Intraseasonal westerly wind events (WWEs), which are believed to be a precursor of El Niño, are generally associated with atmospheric noise. In this paper, Hilbert–Huang transform (HHT)/Ensemble Empirical Mode Decomposition (EEMD) have been applied to illustrate that WWEs may not be considered as stochastic noise. Amplitude modulation is the result of the multiplicative interaction between atmospheric noise, intraseasonal, annual, and interannual signals. Multiplicative interaction needs to be invoked in order to explain the phenomenon, observed in previous studies, constituted by WWEs with the amplitude of 7 m/s, lasting 5-20 days. The decomposition of WWEs, obtained by using EEMD, provides 12 Intrinsic Mode Functions (IMFs). We exclude the first six IMFs that are associated with high-frequency noise and unrelated to El Niño, as well as the higher rank IMFs that are associated with decadal and multi-decadal signals and, hence unrelated to WWEs. We focus on signals with periods between the seasonal (100-240 days, IMFs7-8) and the interannual time-scales, and examine their underlying physical processes.  

This approach foresees that WWEs are initiated by the intraseasonal signal, associated with variability in the atmospheric field over the extratropical from the northern Pacific. The WWEs trigger an eastward propagating and downwelling Kelvin wave, which results in SST warming in the equatorial central‐eastern Pacific, such warming supports WWEs to last two-three month after initiated from the extratropics. Results found using our approach is compared with results obtained with Fourier Filtering, which is commonly used to extract WWEs but with periods between 100 and 250 days. Finally, we discuss the different roles of the seasonal and annual WWEs forcing on El Niño-Southern Oscillation (ENSO) dynamics.