Forecasting of the Southern Oscillation Index

Abstract

The possibility of the Southern Oscillation Index (SOI) forecast on the basis of the artificial neural networks model is studied in the paper. The ensemble of global climatic indices from 1950 to 2010 were used as the inputs. The model verification was done by the control period of 2007–2015. The possibility of successful SOI forecasting in 1–5 months advance for the late autumn-early winter and for early spring was demonstrated. Taking into account that two types of El-Nino and La-Nina exist, the primary concern of April and May SOI forecast for the eastern type events has been shown.