Consistency of square‑root AMOC scaling with absent critical slowing down

Investigate whether the approximately square‑root scaling of the mean Atlantic Meridional Overturning Circulation (AMOC) strength with freshwater forcing in the Community Earth System Model can be consistent with the observation that the same AMOC‑strength observable does not exhibit critical slowing down; determine the theoretical and model conditions under which both behaviors can coexist.

Background

Early‑warning signals for tipping points often rely on the saddle‑node normal form, which predicts a characteristic square‑root scaling of the system’s mean state with the control parameter and critical slowing down (CSD) in fluctuations. A recent study argued that AMOC strength in a CMIP6‑class model scales approximately with the square root of freshwater forcing, suggestive of saddle‑node behavior.

However, the present paper emphasizes that some observables, including AMOC strength, may not display CSD even near collapse. This raises a direct consistency question: can the same observable simultaneously show square‑root mean‑state scaling while lacking detectable CSD? Answering this requires clarifying theoretical and modeling conditions and possibly reinterpreting the role of chosen observables.