The role of tropical inter-basin interactions in initiating the 2020–2023 La Niña: a hindcast-based predictability study

The recent triple-dip La Niña event from 2020 to 2023 in the equatorial Pacific, which occurred without a preceding strong El Niño, challenges the traditional understanding of multi-year La Niña dynamics governed solely by the recharge oscillator theory. This study investigates the mechanisms behind the initiation of the 2020–2023 La Niña and evaluates the predictive capabilities of a fully coupled climate model. Utilizing a multi-year climate prediction system based on the Community Earth System Model version 2 with ocean data assimilation techniques, we assess the contributions of various climate factors, including inter-basin forcings such as the Indian Ocean Dipole (IOD) and tropical Atlantic warming (TAW). Our findings reveal that, while the model underestimates the role of the IOD, the combination of positive IOD and TAW substantially influenced first-dip La Niña initiation through the eastward propagation of velocity potential anomalies. Partial ocean data assimilation experiments corroborate these results and further suggest that accurately predicting the phases of the Pacific Meridional Mode (PMM) is essential for forecasting the initiation process. Our hindcast experiment indicates that the La Niña event typically decays within one year rather than persisting, suggesting that distinct drivers, such as PMM, South Pacific, or aerosol forcing, may have influenced the second and third phases of cooling. These insights highlight the critical role of inter-basin interactions (IOD, TAW, and PMM forcings) in extending ENSO predictability and improving long-term climate forecasting capabilities.