Determine the mechanism of ENSO warm–cold asymmetry

Determine whether the observed warm–cold asymmetry of the El Niño–Southern Oscillation is primarily caused by deterministic nonlinearity in the coupled ocean–atmosphere dynamics, such as a nonlinear wind response to sea-surface temperature, or by multiplicative stochastic forcing whose amplitude depends on the system state.

Background

The paper identifies ENSO’s warm–cold asymmetry as a prominent unresolved issue. Warm events in the eastern Pacific tend to be stronger than cold events, producing a positively skewed sea-surface-temperature distribution. Two competing explanations are discussed: deterministic nonlinearity in the coupled dynamics, particularly a nonlinear atmospheric wind response to sea-surface temperature, and multiplicative noise, in which the amplitude of stochastic forcing varies with the system state.

The authors simplify agent-built ENSO models to assess which mechanisms are necessary to reproduce the asymmetry. The resulting models fall into two families—linear dynamics with multiplicative noise and nonlinear dynamics with additive noise—but the benchmark cannot distinguish between them. Resolving their relative contributions would clarify the physical processes responsible for ENSO complexity and asymmetry.

References

A prominent open question concerns its asymmetry, as warm events in the eastern Pacific tend to be stronger than cold events and give sea surface temperature a positively skewed distribution. One explanation attributes this asymmetry to deterministic nonlinearity in the coupled dynamics, such as a nonlinear wind response to sea surface temperature. Another attributes it to multiplicative noise, stochastic forcing whose amplitude depends on the state, with which even a linear model reproduces the observed asymmetry. Both are physically plausible, and their relative contributions remain unresolved.

— SciExam for ENSO: Can AI Agents Build Climate Models?  (2610.10513 - Zhang et al., 7 Oct 2026) in Section 4.1, “Agent-Built Models Reenact the ENSO Mechanism Debate” (Analysis, subsection “Agent-Built Models Reenact the ENSO Mechanism Debate”)