Independent component analysis of multivariate time series: Application tothe tropical SST variability

With the aim of identifying the physical causes of variability of a given d ynamical system, the geophysical community has made an extensive use of cla ssical component extraction techniques such as principal component analysis (PCA) or rotational techniques (RT). We introduce a recently developed alg orithm based on information theory: independent component analysis (ICA). T his new technique presents two major advantages over classical methods. Fir st, it aims at extracting statistically independent components where classi cal techniques search for decorrelated components (i.e., a weaker constrain t). Second, the linear hypothesis for the mixture of components is not requ ired. In this paper, after having briefly summarized the essentials of clas sical techniques, we present the new method in the context of geophysical t ime series analysis. We then illustrate the ICA algorithm by applying it to the study of the variability of the tropical sea surface temperature (SST) , with a particular emphasis on the analysis of the links between El Nino S outhern Oscillation (ENSO) and Atlantic SST variability. The new algorithm appears to be particularly efficient in describing the complexity of the ph enomena and their various sources of variability in space and time.