The ocean off Peru is not just warm: it pulses. Every few years, the winds die down, warm water creeps eastward, and fishermen are left empty-handed. This is El Niño, and its arrivals do not follow a simple schedule. Some researchers blamed random weather gusts, others sought chaos in the climate system itself. The new work points to a third path: even in a calm, 'stable' zone, the model can hide several rhythms at once, and switching between them looks like chaos.

The classic Suarez–Schopf equation describes sea surface temperature with a single number, but with a delay: today's warming depends on what happened several months ago. Such a delay works like an echo in the mountains: you shout, and the reply comes later, and if you shout in time with the echo, the sound amplifies. It was long thought that oscillations arise only when the echo is too loud, that is, under linear instability. The authors looked into the quiet zone and found hidden orbits there: closed trajectories that are not visible from the neighborhood of equilibrium but attract trajectories from afar.

Between the self-excited and hidden orbits lies a homoclinic 'figure-eight': a trajectory that leaves the equilibrium point and returns to it, but from different sides. This shape acts as a crossroads where the system can choose one of two paths. Under small periodic forcing, for example, the annual cycle of solar heating, the crossroads turns into chaos: the trajectory darts between regimes, and predicting the next step becomes impossible.

To prove this, the authors reduced the infinite-dimensional dynamics to two dimensions using an inertial manifold: a surface that attracts all solutions, like a funnel collects water. On this surface, the classical Poincaré–Bendixson theorems apply, and the qualitative picture becomes rigorous. Numerical calculations confirm: for parameters α=0.75 and τ=1.65, the self-excited orbit has a period of about 8 years, while the hidden one at τ=1.58 has a period of about 10.6 years. Both numbers fall within the observed range of real El Niño events.

This changes the view on forecasting. If El Niño's irregularity is inherent in the deterministic model itself, then even perfect observations will not yield an accurate long-term forecast: there will always be uncertainty in choosing between regimes. The practical benefit is not limited to climate: the same mathematical structures arise in ring arrays of coupled transmission lines used in electronics. Hidden rhythms lurk in the delay, and now we have a tool to see them.