Résumé
Westerly wind bursts (WWBs) that occur in the western tropical Pacific are believed to play an important role in the development of El Nino events. Here, following the study of Lengaigne et al. (ClimDyn 23(6): 601-620, 2004), we conduct numerical simulations in which we reexamine the response of the climate systemto an observedwind burst added to a coupled general circulation model. Two sets of twin ensemble experiments are conducted (each set has control and perturbed experiments). In the first set, the initial ocean heat content of the system is higher than the model climatology (recharged), while in the second set it is nearly normal (neutral). For the recharged state, in the absence of WWBs, a moderate El Nino with a maximum warming in the central Pacific (CP) develops in about a year. In contrast, for the neutral state, there develops a weak La Nina. However, when the WWB is imposed, the situation dramatically changes: the recharged state slides into an El Nino with a maximum warming in the eastern Pacific, while the neutral set produces a weak CP El Nino instead of previous La Nina conditions. The different response of the systemto the exact same perturbations is controlled by the initial state of the ocean and the subsequent ocean-atmosphere interactions involving the interplay between the eastward shift of the warm pool and the warming of the eastern equatorial Pacific. Consequently, the observed diversity of El Nino, including the occurrence of extreme events, may depend on stochastic atmospheric processes, modulating El Nino properties within a broad continuum.