Attribution of the Asymmetry and Its Variation in the ENSO Climate Network

The asymmetric characteristics of ENSO events have attracted widespread attention due to their complex internal dynamics and distinct impacts during El Niño and La Niña phases. Recent advances in complex network theory have provided powerful tools for diagnosing and predicting ENSO. However, whether ENSO‐related climate networks exhibit asymmetry remains unclear. This study constructs a time‐evolving climate network based on time‐delayed cross‐correlations of surface air temperature anomalies (SATA), and systematically investigates the evolution of network strength associated with El Niño and La Niña, where network strength measures the degree of the key region dynamically integrated into prominent lagged co‐variability patterns associated with ENSO evolution. Relative to SATA sequences, network strength shows clear asymmetry during the ENSO pre‐onset development phase. It increases markedly before El Niño and decreases before La Niña, highlighting its potential as an early warning indicator. Diagnostic analyses indicate that changes in covariance among nodes within the equatorial eastern Pacific are the main drivers of the network strength asymmetry. Before El Niño events, the phases of SATA sequences at maximum lagged correlation exhibit decreased coherence, reflecting intensified warm‐cold phase interactions and signalling an early rapid response of the event. Conversely, La Niña events are preceded by enhanced coherent warm anomalies, suggesting a slower and more delayed evolution path. This asymmetric behaviour is further reflected in its long‐term evolution, with a marked intensification since the 1970s, potentially linked to warming in the tropical Pacific. This study provides new insight into the physical basis of ENSO network asymmetry and offers theoretical support for improving ENSO prediction skill.