Tipping Bifurcations in Conceptual Ocean Circulation Models

The Atlantic Meridional Overturning Circulation (AMOC) is often analyzed using low-order box models to understand tipping points. Historically, these studies focus on freshwater flux as the primary bifurcation parameter, treating the temperature gradient as a fixed restoring target. However, the erosion of the equator-to-pole temperature contrast due to polar amplification suggests that thermal forcing should be treated as a dynamic control parameter. In this study, we use Cessi’s reduced box model to map the global bifurcation structure of the thermohaline circulation. We relax the assumption of a fixed thermal background and analyze the system’s behavior under joint thermal and haline forcing. We prove the existence of a cusp bifurcation, identifying the specific geometry of pitchfork and saddle-node bifurcations that bound the stable regime. This geometric characterization reveals that thermal erosion acts as a distinct mechanism for destabilization, capable of driving the system across critical thresholds even in the absence of anomalous freshwater forcing.

💡 Research Summary

This paper revisits the classic low‑order box representation of the Atlantic Meridional Overturning Circulation (AMOC) – specifically the reduced model introduced by Paola Cessi – and extends its bifurcation analysis by treating the meridional temperature gradient as a dynamic control parameter rather than a fixed restoring target. Traditional studies have focused almost exclusively on freshwater flux (denoted p) as the bifurcation driver, assuming that the temperature difference ΔT between the northern and southern boxes relaxes instantaneously to a prescribed equilibrium θ. Under this assumption the model collapses to a one‑dimensional ordinary differential equation for the salinity difference y, whose equilibria are identified as minima of a Lyapunov‑type potential V(y). The resulting bifurcation diagram displays a classic saddle‑node (fold) transition as p is increased, which has been widely used to explain abrupt climate events such as Dansgaard‑Oeschger and Younger‑Dryas episodes.

The authors argue that ongoing polar amplification erodes the equator‑to‑pole temperature contrast, making θ itself a variable that can change on climate‑relevant timescales. By re‑introducing θ as an independent parameter, the system is described by two coupled equations for ΔT and ΔS (or their nondimensional counterparts x and y). The exchange flow Q depends quadratically on the density difference Δρ = αSΔS – αTΔT, leading to a nonlinear feedback that couples the fast temperature dynamics (controlled by a large restoring coefficient α = td/tr) with the slower salinity dynamics (controlled by a parameter μ that measures the strength of advective transport relative to diffusion).

A full two‑parameter continuation in the (p, θ) plane reveals three distinct bifurcation curves:

1. Saddle‑node curve – where the Jacobian determinant vanishes while the trace remains negative. This curve corresponds to the classic fold bifurcation that annihilates a pair of equilibria as freshwater forcing grows or as the temperature gradient weakens.

2. Pitchfork curve – where the Jacobian trace passes through zero while the determinant stays positive. Here a symmetric pair of equilibria exchange stability, a feature that only appears when θ is allowed to vary sufficiently.

3. Cusp point – the intersection of the saddle‑node and pitchfork curves. At this codimension‑two singularity, infinitesimal changes in both p and θ can destroy the multistable structure, producing a dramatic loss of resilience.

Numerical continuation (using AUTO/MatCont) shows that for large θ (strong temperature contrast) the system behaves like the traditional picture: a single saddle‑node governs the transition, and a sufficiently large freshwater pulse can trigger a rapid collapse of the overturning. As θ is reduced, the saddle‑node curve recedes and the pitchfork curve emerges, eventually dominating the bifurcation landscape. In the limit of very weak temperature contrast, the model exhibits a smooth exchange between a temperature‑dominated branch and a salinity‑dominated branch without a sharp fold; the system becomes highly sensitive to modest freshwater anomalies.

Physically, a weaker temperature gradient diminishes the stabilizing thermal restoring term (−α(ΔT − θ)), allowing the quadratic exchange term to be driven primarily by salinity differences. Consequently, even modest freshwater inputs can push the circulation across the now‑reduced barrier separating the two branches. Near the cusp, the “potential wells” associated with each stable state become shallow, indicating low resilience: a small perturbation in either p or θ can cause the system to tip irreversibly.

The paper therefore reframes AMOC tipping as a joint thermal‑haline problem. It demonstrates that thermal erosion—i.e., a gradual reduction of the meridional temperature contrast—constitutes an independent destabilizing mechanism capable of driving the circulation across critical thresholds without any anomalous freshwater forcing. This insight expands the conventional freshwater‑centric paradigm and suggests that future climate projections must incorporate the evolving temperature gradient when assessing the risk of AMOC collapse. The authors conclude that the cusp bifurcation provides a useful geometric framework for quantifying how combined thermal and haline changes reshape the stability landscape of the thermohaline circulation.