Using Network Theory and Machine Learning to predict El Niño

Abstract

This article aims at improving the El Niño Southern Oscillation (ENSO) prediction, which is currently not reliable more than six months ahead. Topological properties of Climate Networks of both a Zebiak–Cane-type model and observations are described and a hybrid model is introduced for ENSO prediction. This hybrid model combines Autoregressive Integrated Moving Average and an Artificial Neural Network. The predictions of the hybrid model improve the CFSv2 ensemble prediction by the National Centers for Environmental Prediction (NCEP), for predictions up to six months ahead. Moreover, the addition of a network variable as input of the prediction model, results in a twelve month lead time prediction with a comparable skill to the shorter lead time predictions.