The Atlantic thermohaline circulation (THC) can weaken abruptly when freshwater is added to the northern North Atlantic, a mechanism proposed to explain the abrupt cold episodes of the last glacial period and deglaciation, such as the Younger Dryas and the Heinrich stadials. Manabe and Stouffer (1995) showed in a coupled ocean-atmosphere model that a 1-sverdrup freshwater pulse between 50°N and 70°N for 10 years drives an abrupt weakening, fluctuation, and multi-century recovery of the THC, with cooling concentrated around the northern North Atlantic. They left open: (i) whether the THC responds comparably under true glacial boundary conditions rather than the interglacial-like state used; (ii) the quantitative sensitivity of the response to meltwater magnitude, duration, and delivery mechanism (prolonged low-flux discharge, episodic iceberg armadas); (iii) how model formulation and flux adjustment affect the transient response; and (iv) which dynamical regime — the steady inactive ('off') equilibrium or a transient weak-and-shallow overturning mode — best explains the palaeoceanographic signatures of the abrupt cold events. Later work has run glacial-boundary and transient deglacial meltwater experiments, largely eliminated flux adjustment, and intercompared model responses, while proxy reconstructions of the glacial deep Atlantic variously support near-collapse (231Pa/230Th{}^{231}\mathrm{Pa}/{}^{230}\mathrm{Th}) and persistent, reorganized North Atlantic Deep Water (multi-proxy studies in 2025-2026 that still report conflicting signals). The surviving open problem is the quantitative model-proxy reconciliation: which meltwater forcings produce which overturning regimes under glacial and deglacial conditions, and which regime and forcing history explain the observed signatures of the specific abrupt cold events.

Background

The Atlantic thermohaline circulation (THC, also called the Atlantic meridional overturning circulation, AMOC) is the large-scale flow in which warm, salty near-surface water moves northward in the Atlantic, cools and becomes dense enough to sink in the Labrador and Nordic seas, forms North Atlantic Deep Water (NADW), and returns southward at depth. Because it is driven by density differences, it is sensitive to freshwater input at the sea surface, which lowers salinity and can suppress the deep convection that feeds it. Broecker (1990) proposed that abrupt releases of meltwater from the decaying Northern Hemisphere ice sheets weakened the THC and caused the abrupt cold swings recorded in Greenland ice cores and North Atlantic sediments. The best-known deglacial example is the Younger Dryas, a roughly millennium-long return to near-glacial conditions between about 12,900 and 11,700 years ago; during the last glacial period the cold extremes coincided with Heinrich events, massive armadas of icebergs discharged from the Laurentide ice sheet that littered the North Atlantic with ice-rafted debris, and the cold intervals they mark are called Heinrich stadials.

Manabe and Stouffer (1995) tested this mechanism in a coupled ocean-atmosphere general circulation model, in which atmosphere, ocean, sea ice, and land-surface components exchange fluxes interactively, and which used prescribed time-invariant flux adjustments — empirical corrections to the surface heat and freshwater fluxes that compensate systematic model biases and prevent artificial climate drift, a standard practice at the time. Starting from the model's own interglacial-like equilibrium, they applied a freshwater flux of 1 sverdrup (1 Sv = one million cubic metres per second) uniformly between 50°N and 70°N in the North Atlantic for 10 years. The THC weakened abruptly, fluctuated, and recovered over the following centuries, with the surface cooling concentrated around the northern North Atlantic and Greenland. The authors noted that the resulting transient weak-and-shallow overturning mode — reduced deep-water formation but continued ventilation of the upper ocean — matched palaeoceanographic signatures of the Younger Dryas better than the fully inactive ('off') equilibrium that coupled models of that family could also sustain as a stable state (Manabe and Stouffer 1988), and Rahmstorf (1995) mapped the bifurcation structure of such transitions under changes in the hydrological cycle.

Four unresolved issues from that study define this problem: (i) the experiment started from an interglacial-like state, so the THC response to comparable freshwater input under true glacial boundary conditions (large continental ice sheets, lower atmospheric CO2_2, altered bathymetry and background circulation) was untested; (ii) the quantitative sensitivity of the response to the magnitude, duration, and delivery mechanism of meltwater — weaker but prolonged discharges, or episodic iceberg armadas rather than a short intense pulse — needed exploration to account for the longer cold episodes in the palaeo-record; (iii) the coarse-resolution, flux-adjusted model left the transient behaviour uncertain; and (iv) different coupled-model outcomes — steady inactive equilibria versus transient weak-and-shallow states — had to be reconciled with palaeo-proxies to establish which dynamical regime explains the observed signatures.

Later work transformed the modelling side: Manabe and Stouffer (1997) extended the experiment to 500-year discharges at high versus subtropical latitudes, finding a qualitatively similar but 4-5 times smaller THC response for subtropical delivery; Ganopolski and Rahmstorf (2001) ran a coupled model of intermediate complexity under glacial boundary conditions, finding a single stable cold glacial circulation mode with deep-water formation south of Iceland and a temporary switch-off of deep-water formation for large freshwater (iceberg-type) inputs, reproducing Heinrich- and Dansgaard-Oeschger-type behaviour (Dansgaard-Oeschger events being the abrupt millennial warming-cooling oscillations of the last glacial period seen in Greenland ice cores); Liu et al. (2009) performed the first synchronous transient simulation of the whole last deglaciation with a full coupled general circulation model, including meltwater routing scenarios (the TraCE-21ka experiment); and Stouffer et al. (2006) intercompared the THC response to idealized freshwater perturbations across the coupled models of CMIP/PMIP (the Coupled Model Intercomparison Project and the Palaeoclimate Modelling Intercomparison Project), finding large inter-model spread in the strength and the recovery of the response. On the data side, the sedimentary protactinium/thorium ratio 231Pa/230Th{}^{231}\mathrm{Pa}/{}^{230}\mathrm{Th}, a kinematic proxy for overturning rate, indicated that the overturning was nearly or completely eliminated during Heinrich event H1 about 17,500 years ago and weakened into the Younger Dryas (McManus et al. 2004), whereas recent multi-proxy reconstructions — combining proxies such as the benthic carbon-isotope ratio δ13\delta^{13}C (a nutrient and ventilation tracer), neodymium isotopes (a water-mass source tracer), benthic Mg/Ca temperatures and δ18\delta^{18}O (water-mass temperature and salinity), and authigenic 231Pa/230Th{}^{231}\mathrm{Pa}/{}^{230}\mathrm{Th} — find that NADW remained widespread, though with shifted properties, during the Last Glacial Maximum and Heinrich Stadial 1 (Blaser et al. 2025; Wharton et al. 2026), and explicitly report the state of the glacial deep North Atlantic as still uncertain and the proxy reconstructions as conflicting.

Problem Statement

Determine, quantitatively and under realistic glacial and deglacial boundary conditions (continental ice sheets, lower atmospheric CO2_2, altered bathymetry and background circulation), how the Atlantic thermohaline circulation responds to meltwater forcing of different magnitudes, durations, and delivery mechanisms — short intense pulses, prolonged low-flux discharges, and episodic iceberg-discharge (Heinrich-type) events, delivered at high latitudes or further south — and which dynamical regime of the overturning each such forcing produces in coupled climate models: a fully inactive ('off') equilibrium, a transient weak-and-shallow mode with reduced deep-water formation but continued upper-ocean ventilation, or a reorganized but active deep-water formation mode. Then reconcile these modelled regimes with the palaeoceanographic proxy record — including 231Pa/230Th{}^{231}\mathrm{Pa}/{}^{230}\mathrm{Th}, benthic δ13\delta^{13}C, neodymium isotopes, and benthic Mg/Ca-temperature and δ18\delta^{18}O constraints — to establish which regime, and which meltwater forcing history, best explains the specific observed signatures of the abrupt cold episodes of the last glacial period and the last deglaciation, in particular the Younger Dryas and the Heinrich stadials. This is the model-proxy reconciliation question left open by Manabe and Stouffer (1995). The question of model-structure dependence (resolution, flux handling, convection and mixing parameterization) is part of the problem to the extent that it controls which regimes coupled models can produce and hence which reconciliations are admissible.

The problem is joint across its parts: it spans the magnitudes, durations, delivery mechanisms, and delivery locations named above rather than a single meltwater case; assigns one of the three named dynamical regimes to each such forcing; reconciles each of the named proxies with the modelled regimes for the Younger Dryas and for the Heinrich stadials; and states how model-structure dependence restricts which regimes, and hence which reconciliations, are admissible. The evidence may come either from a synthesis of published coupled-model experiments and proxy reconstructions or from new simulations.

References

  1. Manabe, S. and Stouffer, R. J., 'Simulation of abrupt climate change induced by freshwater input to the North Atlantic Ocean', Nature 378, 165-167 (1995). DOI: 10.1038/378165a0. URL: https://doi.org/10.1038/378165a0
  2. Broecker, W. S., 'Salinity history of the northern Atlantic during the last deglaciation', Paleoceanography 5, 459-467 (1990). DOI: 10.1029/PA005i004p00459. URL: https://doi.org/10.1029/PA005i004p00459
  3. Manabe, S. and Stouffer, R. J., 'Two Stable Equilibria of a Coupled Ocean-Atmosphere Model', Journal of Climate 1, 841-866 (1988). DOI: 10.1175/1520-0442(1988)001<0841:TSEOAC>2.0.CO;2. URL: https://doi.org/10.1175/1520-0442(1988)001%3C0841:TSEOAC%3E2.0.CO%3B2
  4. Rahmstorf, S., 'Bifurcations of the Atlantic thermohaline circulation in response to changes in the hydrological cycle', Nature 378, 145-149 (1995). DOI: 10.1038/378145a0. URL: https://doi.org/10.1038/378145a0
  5. Manabe, S. and Stouffer, R. J., 'Coupled ocean-atmosphere model response to freshwater input: Comparison to Younger Dryas Event', Paleoceanography 12, 321-336 (1997). DOI: 10.1029/96PA03932. URL: https://doi.org/10.1029/96PA03932
  6. Ganopolski, A. and Rahmstorf, S., 'Rapid changes of glacial climate simulated in a coupled climate model', Nature 409, 153-158 (2001). DOI: 10.1038/35051500. URL: https://doi.org/10.1038/35051500
  7. Stouffer, R. J., Yin, J., Gregory, J. M., Dixon, K. W., Spelman, M. J., Hurlin, W., Weaver, A. J., Eby, M., Flato, G. M., Hasumi, H., Hu, A., Jungclaus, J. H., Kamenkovich, I. V., Levermann, A., Montoya, M., Murakami, S., Nawrath, S., Oka, A., Peltier, W. R., Robitaille, D. Y., Sokolov, A., Vettoretti, G. and Weber, S. L., 'Investigating the Causes of the Response of the Thermohaline Circulation to Past and Future Climate Changes', Journal of Climate 19, 1365-1387 (2006). DOI: 10.1175/JCLI3689.1. URL: https://doi.org/10.1175/JCLI3689.1
  8. Liu, Z., Otto-Bliesner, B. L., He, F., Brady, E. C., Tomas, R., Clark, P. U., Carlson, A. E., Lynch-Stieglitz, J., Curry, W., Brook, E., Erickson, D., Jacob, R., Kutzbach, J. and Cheng, J., 'Transient Simulation of Last Deglaciation with a New Mechanism for Bølling-Allerød Warming', Science 325, 310-314 (2009). DOI: 10.1126/science.1171041. URL: https://doi.org/10.1126/science.1171041
  9. McManus, J. F., Francois, R., Gherardi, J.-M., Keigwin, L. D. and Brown-Leger, S., 'Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial climate changes', Nature 428, 834-837 (2004). DOI: 10.1038/nature02494. URL: https://doi.org/10.1038/nature02494
  10. Blaser, P., Waelbroeck, C., Thornalley, D. J. R., Lippold, J., Pöppelmeier, F., Kaboth-Bahr, S., Repschläger, J. and Jaccard, S. L., 'Prevalent North Atlantic Deep Water during the Last Glacial Maximum and Heinrich Stadial 1', Nature Geoscience 18, 410-416 (2025). DOI: 10.1038/s41561-025-01685-5. URL: https://doi.org/10.1038/s41561-025-01685-5
  11. Wharton, J. H., Kozikowska, E., Keigwin, L. D., Marchitto, T. M., Maslin, M. A., Ziegler, M. and Thornalley, D. J. R., 'Relatively warm deep-water formation persisted in the Last Glacial Maximum', Nature 650, 116-122 (2026). DOI: 10.1038/s41586-025-10012-2. URL: https://doi.org/10.1038/s41586-025-10012-2