Contrasting Impacts of Slow and Fast Wind Variations on ENSO

Winds play a key role in the onset, evolution, and decay of El Niño Southern Oscillation (ENSO) events. At interannual timescales, anomalous westerlies are a key component of the Bjerknes feedback, while high-frequency wind variations in the form of Westerly Wind Events (WWEs) are considered an important triggering mechanism for El Niño events through the excitation of oceanic equatorial Kelvin waves. Despite the existence of a large body of research on the subject, several open questions remain about the sea surface temperature (SST) – wind interaction at different timescales. Specifically, how large is the deterministic wind response to SST anomalies, and how important is the influence of sub-seasonal wind variability (the “atmospheric noise”) in exciting and sustaining those anomalies? Are the characteristics of individual WWEs key factors in the ENSO triggering process, or is the low-frequency tail of the “atmospheric noise” that matters? How state dependent is the atmospheric noise? In this study, we use wind, sea surface height (SSH) and SST observations at high-temporal (daily) and spatial (0.25°) resolutions over the recent two and a half decades, combined with a linear inverse modeling (LIM) framework to address the above questions. The wind product is the recently reprocessed cross-calibrated multi-platform wind vector data set Version 2 (CCMP-V2), which combines inter-calibrated satellite data and buoy winds over the period 1988-2016; SSH anomalies are obtained from the AVISO data set, and SST data are from the National Oceanic and Atmospheric Administration Optimum Interpolation SST (NOAA-OISST) product. Our results show that sub-seasonal wind variations do excite equatorial Kelvin waves, but their contribution to SST variations in the eastern Pacific is minor relative to that of inter-annual wind variations. Area averaged wind anomalies over the western equatorial Pacific, SST anomalies over the eastern equatorial Pacific, and SSH anomalies across the entire basin are used to quantify the relative importance of the Bjerknes feedback and stochastic atmospheric noise in ENSO growth. This provides important insights into ENSO predictability and prediction.