A Bivariate Extension to Traditional Empirical Orthogonal Function Analysis

This paper describes the application of canonical correlations analysis to the joint analysis of global monthly mean values of 1996-1997 sea surface temperature (SST) and height (SSH) data. The SST data are considered as one set and the SSH data as another set of multivariate observations, both with 24 variables.

This type of analysis can be considered as an extension of traditional empirical orthogonal function (EOF) analysis which provides a marginal analysis of one variable over time. The motivation for using a bivariate extention stems from the fact that the two fields are interrelated as for example an increase in the SST will lead to an increase in the SSH.

The analysis clearly shows the build-up of one of the largest El Niño events on record. Also the analysis indicates a phase lag of approximately one month between the SST and SSH fields.