Sévellec, Fedorov and Liu (2017) used an adjoint optimal-flux-perturbation framework and coupled climate-model experiments to argue that declining Arctic sea ice injects positive buoyancy anomalies into the North Atlantic on multidecadal timescales, weakening the Atlantic Meridional Overturning Circulation (AMOC) and possibly explaining the suggested observed slowdown and the subpolar 'warming hole', a region of relative cooling amid greenhouse-driven warming. The paper itself states that the attribution of the recent AMOC slow-down 'remains an open question in view of oceanic natural decadal variability', and its design idealizations — linearized adjoint sensitivities in a coarse-resolution (2° NEMO ORCA2) configuration, and sea-ice radiative-budget perturbations imposed at time zero and maintained — leave open whether the mechanism is robust beyond that experimental design, and whether the suggested AMOC change can be quantitatively attributed to sea-ice decline rather than to internal variability, wind forcing, aerosols, or other freshwater sources; the same authors' follow-up studies (2019, 2021) took up the robustness question as their open target. The two questions remain contested: model responses to identical observed-amplitude ice loss diverge across resolution and mean state (Li et al. 2021; Ferster et al. 2022); attribution studies variously emphasize natural variability, Asian aerosols, or subarctic freshening (Latif et al. 2022; Liu et al. 2024; Pontes & Menviel 2024); and the first observational causal analyses (2025) remain partial, fingerprint-based and method-dependent (Nichita et al. 2025; McMonigal et al. 2025). The problem is to assess quantitatively whether and to what extent the suggested AMOC slow-down is attributable to sea-ice decline rather than to competing drivers — including the possibility that the attribution is indeterminate — and to establish the sign, magnitude and transient behavior of the sea-ice-driven mechanism across models and under realistic evolving forcing.

Background

The Atlantic Meridional Overturning Circulation (AMOC) is the basin-scale circulation in which warm, salty upper-ocean water flows northward in the Atlantic, cools and densifies in the subpolar North Atlantic and Nordic Seas, sinks through deep convection (vertical mixing that forms dense deep water), and returns southward at depth; it carries the bulk of the Atlantic's northward heat transport. Sustained direct monitoring exists only since 2004 (the RAPID-MOCHA array at 26.5°N), so longer-term AMOC changes are inferred indirectly from ocean reanalyses (model reconstructions of the past ocean state constrained by observations) and from sea-surface-temperature (SST) 'fingerprints', the best known being the subpolar North Atlantic 'warming hole' — a region of relative cooling amid greenhouse-driven warming that an AMOC slowdown would produce through reduced heat transport. Arctic sea ice has declined rapidly since 1979. Retreating ice exposes the ocean to anomalous surface heat and freshwater fluxes, and the resulting positive buoyancy anomalies (lightening of surface water) can, if exported to the subpolar North Atlantic, suppress deep convection and weaken the AMOC. Sévellec, Fedorov and Liu (2017) quantified this pathway using an adjoint model (a technique in which the linearized model equations are integrated backward to compute the sensitivity of a chosen output, here AMOC strength, to distributed surface flux perturbations) to construct 'optimal' flux perturbations — the spatially patterned surface buoyancy fluxes that maximize AMOC change — finding that Arctic flux anomalies dominate the AMOC response on timescales longer than about two decades, and confirming with coupled-model experiments that perturbing the sea-ice radiative budget weakens the AMOC over roughly a century. Their conclusion that sea-ice decline 'may explain the suggested slow-down of the AMOC' rests on acknowledged idealizations: linearized sensitivities computed in the coarse-resolution (non-eddying) 2° NEMO ORCA2 ocean configuration, which cannot resolve baroclinic mesoscale eddies (rotating ocean currents of roughly tens-of-kilometre scale that redistribute heat and freshwater) or narrow boundary-current freshwater pathways; idealized perturbations (spatially uniform meltwater; sea-ice radiative-budget modifications imposed at time zero and maintained) rather than the observed, gradually evolving forcing history; and no partition against competing drivers of observed AMOC change. A key structural control is the basin-scale salt-advection feedback — the feedback by which a weaker AMOC imports less salty water into the North Atlantic, further freshening and weakening it — whose strength, often summarized by an AMOC stability indicator based on the freshwater transport the AMOC itself carries, is model-mean-state dependent. The problem formulated here is the one the 2017 study and its follow-up literature leave open, and it is a central instance of the broader, still-unresolved debate over whether the recently suggested AMOC weakening is forced or dominated by internal multidecadal variability.

Problem Statement

Determine, quantitatively where the evidence permits and with defensible uncertainty, whether and to what extent the suggested multi-decadal slow-down of the Atlantic Meridional Overturning Circulation (AMOC) is caused by Arctic sea-ice decline, as opposed to internal multidecadal ocean-atmosphere variability, buoyancy forcing driven by winds and the North Atlantic Oscillation (NAO; the leading pattern of wintertime atmospheric pressure variability over the North Atlantic), anthropogenic aerosol changes, and other freshwater inputs such as Greenland melt and Arctic runoff. The suggested slow-down is evidenced by RAPID transport measurements at 26.5°N since 2004, by proxy and sea-surface-temperature (SST)-based multi-decadal AMOC indices (spatial surface-temperature patterns used as indirect indicators of AMOC change), and by the subpolar North Atlantic 'warming hole'; its magnitude, and even its existence over a given period, are contested. In parallel, establish whether the proposed mechanism — multidecadal accumulation and southward propagation of positive buoyancy anomalies from the ice-loss region, suppressing North Atlantic deep convection (the vertical mixing that forms dense deep water) — is robust in sign, magnitude and long-term (transient, possibly compensating or recovering) behavior across model resolution and mean-state stability, and under realistic, gradually evolving forcing histories rather than idealized instantaneous perturbations. Answers may draw on multi-model transient experiments, observational analyses, causal-inference studies, or combinations thereof. Both questions — the extent of attribution and the robustness of the mechanism — are jointly required, and a complete answer gives a defensible conclusion on each: a quantitative decomposition is required only where the evidence permits, and a demonstrated indeterminacy or non-robustness answers the corresponding question. Establishing with defensible evidence that the suggested weakening is not robustly distinguishable from internal variability, that the attribution is indeterminate, or that the mechanism is not robust across models or forcing histories is an acceptable, non-failing outcome.

References

  1. Sévellec, F., Fedorov, A. V. & Liu, W. (2017). Arctic sea-ice decline weakens the Atlantic Meridional Overturning Circulation. Nature Climate Change 7(8), 604–610. doi:10.1038/nclimate3353, https://doi.org/10.1038/nclimate3353
  2. Liu, W., Fedorov, A. & Sévellec, F. (2019). The Mechanisms of the Atlantic Meridional Overturning Circulation Slowdown Induced by Arctic Sea Ice Decline. Journal of Climate 32(4), 977–996. doi:10.1175/JCLI-D-18-0231.1, https://doi.org/10.1175/JCLI-D-18-0231.1
  3. Li, H., Fedorov, A. & Liu, W. (2021). AMOC stability and diverging response to Arctic sea ice decline in two climate models. Journal of Climate 34, 5443–5460. doi:10.1175/JCLI-D-20-0572.1, https://doi.org/10.1175/JCLI-D-20-0572.1
  4. Ferster, B. S., Simon, A., Fedorov, A., Mignot, J. & Guilyardi, E. (2022). Slowdown and Recovery of the Atlantic Meridional Overturning Circulation and a Persistent North Atlantic Warming Hole Induced by Arctic Sea Ice Decline. Geophysical Research Letters 49(16), e2022GL097967. doi:10.1029/2022GL097967, https://doi.org/10.1029/2022GL097967
  5. Dai, A. (2022). Arctic amplification is the main cause of the Atlantic meridional overturning circulation weakening under large CO2 increases. Climate Dynamics 58, 3243–3259. doi:10.1007/s00382-021-06096-x, https://doi.org/10.1007/s00382-021-06096-x
  6. Latif, M., Sun, J., Visbeck, M. & Bordbar, M. H. (2022). Natural variability has dominated Atlantic Meridional Overturning Circulation since 1900. Nature Climate Change 12(5), 455–460. doi:10.1038/s41558-022-01342-4, https://doi.org/10.1038/s41558-022-01342-4
  7. Jackson, L. C., Biastoch, A., Buckley, M. W., Desbruyères, D. G., Frajka-Williams, E., Moat, B. & Robson, J. (2022). The evolution of the North Atlantic Meridional Overturning Circulation since 1980. Nature Reviews Earth & Environment 3(4), 241–254. doi:10.1038/s43017-022-00263-2, https://doi.org/10.1038/s43017-022-00263-2
  8. Weijer, W., Haine, T. W. N., Siddiqui, A. H., Cheng, W., Veneziani, M. & Kurtakoti, P. (2022). Interactions between the Arctic Mediterranean and the Atlantic Meridional Overturning Circulation: A Review. Oceanography 35(3–4), 118–127. doi:10.5670/oceanog.2022.130, https://doi.org/10.5670/oceanog.2022.130
  9. Lee, S.-K., Kim, D., Gomez, F. A., Lopez, H., Volkov, D. L., Dong, S., Lumpkin, R. & Yeager, S. (2024). A pause in the weakening of the Atlantic meridional overturning circulation since the early 2010s. Nature Communications 15, 10642. doi:10.1038/s41467-024-54903-w, https://doi.org/10.1038/s41467-024-54903-w
  10. Liu, F., Li, X., Luo, Y., Cai, W., Lu, J., Zheng, X.-T., Kang, S. M., Wang, H. & Zhou, L. (2024). Increased Asian aerosols drive a slowdown of Atlantic Meridional Overturning Circulation. Nature Communications 15, 18. doi:10.1038/s41467-023-44597-x, https://doi.org/10.1038/s41467-023-44597-x
  11. Pontes, G. M. & Menviel, L. (2024). Weakening of the Atlantic Meridional Overturning Circulation driven by subarctic freshening since the mid-twentieth century. Nature Geoscience 17(12), 1291–1298. doi:10.1038/s41561-024-01568-1, https://doi.org/10.1038/s41561-024-01568-1
  12. Nichita, D.-R., Dubet, I., Dima, M. & Ionita, M. (2025). Tracing the Observed Causal Impact of Diminishing Summer Sea-Ice Concentration on the Atlantic Meridional Overturning Circulation. Geophysical Research Letters 52(4), e2024GL113293. doi:10.1029/2024GL113293, https://doi.org/10.1029/2024GL113293
  13. McMonigal, K., Larson, S. M., Gervais, M., Klavans, J. M., He, C., Cane, M. A., Corti, S. & Bellomo, K. (2025). Fingerprints of AMOC Decline Are Sensitive to External and Mechanistic Forcing. Geophysical Research Letters 52(12), e2025GL116307. doi:10.1029/2025GL116307, https://doi.org/10.1029/2025GL116307
  14. He, C., Patrizio, C. R., Kwon, Y.-O. & Bellomo, K. (2026). North Atlantic Sea Surface Temperature Variability: Impacts, Mechanisms, and Challenges. WIREs Climate Change 17(3), e70077. doi:10.1002/wcc.70077, https://doi.org/10.1002/wcc.70077