PNNL

Abstract

A Markovian stochastic model is developed for studying variability and predictability of the propagation of the Madden-Julian Oscillation (MJO). This model represents the daily changes in real time multivariate MJO (RMM) indices as random functions of its current state and the background conditions. The probability distribution function from which the RMM changes are drawn is obtained using a machine learning algorithm trained to maximize MJO forecast skills using 40 years of daily observations of indices of RMM, seasonality, El-Niño Southern Oscillation (ENSO), Quasi-Biennial Oscillation (QBO) and the Indian Ocean Dipole (IOD). Skillful forecasts are obtained for lead times between 8 days and 27 days, depending on the initial and background states. Results from large ensembles of simulations by the stochastic model show that, because of monsoonal changes in the background state, MJO propagation across the Maritime Continent (MC) region is most likely to be disrupted in boreal spring and summer when MJO events propagate from favorable conditions over the Indian Ocean to unfavorable ones over the MC. Analysis of signal-to-noise ratio indicates better predictability during spring and summer especially when the MJO activity is away from the MC region.