Chaotic Kramers’ Law: Hasselmann’s Program and AMOC Tipping

J. Deser, R. Römer, N. Boers, C.Kuehn · 2025

Read Chaotic Kramers’ Law: Hasselmann’s Program and AMOC Tipping by J. Deser, R. Römer, N. Boers, C.Kuehn (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-18904.txt.

Introduction

Jakob Deser and co-authors extend a classic law of sudden jumps to systems pushed by chaos. Many systems rest in one of two stable states until noise knocks them over. Kramers' law ties the wait for such a jump to the strength of the noise. Real pushes, the team argues, are often fast chaos rather than pure noise.

They show a like law holds when the push is chaotic, provided the push is strong enough. Below that strength, no jumps occur at all. Their test case is a stripped model of the Atlantic overturning current, a system with real tipping fears. The law held even far from the pure noise limit. The authors note limits and computing hurdles in wide time scale gaps.

Catalog record

CREATORJ. Deser, R. Römer, N. Boers, C.Kuehn
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-18904

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Chaotic Kramers’ Law: Hasselmann’s Program and AMOC Tipping J. Deser, R. Römer, N. Boers, C.Kuehn Abstract In bistable dynamical systems driven by Wiener processes, the widely used Kramers’ law relates the strength of the noise forcing to the average time it takes to see a noise-induced transition from one attractor to the other. We extend this law to bistable systems forced by fast chaotic dynamics, which we argue is in some cases a more realistic modeling approach than unbounded noise forcing. Transitions similar to the noise-driven case can only occur if the amplitude of the chaotic forcing is large enough. If this is the case, in our numerical example - a reduced-order model of the Atlantic Meridional Overturning Circulation (AMOC) - we observe the chaotic Kramers’ law to hold even when the chaotic forcing is far from the stochastic limit. We discuss the limitations of the chaotic Kramers’ law, how to address the numerical issues associated with the timescale separation, and give a possible explanation for the dynamics of recently found AMOC collapses and recoveries in complex climate models. Preprint: AIP/123-QED A common approach when modeling dynamical systems is to separate a system into slow and fast components, and approximate the fast and often chaotic dynamics by a Wiener process that acts as forcing on the slow dynamics. One can ask what happens if the fast dynamics are not quite fast enough for this stochastic approximation to hold, but when, instead, chaotic forcing would be more realistic. For such cases, we derive the chaotic Kramers’ law that relates the strength of the chaotic forcing to tipping times between two attractors of the forced slow dynamics. We explain key assumptions for the chaotic Kramers’ law to hold, its surprisingly wide applicability, and its limitations while illustrating and interpreting our findings in a chaotically forced reduced-order AMOC model. I Introduction The Hasselmann program from the 1970s [ 29 ] is a widely used modeling approach in climate science, and similar techniques are commonly used in many other fields [ 12 , 39 ] . The program proposed to model fast, often chaotic, fluctuations as noise, forcing the slower dynamics. This approach allows one to efficiently model and understand the slow dynamics and many large-scale processes, which would prove difficult if one tried to resolve all small-scale, fast processes in detail. However, in many cases, it is possible - and if one is interested in long or many model runs, even necessary - to describe natural processes with reduced-order models. In these simpler models, it can be feasible to also resolve the faster dynamics, which leads to the question of whether, in these models, a stochastic approximation of the fast dynamics is still a good and justified approach. Instead, one could, for example, use some bounded and fast chaotic forcing that acts on the slower dynamics. These two types of forcing can lead to even qualitatively different results: A multi-stable system driven by sufficiently generic unbounded (e.g., Gaussian) noise will almost surely undergo noise-induced tipping in finite time, i.e., if initialized in one attractor of the unforced system, it will transition to another attractor as a result of the noise realization upon waiting for long enough time. In contrast, a system forced by small-amplitude bounded chaos might only tip for certain values of the forced system’s parameters, giving rise to a “chaotic tipping window” [ 5 , 45 ] . Thus, although fast and small fluctuations may sometimes seem negligible, the choice of how to approximate them is essential. Despite the differences between stochastic and chaotic forcing, homogenization theory shows that, in the limit of infinite time-scale separation and given some assumptions, chaotic forcing converges to stochastic forcing [ 41 , 31 , 28 ] . Close to this limit, one can try to extend results from one case to the other; in this way, we derive the “chaotic Kramers’ law” which generalizes the well-known Kramers’ law from the case of systems driven by stochastic forcing to systems driven by fast chaotic forcing of large-enough amplitude. We check this result in an example system and discuss how to numerically handle the large timescale separation in the chaotically forced system. This example suggests that the chaotic Kramers’ law holds far beyond the stochastic limit, even when increments of the chaotically forced system are clearly not normally distributed. The system that we use as an example is a chaotically forced reduced-order model [ 55 ] of one of the most important potential climate tipping elements, the Atlantic meridional overturning circulation (AMOC). A good understanding of the AMOC is a pressing problem as it strongly affects weather and climate on a global scale. A potential collapse of the AMOC would lead to substantial cooling in Europe and rainfall reductions in tropical African and Asian monsoon systems [ 9 ] . Observational data suggest declining AMOC strength since the mid-twentieth century after a stable period [ 21 ] , but the question of whether the AMOC will undergo an abrupt or large transition (i.e. tipping) due to climate change, and if so on which timescale, remains subject of ongoing research [ 15 , 11 , 10 , 7 , 42 ] . Studies with highly complex general circulation models (GCMs) can be used to assess the response of the AMOC to anthropogenic greenhouse gas forcing or to artificial hosing, where freshwater is added to the North Atlantic, mimicking meltwater influx from the Arctic [ 44 , 40 ] . However, GCMs are computationally expensive and, due to their complexity, sometimes difficult to interpret, which calls for an identification of the most important underlying dynamics using reduced-order models. Due to their computational efficiency, these simpler models can also be used to run larger ensembles or to explore areas in phase space that would be too costly to explore with GCMs. We believe an improved understanding of chaotic forcing, as discussed in this paper, to be a valuable approach to creating useful reduced-order models capturing real-world mechanisms. Our findings indicate that seemingly stochastic transitions could also be observed in purely deterministic GCMs and thus support the use of both deterministic and stochastic AMOC models in future research. The paper is organized as follows. We give an overview of the theory needed to derive and interpret the chaotic Kramers’ law in Section II by summarizing results from homogenization theory and discussing the known Kramers’ law in the case of stochastically forced bistable systems. Then, we generalize this to the case of chaotically forced systems in Section II.2.1 . We consider a reduced-order AMOC model [ 55 , 1 , 23 ] in which we replace the stochastic with chaotic forcing in Section III and discuss how the fast chaotic forcing can be connected to stochastic forcing in a numerically stable way. Then, we use this model to test the chaotic Kramers’ law numerically in Section III.2 and interpret the results in the context of other studies on the AMOC. We end with a general discussion in Section IV . II Chaotic limits and asymptotic transition times II.1 Homogenization In this section, we present one specific result from the theory of averaging and homogenization of ODEs. The general idea of homogenization

Read the complete text