the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Jan 2026
Interaction of AMOC and Intrinsic Multi-decadal Southern Ocean Variability
Abstract. A strongly-eddying version (0.1° horizontal resolution) of the Parallel Ocean Program (POP) shows pronounced intrinsic multi-decadal variability in the Southern Ocean, the so-called Southern Ocean Mode (SOM). This Southern Ocean multi-decadal variability is induced by eddy-mean flow interaction and deep convection. The SOM variability propagates through the global ocean and influences the strength of the Atlantic Meridional Overturning Circulation (AMOC) by about 3 Sv. The opposite role on how the AMOC influences the SOM is unknown, as this requires long simulations and preferably with different AMOC background states. Here, using the results of a simulated AMOC collapse in the strongly-eddying ocean-only POP version, we find that the amplitude of the SOM is substantially reduced following an AMOC collapse. Associated changes in horizontal and vertical density variations lead to a weakening of the Antarctic Circumpolar Current transport and a shutdown of deep convection in the Weddell Sea. In contrast, these changes promote deep convection events and the emergence of multi-decadal variability in the Pacific sector of the Southern Ocean. A mechanical energy budget analysis shows both a reduction in the wind input and a disruption of the phase difference between wind work and the potential to kinetic energy conversion. The results highlight the strong connection between the AMOC and intrinsic multi-decadal variability in the Southern Ocean.
- Preprint
(19795 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 20 Mar 2026)
- RC1: 'Comment on egusphere-2026-209', Anonymous Referee #1, 17 Feb 2026 reply
-
RC2: 'Comment on egusphere-2026-209', Quentin Dalaiden, 26 Feb 2026
reply
Smolders et al. make use of a multi-century eddy-permitting ocean simulation with North Atlantic freshwater forcing to analyze the relationship between the Atlantic Meridional Overturning Circulation (AMOC) and the Southern Ocean Mode (SOM). First, I would like to note that I do not have a formal background in oceanography. Nevertheless, I greatly enjoyed reading this study and found it highly interesting. In particular, we have also observed a strong relationship between AMOC variability and Antarctic Circumpolar Current (ACC) variability (positive correlation) in our paleo-forced ocean simulation. While the model configuration we used is coarse (1° resolution), the similarity of this relationship with the eddy-permitting configuration presented here may provide further support for further investigation. This relationship may be linked to historical atmospheric forcing that synchronizes ocean variability between the North Atlantic and Southern Ocean, or to other dynamical mechanisms. Additionally, we found similarities between temperature changes in the Pacific and ACC variability.
Overall, I found the study very clear in its choice of diagnostics, and the results are presented in a logical way. However, I believe the manuscript would benefit from additional clarifications and further discussion of several aspects, particularly regarding the emergence of deep convection activity in the Pacific. My specific comments are listed below:
- Line 151. At the end of the simulation, when the AMOC is strongly reduced relative to the initial state, an increased SST variability is observed in the Pacific sector of the Southern Ocean, while decreased variability is observed in the Atlantic sector. One possible interpretation is that the increased variability in the Pacific sector may be related to a weaker or more relaxed ACC, potentially allowing enhanced meridional heat exchanges. A discussion of the role of the ACC in this context would be valuable. In addition, examining the relationship with Ross Gyre variability could help clarify the mechanisms involved. Similarly, an analysis of the Weddell Gyre dynamics would help explain the reduced variability in the Atlantic sector.
- Figure 3. If I understand correctly, the subsurface warming and deep ocean cooling, along with increased salinity at subsurface and fresher conditions at depth, are consistent with a weakening of the overturning circulation. It would be helpful to include a more detailed explanation of the mechanisms responsible for the simulated temperature and salinity changes, rather than relying primarily on references to previous studies, to improve accessibility for readers.
- Line 198. The increase in potential energy toward the end of the simulation is somewhat "surprising". Instead of computing potential energy using a fixed reference density, would it be more appropriate to use a time-varying density reference to better capture changes in multi-decadal variability?
- Line 207. Related to one of my previous comments, a discussion of the relationship between ACC strength and the non-zonality parameter would be useful here, as this could help clarify the dynamical interpretation of the results.
- Deep convection section. The final section of the manuscript, which discusses deep convection, is particularly interesting but raises several questions (not concerns!). The emergence of deep convection activity in the Pacific sector is not entirely clear to me. For example: Why is the NZ region characterized by large stratification changes? How do these stratification changes contribute to the increased SST variability in the Pacific sector shown in Figure 2? How are these stratification changes related to ACC strength and Ross Gyre dynamics (including strength, position, and extent)? Further clarification of these mechanisms would greatly strengthen this section.
- Coupled ocean-atmosphere feedback. The large oceanic changes associated with the weakening of the AMOC (likely) have important atmospheric consequences. Through coupled ocean-atmosphere feedback, these oceanic changes could be amplified or damped. I believe the conclusion section would benefit from a discussion of these coupled feedback, particularly regarding potential atmospheric responses (e.g., wind changes) and their influence on Southern Ocean circulation and variability.Minor comments:
- Line 201: Consider adding a reference to Figure 3 earlier in the sentence, before “increased P,” to guide the reader.
- Line 204: I suggest citing Figure 4e explicitly at the end of the sentence.Citation: https://doi.org/10.5194/egusphere-2026-209-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 186 | 94 | 16 | 296 | 12 | 10 |
- HTML: 186
- PDF: 94
- XML: 16
- Total: 296
- BibTeX: 12
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments:
This study examines the impact of the AMOC on the intrinsic multi-decadal variability in the Southern Ocean, called the Southern Ocean Mode (SOM), using an eddy-resolving ocean model. The authors added the freshwater forcing in the North Atlantic and demonstrated that the SOM in the Atlantic sector gets weaker after the AMOC collapse, while in the Pacific sector, deep convection starts to emerge with multidecadal variability. The weakening of the SOM in the Atlantic sector may be related to the strengthening of the stratification and the weakening of deep convection in the upper ocean, while the emergence of the SOM in the Pacific sector may be linked to the weakening of the stratification and the strengthening of deep convection.
Although these results are solely based on the model simulations, the SOM changes driven by the AMOC collapse are intriguing and worth investigating, given a possible influence of the weakening AMOC on the Southern Ocean and sea ice under future global warming. However, the physical processes underlying the changes in the ocean stratification responsible for the SOM changes are not well examined and remain unclear. For example, the relative contributions of the ocean temperature and salinity to the stratification changes and their sources and pathways are not clarified. The weaker AMOC is expected to induce the weakening of the poleward advection of warm and saline subsurface water (i.e., Antarctic Circumpolar Deep Water) and hence weakening of the deep convection in the Weddell Sea. However, in the Pacific sectors, there are different stratification changes in the NZ, AU, and PA sectors. Why do these stratification changes happen and what are the links with the deep convection changes? This is worth further investigation in this study. Below are other major and specific comments to further improve this study before possible publication in this journal.
1. Changing SOM variability (Section 3.1)
It is interesting to find the change in the SOM variability before and after the freshwater input, but there are several concerns to be addressed before drawing conclusions.
First, the model shows a gradual decrease in the AMOC strength until 300 years when the freshwater forcing is implemented. This gives me the impression that the model does not reach an equilibrium state and requires further spin-up period before conducting the freshwater forcing experiment.
Second, the authors discuss the changes in the frequency of the SOM and Drake Passage transport qualitatively (e.g., L136 and L141), but are these changes statistically significant? I would recommend the authors should conduct wavelet power spectrum analysis with statistical test to the timeseries and describe how the period of the timeseries change before and after the freshwater input.
Third, the authors applied the EOF analysis to the SST anomalies in the Southern Ocean, but are the SST anomalies used in the EOF analysis monthly or yearly mean values? I would recommend the authors should consider strong seasonality of the SST variability in the Southern Ocean.
Fourth, the explained variances of the first EOF mode are very small at 13 % and 7 % before and after the freshwater forcing. Are these EOF modes well separated from the corresponding second EOF modes? I would recommend the authors should perform a statistical test introduced by North et al. (1982).
Lastly, the blue outlined region used for the SOM-P index includes the negative SST variability, so it would be better to modify the box that covers a core region of the positive SST variability, for example, 60-45ºS, 170ºE-150ºW.
- North, G. R., Bell, T. L., Cahalan, R. F., & Moeng, F. J. (1982). Sampling errors in the estimation of empirical orthogonal functions. Monthly Weather Review, 110(7), 699-706.
2. AMOC-SOM coupling (Section 3.2)
This subsection discusses the impact of the freshwater input on the ocean background mean state in the Atlantic sector, but does not discuss the coupling between the AMOC and SOM. The authors mention that the density anomalies associated with the SOM influence the AMOC through their northward propagation within the Atlantic basin (van Westen & Dijkstra 2017). If this is true in the model world, the change in the SOM variability after the freshwater input can also affect the AMOC strength. Is there any feedback processes between the SOM and AMOC changes?
Second, the subsurface ocean gets warmer and saltier after the AMOC collapse, while the deeper ocean gets cooler and fresher (Fig. 3). Are these changes statistically significant and large enough compared to the intrinsic variability? The authors should carefully explain why these temperature and salinity changes occur in the subsurface and deep oceans after the freshwater input.
Third, the authors describe three different isopycnals in Fig. 3c. However, some of the values before and after the AMOC collapse (i.e., 1026.2 vs 1025.7, 1027.7 vs 1027.6) are different and hard to compare with each other. Could the authors plot the same isopycnals in the figure?
Fourth, which of the ocean temperature and salinity changes is more important for the meridional density gradient changes? This can be further investigated by decomposing the density anomalies into the temperature and salinity-driven components.
Lastly, the ACC does not uniformly change in response to the AMOC collapse. For example, the ACC gets stronger north of 40ºS and south of 50ºS, while it gets weaker between 40-50ºS. This may be related to the meridional density gradient change, but why does this heterogeneous change in the ACC happen?
3. Mechanisms of SOM changes (Section 3.3)
The authors draw conclusions from the energetics of three different SOM cycles that the increase in P at the end of the simulation is not driven by enhanced mean wind energy input (i.e., G), but instead is a consequence of changes in the density field associated with the AMOC collapse (i.e., C). This appears to be true, but the relationship with the non-zonality parameter is unclear. For example, the decrease in G and C over time indicates less zonal flow with narrower isopycnals and reduced eddy generation through baroclinic instability, respectively. However, the non-zonality parameter increases, indicating more eddy generation through baroclinic instability. This looks puzzling and contradictory results. Is my understanding correct?
Another intriguing aspect is the periodicity of the cycle and time lag among the variables. In the SOM cycle 1 and 2, P shows a clear cycle with a period of 40-50 years, but what determines the timescale of this cycle? According to Eq. (4), G and C should contribute to the rate of P, and both G and C show similar cycles with a period of 40-50 years. Also, there is 10-yr lag between K and C (L224), but what causes this time lag? By definition in Eq. (5), C should contribute to the rate of K, so the time lag looks very long.
4. Changes in the Southern Ocean deep convection (Section 3.4)
It is interesting to find different responses of the ocean stratification in different sectors to the AMOC collapse. However, the underlying mechanisms need to be clarified. In the NZ and WGKP regions, the stratification increases in the upper 100 m, decreases in the middle layer (250-500 m), and increases again in the deeper layer down to 1500 m. I am speculating that poleward advection of warm and saline subsurface water (i.e., Antarctic Circumpolar Deep Water) should decrease after the AMOC collapse thereby increasing the stratification in the deeper layer and inducing the shallower mixed-layer. Despite this expectation, why does the middle layer show the weaker stratification in both sectors? Does it also affect the initiation of multidecadal oscillation in the NZ sector?
In contrast, both AU and PA regions show a steady decrease in the ocean stratification thereby inducing the deeper mixed-layer. What causes this weaker stratification in these regions? The mixed-layer depths in the AU and PA regions appear to oscillate on a multidecadal timescale after the AMOC collapse. Is it related to the initiation of the multidecadal oscillation in the NZ sector or is there any time lag in the multidecadal oscillation among the different sectors? Potential sources of multidecadal oscillation and their influence on the pathway to other sectors should be carefully examined and described in this paper.
5. Conclusions (Section 4)
Some of the statements are not yet examined in this study. For example, a weakening of the AMOC increases the stratification in the WGKP region, mainly due to reduced upper-layer salinities (L304-305). However, the relative contribution of temperature and salinity to the density anomalies and the potential sources of them are unexplored in this study. In the Pacific sector, the deep convection events begin to emerge only after the AMOC has collapsed (L307). It remains unclear why the deep convection events start to emerge after the AMOC has collapsed, while it does not appear before the AMOC collapse. Although the role of deep convection in driving multi-decadal oscillations in the Southern Ocean remains uncertain, the reasons for the changes in the deep convection should be clarified in this study.
Specific comments:
L31: The timescale of ENSO is much shorter than the SOM, so I am wondering if the interdecadal variability of ENSO such as the Interdecadal Pacific Oscillation (IPO; Power et al. 1999) may affect the SOM through the atmospheric teleconnection to the South Atlantic, known as the Pacific-South American pattern (PSA; Mo and Paegle 2001)
- Power, S., Casey, T., Folland, C., Colman, A., & Mehta, V. (1999). Inter-decadal modulation of the impact of ENSO on Australia. Climate Dynamics, 15(5), 319-324.
- Mo, K. C., & Paegle, J. N. (2001). The Pacific–South American modes and their downstream effects. International Journal of Climatology: A Journal of the Royal Meteorological Society, 21(10), 1211-1229.
L35-37: How do the effects of the SOM extend into the North Atlantic? Does the SOM affect the tropical Atlantic SST thereby inducing the atmospheric teleconnection to the North Atlantic or influence the ocean circulation in the tropical Atlantic and the North Atlantic, but how?
L43: The role of convection is unclear. Do you mean the Southern Ocean deep convection (south of 60S in Antarctic Sea) is modified along this cycle? I am wondering how the Southern Ocean deep convection affects the mid-latitude phenomenon of the SOM (35-50S)
L70: I think the 300-yr simulation is not sufficiently long for the ocean model to reach the equilibrium state in the Southern Ocean where the deep convection plays a key role in formation of Antarctic Bottom Water and ocean background mean state. Ideally, a few thousand years are required, but it would not be feasible to run such a long simulation using the eddy-resolving ocean model because of the computational constraint. Can the authors examine to what extent the model is spun up by checking the Southern Ocean temperature and salinity from the surface to bottom?
L76-77: Eastward propagation of the temperature anomalies from the South Atlantic to the Indian Ocean Sector along the ACC is also reported in another study by Morioka et al. (2017). However, it is unclear how these signals propagate southward and enter the Weddell Gyre. Is there any mechanism for the poleward migration?
- Morioka, Y., Taguchi, B., & Behera, S. K. (2017). Eastward propagating decadal temperature variability in the South Atlantic and Indian Oceans. Journal of Geophysical Research: Oceans, 122(7), 5611-5623.
L79: What is 60 ZJ?
L124: Why did you take the 5-yr time average? Is there any objective criterion for the length of the average? Are the results sensitive to this selection?
L133: Is the AMOC strength during the first 50 model years reasonable compared to the observation?
Figure 3: I would recommend plotting only significant values in color or with dots and adding the vertical dashed lines that denote the mean ACC latitude band for Figs. 3a-c to compare against Fig. 3d.
L162: Could the authors elaborate more on how the SOM influences the AMOC?
L212-213: In Fig. 6a, Regime A does not correspond well with this statement. The end of Regime A does not correspond to the minimum of the total kinetic energy K.
L231, 233: Please show the figures on the energetics of the Atlantic and Pacific sectors as the supplementary figures.
L254: I guess AU stands for Australia, but there is no mention in the sentence. Please specify the acronym.
L264: It looks to me that the MLD response is deepening in association with the weakened stratification, although the amplitude is smaller than the other sectors.