El Niño-Southern Oscillation (ENSO) is the strongest interannual climate signal, affecting weather worldwide. Its phases (El Niño and La Niña) are associated with weakening or strengthening of trade winds and upwelling off the coast of Peru. There is a long-standing debate: is ENSO a stochastic system excited by noise, or a nonlinear oscillatory system with internal instability? The work by Ma and Wang belongs to a line of research using simplified coupled ocean-atmosphere models to identify essential features of ENSO dynamics. Unlike studies focused on statistical or linear approaches, here they apply the rigorous theory of dynamic transitions, previously successfully applied to phase transitions in superfluid helium. This allows not only describing observed variability but also deriving the mechanism from first principles of fluid dynamics.
The foundation is the theory of dynamic transitions developed by the authors earlier. They consider the Boussinesq equations for atmospheric circulation over the equator, restricted to latitude θ=0 (meridional velocity is zero). Turbulent friction is accounted for through scale-dependent coefficients. In the idealized case (uniform surface temperature), the problem has translational symmetry, and the bifurcation leads to a continuum of steady solutions, a circle S¹. When a small inhomogeneity in surface temperature is added (a realistic condition), the symmetry is broken, and the bifurcation becomes saddle-node (type III) or continuous (type I) depending on the parameter b. The analysis shows that under typical conditions, two stable steady states emerge, corresponding to El Niño and La Niña phases, with basins of attraction separated by an unstable manifold. For realistic parameters (troposphere height 8 km, physical constants of air), the critical Rayleigh number is estimated as Rc ≈ 2.7×10²¹, which corresponds to a temperature difference of about 60°C, consistent with observations.
Theorem 4.1 establishes that under natural boundary conditions, the system undergoes a type III (mixed) transition with a saddle-node bifurcation at R* < Rc. In the interval R* < R < Rc, there are two stable states: normal (weak) and anomalous (La Niña). For R > Rc, two stable states appear, corresponding to El Niño and La Niña. Each state has its own basin of attraction, and the boundaries between them are narrow. The ENSO mechanism is described as the interaction of two oscillatory processes: (1) transitions between metastable states (El Niño, normal, La Niña) and (2) spatiotemporal variations in sea surface temperature (SST), caused by solar heating and upwelling. SST modulates the atmospheric Rayleigh number: during El Niño, SST rises, the Rayleigh number increases, bringing the system closer to the critical value Rc and making a transition to La Niña more likely. During La Niña, enhanced upwelling cools the SST, the Rayleigh number falls, and the basin of attraction of La Niña narrows, favoring a transition back. Randomness arises from uncertainty in initial data near narrow basins of attraction, while determinism comes from the predictable dynamics of the Rayleigh number and SST. Thus, ENSO appears as a self-exciting system with natural feedback that generates sporadic oscillations.
This work offers a new perspective on the nature of ENSO, unifying stochastic and deterministic features within a single mathematical structure. Unlike neighboring studies, where ENSO is modeled as a linear system excited by noise or as a chaotic oscillator, here the mechanism is derived from bifurcation analysis of the Walker circulation. This is the first rigorous mathematical demonstration that metastable states and narrow basins of attraction can explain the sporadic nature of ENSO without invoking external stochastic forcing. This approach could be extended to other climate modes, such as the thermohaline circulation, and contribute to more accurate predictive models.
The proposed mechanism opens the way to building forecasting schemes based on estimating the system's current position relative to basins of attraction. If these basins can be identified from observational data, the probability of transitions between phases could be assessed. Further development may include accounting for a more complete spatial structure of SST and interaction with other basins, as well as applying the theory to paleoclimate data to test the universality of the mechanism. It may also refine the role of random perturbations near critical Rayleigh numbers.
The results could influence climatology, physical oceanography, and dynamical systems theory. They also have implications for forecasting ENSO and associated extreme weather events.
The theory's predictions need to be compared with observational data, such as the Southern Oscillation Index, and the realism of the basins of attraction assessed. The sensitivity of the mechanism to parameters should also be investigated, and a more detailed description of ocean dynamics included.
This work directly addresses the fundamental problem of predictability of climate variability. The question of whether ENSO is stochastic or deterministic remains open, and this theory offers a resolution through the concept of metastable states. Furthermore, it touches on the general problem of the emergence of low-frequency variability in geophysical flows, where the interaction of fast and slow processes generates complex dynamics.
The name 'El Niño' (Spanish for 'the Christ child') originated with Peruvian fishermen, who noticed that a warm current sometimes appeared off the coast during the Christmas season. The phenomenon has been known since the 19th century, but its global significance was only realized in the late 20th century.
The idea of metastable states and sudden transitions between them echoes the concept of 'climate tipping points' in science fiction, such as in Kim Stanley Robinson's novel 'Forty Signs of Rain', where abrupt climate changes occur when critical thresholds are crossed.