Abstract
A nonlinear time series forecasting scheme developed by Sugihara and May has been applied to the Southern Oscillation index. Although forecast skill is comparable only to persistence or linear (autoregressive) methods, the scheme has the advantage of identifying close analogs to the current situation, if these exist. The operational implementation of the scheme in the Seasonal Climate Outlook issued by the National Climate Centre of the Australian Bureau of Meteorology is described and its performance during the 1991/92 El Niño–Southern Oscillation event is examined.
