Résumé
The western equatorial Pacific oceanic heat content (warm water volume in the west or WWVw)is the best El Niño–Southern Oscillation (ENSO) predictor beyond 1-year lead. Using observations andselected Coupled Model Intercomparison Project Phase 5 simulations, we show that a discharged WWVwinboreal fall is a better predictor of La Niña than a recharged WWVwfor El Niño 13 months later, both in terms ofoccurrence and amplitude. These results are robust when considering the heat content across the entireequatorial Pacific (WWV) at shorter lead times, including all Coupled Model Intercomparison Project Phase 5models or excluding Niño-Niña and Niña-Niño phase transitions. Suggested mechanisms for this asymmetryinclude (1) the negatively skewed WWVwdistribution with stronger discharges related to stronger windstress anomalies during El Niño and (2) the stronger positive Bjerknes feedback loop during El Niño. Thepossible role of stronger subseasonal wind variations during El Niño is also discussed.