Identifying robust observables for AMOC early‑warning detection

Identify AMOC‑related observables that are robust and representative for early‑warning detection, i.e., observables that reliably exhibit critical slowing down and strongly project onto the critical degree of freedom near tipping in order to enable interpretable EWS.

Background

The authors demonstrate that seemingly obvious choices such as AMOC strength can perform poorly as EWS, whereas specially designed or physically motivated observables may capture critical slowing down. However, which observables are robust in practice remains unknown.

Resolving this will require integrating physical insight, dynamical‑systems analysis (e.g., edge states), and model‑based evaluation to select observables that provide reliable and interpretable early‑warning signals for AMOC tipping.