Modeling El Niño southern oscillation index using time series forecasting

Abstract

Southern Oscillation Index (SOI) is measured as the difference in air pressure across the Pacific Ocean, between Tahiti in the south-east and Darwin in the west. SOI is one of the indices that are often used to analyze and predict the changes in El Niño Southern Oscillation phenomenon. Many statistical models have been developed using SOI indices in forecasting. The objective of this study is to find the best method among the Box-Jenkins Autoregressive Integrated Moving Average (ARIMA), Single Exponential Smoothing and Double Exponential Smoothing in forecasting the monthly SOI. SOI data from January 1990 to December 2015 with a total of 25 years were employed in this study. Akaike Information Criterion (AIC) and the Sum of Square Error (SSE) were used as goodness of fit test in selecting the best model. The result indicated that the Box-Jenkins ARIMA is a suitable method in forecasting SOI values compared to others based on the smallest SSE.