heat flux datasets lack representation of sea spray-mediated component, which lead to a major problem in assessing and understanding the role of sea spray in the Southern Ocean. Here, we present a new gridded dataset of spray-mediated latent and sensible heat fluxes over the Southern Ocean that incorporates a full sea spray microphysical model, which considers spray generation that depends on regional wind and wave states. Our data allows a quantitative assessment of the spray-mediated contribution to the total air-sea heat flux exchange. Climatologically, spray-mediated heat fluxes range from 0.04 to 39.6 W m-2, with large fluxes in the circumpolar region (50° S–60° S), especially in the Indian sector during austral winter. Validation of an extreme cyclone shows that, within the high wind sector of the cyclone, spray-mediated heat fluxes double the regional fluxes compared to conventional bulk algorithm estimates. This dataset serves a critical foundation for examining the Southern Ocean’s role in the global energy budget.
- Preprint
(2397 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on essd-2026-476', Anonymous Referee #1, 29 Sep 2026
-
RC2: 'Comment on essd-2026-476', Anonymous Referee #2, 02 Nov 2026
The manuscript "A Spray-Mediated Surface Turbulent Heat Flux Dataset for the Southern Ocean" provides a unique dataset that demonstrates the potential significance for spray induced air-sea heat exchange across the Southern Ocean. While I believe this work is interesting and should be considered for publication, there are some key points I would like to raise.
- L15, I think before it can be claimed that without sea spray mediated fluxes there is a major problem in understanding air-sea heat fluxes in the Southern Ocean, it should first be shown that observations and reanalysis or data products differences are considerably large.
- L20, is the cyclone statistically extreme or is this just an adjective?
- Perhaps avoid using rainbow color schemes in the figures. This would help colorblind people interpret the figures better.
- Figures 1 and 2, are these winter climatologies? Summer in the Southern Hemisphere is DJF.
- Eq. 1 (L109) and L111. The linear drag coefficient C_d is given without a citation. This formulation increases with wind speed even at very high winds. Observations and more recent parameterisations instead show C_d levelling off at high winds, roughly above 25 to 30 m s-1 (e.g. Edson et al., 2013). The C_d thus likely overestimates R_B at high winds.
- Eq. 2 (L110). The authors compute the kinematic viscosity as a function of sea surface temperature T_s. Yet, the kinematic viscosity is a property of air. The polynomial in Eq. 2 (Andreas, 1989) is a fit for air temperature at standard pressure. Using T_s in place of T_air may be a reasonable approximation for the air in the viscous sub-layer near the wave surface, but this should be stated and justified. Also, add units please in the description.
- Eq. 4 (L136). There seem to be some important typos compared to Zhao et al. (2024). The coefficients for the third and fourth size ranges look like typos. For 75 ≤ r_0 < 200 um, the coefficient should be 4.41 x 10^1, not 1.40 x 10^1. For 200 ≤ r_0 ≤ 500 um, it should be 1.41 x 10^13, not 1.40 x 10^13. If the printed values were used, the results would need to be recalculated. Then, please make the inequalities consistent and include the equals sign where needed, so the upper limit is closed (r_0 ≤ 500 um). Check the Zhoa et al. (2024) paper.
- Eq. 6 (latent heat). The second expression, (28.34 - 0.00149 T_air) x 10^5, is the latent heat of sublimation. Spray droplets are liquid seawater, so the latent heat of vaporization should apply in both cases.
- Eq. 6 (latent heat). T_s < 0 degC doesn’t mean the water is frozen. Seawater freezes at about -1.8 degC, and droplets can stay liquid well below 0 degC. The switch between the two expressions depends on T_s, but the formulas use T_air. Please explain this choice or make it consistent. The second line condition should be T_s < 0.
- L146, rho_w = 1000 kg m-3, which is the value for fresh water. Seawater is about 1025 kg m-3. The difference is small, but it’s worth correcting.
- L148, Is r_m the same as r_0, just in metres? If so, please use one symbol with consistent units throughout. If not, define how r_m relates to r_0.
- There are quite a few grammatical errors. Check the writing again carefully. Also, ensure that each acronym has been described before using it. I found more than one instance where acronyms are used without explanation.
-
Choice of bulk algorithm (L290, Fig. 7a). The bulk fluxes come from Smith (1988), which uses nearly constant heat and moisture transfer coefficients (C_H and C_E of about 1.0-1.2 x 10^-3). These were derived mostly from moderate wind data. So the underestimation in Fig. 7a may come from the outdated bulk parameterisation as much as from missing spray processes. The comparison then can’t show that the spray model is what fixes the bias. Please redo the analysis with a current algorithm such as COARE 3.6, which better represents transfer at high winds. Then show whether a systematic low bias still remains for the spray contribution to correct. Without this step, the conclusion that “the spray microphysical model provides a significant correction” isn’t supported.
-
L292 (underestimation for THF > 500 W m⁻²). This claim rests on about 5 points in Fig. 7a. Please state the number of observations above 500 W m-2 and the wind speeds where it occurs. Five points aren’t enough to establish a systematic bias. The point that the bias “becomes particularly clear” above 20 m s-1 also needs the number of points in that range.
-
Fig. 7b and L294, is there an overcorrection? Adding spray does improve the few high-flux points. But it also pushes the main cluster (roughly 200–600 W m-2) above the 1:1 line. In Fig. 7a, that cluster lies close to the line. So the spray term seems to cause an overestimate for most of the data. The text should acknowledge this rather than say the bias is “largely corrected.” Please give summary statistics such as mean bias, RMSE and the regression slope for both panels. Ideally, also split them by wind-speed bin (for example 15-20 and >20 m s-1) so the reader can see the net effect.
- L330: The spatial match between spray-mediated flux and significant wave height is expected by construction, since the spray model is forced by wind and wave fields. It is not independent evidence that spray fuels the cyclone. No observations (flux measurements) support a causal role in cyclone development. Please remove the claim or support it with observations.
- L336, the bulk algorithm “not capable to allow the development of the record cyclone”. This is not proven. ERA5, which uses a conventional bulk scheme, does reproduce the cyclone. That is how it is identified and analysed here. In addition, bulk algorithms such as COARE are tuned against observed fluxes that already include some spray contribution. Adding the spray term on top of the interfacial flux risks double counting.
- Section 4.4 reports correlations without a physical explanation. Please explain the mechanism linking each mode to spray-mediated THF, for example through wind speed, wave height, air-sea temperature or humidity contrasts. Have you checked regressions of wind speed, wave height and the air-sea temp/humidity differences on the indices? That would show whether spray adds anything beyond the known wind response to SAM.
- Figure 9, please states the timescales used for the correlations. Are these monthly? How was the significance computed and was autocorrelation accounted for?
Citation: https://doi.org/10.5194/essd-2026-476-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 425 | 149 | 102 | 676 | 167 | 137 |
- HTML: 425
- PDF: 149
- XML: 102
- Total: 676
- BibTeX: 167
- EndNote: 137
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors present a new dataset of spray mediated latent and sensible heat fluxes for the Southern Ocean. The authors combine previously published flux and sea spray generation models to produce a physically-based, droplet-size dependent model, and use it to determine spray mediated fluxes for the Southern Ocean (South of 35S), 1993 – 2022, at monthly, ¼ degree resolution.
Such a dataset would be valuable to the community – the paper itself justifies this need and is in general very well written. I have no specific corrections or suggestions for the paper itself. The dataset and code are provided in a easy-to-use format and their use and access are well described.