the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 31 Jul 2025
Mesoscale eddies heterogeneously modulate CO2 fluxes in eddy-rich regions of the Southern Ocean
Abstract. Mesoscale eddies are known to influence the Southern Ocean biogeochemistry. However, the distinct contributions of cyclonic and anticyclonic eddies to air-sea CO2 fluxes, as well as their longer-term effects remain poorly studied. We present results from a 27-year global eddy-resolving ocean-biogeochemical simulation. We used the Okubo-Weiss parameter to classify the modeled flow regimes into cyclonic and anticyclonic eddies, peripheries, and the surrounding background waters. Our results reveal a heterogeneous influence of eddies depending on the region, driven by regional differences in eddy intensity and the gradients in background properties. The factors controlling CO2 fluxes within eddies follow the same degree of importance as in background waters, with ∆pCO2 being the dominant factor. This is driven primarily by changes in dissolved inorganic carbon. Our analysis shows that eddies act as a persistent carbon sink on decadal timescales, while their influence on shorter timescales is more variable and strongly shaped by eddy polarity. Anticyclonic and cyclonic eddies and periphery account for around 10 % of the Southern Ocean’s carbon uptake, with anticyclonic eddies showing the highest carbon uptake per unit area. The ability of eddies to absorb carbon computed in our results is consistent with recent observational estimates, confirming that the model realistically represents the influence of mesoscale eddies on CO2 fluxes. Above all, our results underscore the role of mesoscale eddies in enhancing carbon uptake across the Southern Ocean.
- Preprint
(12625 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 17 Sep 2025)
-
RC1: 'Comment on egusphere-2025-3067', Anonymous Referee #1, 31 Aug 2025
reply
Salinas-Matus et al. present an investigation into the effects of mesoscale eddies on the air-sea CO2 flux within eddy rich regions of the Southern Ocean. They find geographical differences in the eddy effects on air-sea CO2 flux between the Brazil-Malvinas confluence, Agulhas Retroflection and the region South of Tasmania. They investigate the factors driving these differences in the air-sea CO2 fluxes and find that the pCO2 gradient (ΔpCO2) is the primary driver, which itself is mainly driven by differences in dissolved inorganic carbon (DIC). Using a temporal decomposition a further result shows that eddies act as persistent CO2 sinks on decadal timescales but are more variable at shorter timescales. I find this study to provide a comprehensive analysis of the eddy rich regions of the Southern Ocean and the modulation of the air-sea CO2 fluxes and can recommend for publication once my remaining comments below have been addressed (especially the analysis in Figure 5 which is unclear).
General comment: I’d suggest to aid readers, to change the cyclonic eddy results to be coloured blue, and the anticyclonic eddies to be coloured red. Although the legends are clear in defining the colours, the common conventions in previous work are for anticyclonic (generally ‘warm’) eddies are coloured red, and cyclonic (generally ‘cold’) eddies blue.
General comment: I’d suggest picking a different colour for the periphery to avoid having red and green to aid readers that are colour-blind.
Line 27: “Cyclone” should be “Cyclonic”
Line 86: Can some details of the air-sea CO2 flux parameterisation be mentioned? What was the parameterisation used for kw (and Schmidt number), and the solubility.
Line 92: What was the native time resolution of the model, that was then collated into the daily averages?
Line 113: What is the sensitivity of the results to the Okubo-Weiss parameter values used to define the background, eddy and periphery. Is this a commonly applied value (I don’t see a reference for this selection)? The selection appears to identify large regions far from negative OW values, that get identified as the “periphery” in Figure 1a, which may suggest that this value is too relaxed?
Line 130: I’d suggest the final sentence isn’t required - it could be moved to the conclusion or introduction if the authors would like to keep it.
Figure 1: Suggest changing colour map in Figure 2b (bottom panel) for colour-blindness.
Line 170: Id suggest more is said surrounding the “background” conditions in the Agulhas retroflection region being cooler than the cyclonic eddies. The cyclonic eddies would be forming from the cool side of the retroflection and therefore should be cooler than the background as this water wouldn’t be originating in the Indian Ocean. Figure 3a shows the cooler water on Southern Ocean side of the retroflection, and Figure 2b shows SST anomalies for cyclonic eddies being negative. The background conditions also appear based on Figure 1a to be originating from the Indian South Subtropical gyre with warmer temperatures.
Line 201: Id suggest more is said on the Tasmania region eddies and the sporadic events. These sporadic high uptake events seem to be more prevalent in the Tasmania region, compared to the other regions, and have a large effect on the Figure 4 uptake results. As pCO2 is the dominant driver in the integrated fluxes, could you elaborate on a mechanism?
Table 2: It is unclear what the ± values denote. Is it mean ± the standard deviation?
Line 253: Could the greater influence of kw be due to the increase in wind speeds generally observed over anticyclonic eddies? (Frenger et al., 2013) and the opposite for cyclonic eddies?
Line 264: How was the pCO2 (atm) prescribed? Can details of this be added in the methods for model setup?
Line 271: Figure 5 captions indicate these regressions are for ΔpCO2 regressed against the thermal and non-thermal components separately, where as the text indicates this is the total flux. Based on the regressions I think this is each component regressed against the total flux. The analysis in this form may not answer the aim as the contributions of the thermal and non-thermal components to the changes in ΔpCO2 would be combined with variability in kw (and other inputs to the fluxes) when regressing against fluxes. I am unsure of the aim of this portion of the analysis and suggest the authors should clarify this analysis (Lines 264-275).
Code availability: Code is available to do sections of the analysis but was unable to find plotting scripts to complete the analysis.
Data availability: I note no data availability statement. Are these fields available or can they be requested? I expect not due to data volume, but this should be stated. These fields appear very useful for studying mesoscale eddies over a long time period, so could be useful for the community.References
Frenger, I., Gruber, N., Knutti, R., & Münnich, M. (2013). Imprint of Southern Ocean eddies on winds, clouds and rainfall. Nature Geoscience, 6(8), 608–612. https://doi.org/10.1038/ngeo1863Citation: https://doi.org/10.5194/egusphere-2025-3067-RC1 -
RC2: 'Comment on egusphere-2025-3067', Anonymous Referee #2, 06 Sep 2025
reply
Salinas-Matus et al. investigated the influence of anticyclonic and cyclonic eddies, periphery and “background” in modulating air-sea CO₂ fluxes. For this, three different regions with differing characteristics in the Southern Ocean were picked, namely the Agulhas Region, the Brazil-Malvinas Confluence Region and the region south of Tasmania. The aim of their study is to investigate the physical and biogeochemical drivers of the air-sea fluxes at daily to decadal timescales.
Mesoscale eddies in the Southern Ocean play a key role in shaping air–sea CO₂ fluxes and carbon uptake. Using a 27-year, global eddy-rich (nominal resolution 8.4-10 km) ocean biogeochemical simulation (ICON-O with HAMOCC), eddies were identified via the Okubo–Weiss parameter and analyzed across their cores, peripheries, and surrounding waters (“background”). Since the eddies were not tracked, the lifetime of the eddies was not considered in their study.
The study highlights that the influence of eddies on CO₂ fluxes is strongly region-dependent, with controlling mechanisms largely consistent with those in background waters. Variations in air–sea CO₂ exchange are primarily driven by changes in the ocean–atmosphere pCO₂ disequilibrium (ΔpCO₂), itself dominated by dissolved inorganic carbon (DIC) dynamics.
Anticyclonic eddies consistently show the strongest ingassing per unit area, which they attribute to eddy-pumping processes that transfer carbon to deeper layers, thereby enhancing long-term uptake capacity. Importantly, eddy peripheries emerge as a major contributor to regional CO₂ fluxes: while their uptake efficiency is intermediate between eddy cores and background waters, their much larger spatial coverage makes them an integral component of the Southern Ocean carbon budget which requires further analysis.
On decadal timescales, mesoscale eddies emerge as a net carbon sink, contributing roughly 10% of the Southern Ocean’s total carbon uptake if integrated over the area that the eddies cover. On shorter timescales, however, flux anomalies are more variable and reflect contrasting behaviors of cyclones and anticyclones.
General comments:
The article reads very smoothly and is easily understandable. The Abstract and Introduction highlight the relevance of the research interest which is also justified by profound literature references. General information about the model and the experiment design which is crucial in order to evaluate the results is given. The presentation of the results is in a logical order, easy to follow with the presented Figures and sufficiently substantiated with observational literature. The summary rounds up the paper in a very nice way, clearing all questions presented in the introduction and pointing out the main findings. Overall, Salinas-Matus et al. is an enjoyably written paper with interesting new insights on the driving mechanisms of eddies on CO₂ fluxes that may need some clarification in some methodological details as well as a more detailed discussion about the overall impact of eddies on the Southern Ocean CO₂ fluxes (see comments below) .
Specific comments:
- L115ff: Is there literature based on which the threshold values were defined? (Common definition e.g. 0.2* stdev(OW) as Schütte et al. 2016)
- Fig.2.: Some eddies look very deformed and unusually large → did the authors check for several local maxima/minima?
- L126: could you elaborate on how you preprocess the data prior carrying out the Fourier analysis? The eddies are at different locations and several for each timestep. Is the average of “eddy composites” (defined as OW below the chosen threshold) taken for each time step (daily) and then the Fourier analysis on the such obtained time series applied? If this is the case, maybe a different term than “composite” should be used or at least defined.
- E.g., Fig. 2: How are anomalies defined? Also, tell how is the sign of the flux defined wrt ingassing/outgassing?
- Eddy contribution to the Southern Ocean carbon sink: the authors provide % numbers for the contribution of eddies given as flux integrated over the “eddy” (= OW below threshold) area. How much bigger is this flux than if one assumed “background flux conditions” over the same area (in %)?
Along similar lines, the per square meter fluxes (Fig. 4c), how different are they compared to the background fluxes (in %)? - I am asking as such a statement (of 10-20% contribution by eddies to the Southern Ocean carbon sink) makes the eddy contribution sound very important. The anomalous contribution (i.e., “what if” the area covered by eddies was not covered by eddies) is more interesting. There is a lot of compensation between anticylones, cyclones, and perhaps seasonality, as the authors point out themselves. E.g., they cite Song et al., 2016, in the introduction (polarity dependence of fluxes), but do not get back to it in the discussion. How do the authors’ findings compare?
- Given the above two points, I do not feel that comfortable with a strong statement that eddies are very important for the Southern Ocean carbon sink (because of the local air-sea flux anomalies they cause, which is the focus of the paper); are the authors? I would be more comfortable if they discussed the above aspects in their manuscript more clearly/transparently.
- L86 or L252ff: mention somewhere, that, given that this is an ocean-only model (no coupling to the atmosphere) there is no eddy feedback on winds, i.e., a bit of the effect eddies might have on air-sea CO2 fluxes is missing?
- L103: the authors discuss the drift of the coarser resolution model, can they also briefly comment on the (likely) drift of the 10 km model (which was run for a much shorter time period)
- L151: Can the authors elaborate on how/why they expect the eddy intensity to relate to air-sea CO2 fluxes?
- L189 “mesoscale flow regimes (anticyclones, cyclones, and periphery) have a greater capacity for carbon uptake compared to the "background" (Fig. 4a)”: not true for all regions, and not for all of the SO?
- L192 “Among these regimes, anticyclonic eddies show an enhanced ability to take up CO2 (Fig. 4a).”: see comment above, how large is the enhanced ability, in absolute numbers and relative to the background flux?
- L197 “and the stronger vertical gradient in those two regions (Fig. 3b).”: do one clearly see this from the Figure?
- L202 “a pattern influenced by sporadic but intense carbon uptake events (Fig. 3).”: can you elaborate on how one can conclude this from Fig. 3?
- L220 “it is worth mentioning that both anticyclonic and cyclonic eddies do exhibit episodes of CO2 outgassing (Fig. 3c).”: can we see this in Fig. 3c? If so, how, can you elaborate?
- L292ff importance of eddy type: mention that there is compensation of negative and positive flux anomalies?
- L295 “despite the smaller region they occupy”: could you provide the number here (area coverage in %, you should have the numbers in your analysis), see comment above, to get a sense of to which extent the eddies enhance the flux.
- Given that the simulation is a historical simulation (i.e., natural fluxes, with the anthropogenic perturbation of increasing atmospheric CO2 on top): curiosity: do you have a hypothesis if the flux change (reduced outgassing) is more due to enhanced anthropogenic carbon uptake, or suppression of outgassing of carbon rich subsurface waters? If not, could you discuss these aspects in the discussion?
Technical comments:
- L27. the role of cyclonic and anticyclonic eddies
- L93: mention specifically that it is eddy-rich
- L236f: repetitive with lines 208-210
- larger Figure labels, some not readable in when printed
- add sign convention to Figure description: negative values outgassing, positive ingassing
- Fig.2 b) take care of colorblind-friendlier choice of colormap and add whether AE or CE are blue or red
- AE are presented in blue colour, CE in red colours. This is a bit confusing, the other way around would be way more intuitively
- Fig. 1a: standard deviation vorticity
- Fig. 4f: Frequency
- L323: I suggest to add units to the variables.
- L338f: pCO2ocean
References
Schütte, F., Brandt, P., and Karstensen, J.: Occurrence and characteristics of mesoscale eddies in the tropical northeastern Atlantic Ocean, Ocean Sci., 12, 663–685, https://doi.org/10.5194/os-12-663-2016, 2016.
Citation: https://doi.org/10.5194/egusphere-2025-3067-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 700 | 50 | 8 | 758 | 19 | 28 |
- HTML: 700
- PDF: 50
- XML: 8
- Total: 758
- BibTeX: 19
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1