Spatial and frequency dependence of westerly wind events causing ENSO diversity

Abstract

<p>The El Niño Southern Oscillation (ENSO) exhibits a large diversity of events characterized by the location of extreme sea surface temperature anomalies either in the Eastern Pacific (EP) or Central Pacific (CP). While broadband stochastic wind forcing is one of the known drivers of ENSO, the relative influence of its high- and low-frequency component on ENSO diversity remains unclear. </p><p>We conduct a spectral analysis of westerly wind anomalies, yielding high- and low-frequency wind components for six different regions in the equatorial Pacific. The influence of these high- and low-frequency westerly wind anomalies, as well as their spatial location, is used for predicting ENSO diversity at different lead times. Using causal network discovery combined with multiple linear and non-linear regression analyses, we obtain different predictors for the different types of ENSO. Our results identify causally relevant, spatial and frequency-band restricted westerly wind anomalies at different lead times, which might improve early forecasting of El Niño events and improve our understanding of stochastic wind forcing energizing ENSO. </p>