The Effect of the Madden-Julian Oscillation on the Energetics and Prediction of El Niño - Southern Oscillation



Journal Title

Journal ISSN

Volume Title



Previous studies have documented extensive interactions between the El Niño-Southern Oscillation (ENSO) and the Madden-Julian Oscillation (MJO). These studies have primarily focused on interactions during individual El Niño events, and a comprehensive quantitative framework which captures the mechanisms responsible for the interactions between these sources of variability has yet to be specified. Expanding upon the known relationship between wind power and available potential energy with respect to El Niño, this study exploits the known relationship between MJO wind forcing and Kelvin wave activity to develop such a framework. A statistically significant relationship between MJO influence and Kelvin wave activity from January to July and the ultimate strength of the resultant El Niño event is found, both in an observational dataset and global climate model output. Associated with these stronger events are stronger wind power, available potential energy, and Kelvin wave development. This coupled atmosphere-ocean index of MJO influence is shown to differentiate between composite events more strongly than the purely atmospheric index of MJO strength and a purely oceanic index of thermocline variability. Sensitivity experiments performed on El Niño events which are influenced by the MJO and on those which are not demonstrate that a necessary condition for MJO-ENSO interactions is a coherent phasing between those phenomena. However, these experiments also show that although MJO can have significant influence on the timing and strength of El Niño, it is not the triggering mechanism of these events. Holistic indices which capture covariability between MJO winds, Kelvin wave activity, and SST anomalies are then shown to have prediction skill with respect to ENSO. Specifically, these indices strongly outperform Niño 3.4 as an ENSO predictor when applied to CFS-v2 hindcasts of 1980-2014.