Failure to track a stable AMOC state under rapid climate change

Abstract

The Atlantic Meridional Overturning Circulation (AMOC) is a tipping element of the climate system. The current estimate of the global warming threshold for the onset of an AMOC collapse is +4C. However, such a threshold may not be meaningful because AMOC stability depends on the rate of radiative forcing and background climate state. Here, we identify an AMOC stabilising mechanism that operates on timescales longer than present-day radiative forcing increase. Slow forcing permits coherent adjustment of surface and interior ocean properties, supported by enhanced evaporation and reduced sea-ice extent, counteracting destabilising feedbacks. This mechanism is explicitly demonstrated in a slow CO2 increase experiment (+0.5 ppm/yr), in which the AMOC remains stable up to +5.5C of global warming. By contrast, under intermediate- and high-emission scenarios, the AMOC collapses at substantially lower warming levels (+2.2C and +2.8C, respectively). Our findings demonstrate the strong radiative forcing path dependence of AMOC tipping and imply that limiting the rate of radiative forcing is critical for reducing the near-term risk of an AMOC collapse.

Figures (14)

• Figure 1: Climate model simulations with the CESM. (a): The AMOC strength (at 26$^{\circ}$N and 1,000 m depth) for the quasi-equilibrium PI hosing simulation, including four statistical equilibria. (b): The AMOC strength for the historical and three extended RCP scenarios and $\overline{F_H} = 0.45$ Sv. The inset shows the GMST anomaly compared to the period 1850 -- 1899. (c): The AMOC strength and GMST anomaly (compared to the first 50 model years) for the CO$_2$ ramp simulation. (d): The AMOC versus GMST anomaly for the simulations presented in panels b & c. The dashed lines indicate the GMST thresholds for AMOC tipping for the RCP4.5 (+2.2$^{\circ}$C) and RCP8.5 (+2.8$^{\circ}$C) scenarios vanWesten2025e. In all panels, the thin curves are yearly averages, whereas the thick curves are smoothed versions (25-year moving averages).
• Figure 2: Oceanic Responses in the Atlantic Ocean (a&b): The time-mean AMOC in depth coordinates for model years a) 1 -- 50 and b) 1701 -- 1750. The lower panel shows the meridional heat transport (MHT). (c&d): The time-mean AMOC in density coordinates for model years c) 1 -- 50 and d) 1701 -- 1750. The three curves represent the (section-averaged) depth level, whereas the 20 m depth contour is smoothed to reduce its meridional variability. (e): The zonally-averaged temperature difference between model years 1701 -- 1750 and 1 -- 50 (shading). The curves are three isotherms of model years 1 -- 50 for reference. (f): Similar to panel e, but now for the zonally-averaged salinity.
• Figure 3: Surface-forced AMOC (a): The surface-forced AMOC between 40$^{\circ}$N and 65$^{\circ}$N ($\Delta \Psi_{\mathrm{surf}}$, sinking rates) over the first 50 model years in the CO$_2$ ramp, including the temperature and salinity contributions. The AMOC at 40$^{\circ}$N in density coordinates with its maximum indicated by the dashed red line ($\sigma_2^{\mathrm{max}}$) are also shown. (b): The $\Psi_{\mathrm{NADW}}$ and $\sigma_2^{\mathrm{max}}$ (inset). (c): The spatially-averaged (40$^{\circ}$N and 65$^{\circ}$N) surface buoyancy flux, decomposed into the heat and freshwater fluxes. (d): The $\Psi_{\mathrm{NADW}}$ differences compared the first 50 model years, including the temperature and salinity contributions. The differences for the maximum AMOC strength at 40$^{\circ}$N are also shown. The time series in panels b,c,d are smoothed through a 25-year running mean (to reduce the variability).
• Figure 4: The Atlantic freshwater budget (a): The Atlantic Ocean (34$^{\circ}$S to 65$^{\circ}$N) freshwater budget for the CO$_2$ ramp simulation with the freshwater content ($\overline{W}$), freshwater convergence ($F_{\mathrm{con}}$), surface freshwater fluxes ($F_{\mathrm{surf}}$) and changes in the freshwater content ($\frac{\mathrm{d}\overline{W}}{\mathrm{d}t}$). The quantity $\overline{W}$ is split into an upper 1,000 m contribution ($\overline{W}_{1000\uparrow}$) and below 1,000 m contribution ($\overline{W}_{1000\downarrow}$) and are displayed as their differences compared to their time mean over the first 50 model years. (b): The meridional freshwater convergences (omitting the contribution by the Strait of Gibraltar) for the different freshwater transport components. (c): The salinity along 34$^{\circ}$S over the first 50 model years, where the dashed lines indicate the different water masses which are based on the meridional velocity profile (see procedure outlined in vanWesten2024b). (d): The zonally-averaged salinity at 34$^{\circ}$S over time, which are displayed as differences compared to the first 50 model years. The dashed lines are the different water masses. The arrows are indicative of the meridional velocity (right = northward, left = southward) associated with the AMOC.
• Figure 5: Salinity and freshwater flux responses (a): The depth-averaged (upper 100 m) salinity differences between model years 1701 -- 1750 and 1 -- 50. (b -- f): The surface freshwater flux differences and its decomposition between model years 1701 -- 1750 and 1 -- 50 (positive = relative surface freshening). In all panels, the circled markers indicate non-significant ($p \geq 0.05$, two-sided Welch t-test) differences, markers were not displayed when no sea ice was present (panel f).