A method for constructing skillful seasonal forecasts using slow modes of climate variability

Abstract

A methodology for constructing skillful statistical seasonal forecasts of climate fields is described and applied to predict the Southern Hemisphere summer mean sea level pressure anomalies for the period 1993--2004. The method employs a recently developed variance decomposition approach, which allows a separation of the predictable and unpredictable components of climate variation. The proposed forecast scheme is based on finding predictors for the amplitude time series of the dominant slow (or predictable) modes of interannual climate variation.