And also presenting a poster the last few evenings on my most recent paper looking at AMOC tipping using quasipotentials and exploring the idea of a geometric early warning

I had a wonderful time convening a session yesterday on Early Warning Signals: from Detection to Action. Some really excellent talks and discussion on such an important topic
@gsiexeter.bsky.social @exeter.ac.uk

And a fun time presenting some of my PhD work to a new audience

Finally, we speculate on the effects of combining noise and transient forcing, where the geometric signal needs to be understood in a probabilistic sense.

We also explore this model in the presence of noise by numerically computing quasipotentials that characterize the noise-induced transitions between different attractors.

We apply this to a simple but data-adapted model of the Atlantic Meridional Overturning Circulation to highlight cases of large transient forcing where one may be able to predict the future state in advance, using this geometric early warning signal.

We introduce a geometric signal that enables one to understand behavior near an edge state on a threshold that lies between two basins of attraction.

Transient forcing of a nonlinear dynamical system can give rise to very different future stable behaviors for small changes in forcing. In this paper, we explore the potential to find early warnings of parameter-dependent tipping between such different behaviors for such small changes in forcing.

I am convening a session with @paulritchie.bsky.social (University of Exeter) and Florian Diekert (University of Augsburg) on ‘Early Warning Signals for Tipping Points: From Detection to Action’. We welcome all abstract submissions and look forward to an interesting discussion around this topic.

The call for abstracts and registration for the Global Tipping Points Conference 2025 (30 June – 3 July, Exeter, UK) are now open.
global-tipping-points.org/conference-2...
The deadline for abstract submission is Wednesday 19th March.

Thanks to my collaborators on this work, Peter Ashwin, @jonbaker303.bsky.social and Richard Wood. @uniofexeter.bsky.social @metoffice.bsky.social

In this study, we quantify noise for a conceptual AMOC model from CMIP6 experiments. We find that the noise on relevant time and spatial scales is small and unlikely to induce a tipping of this model unless combined with some freshwater forcing.