A rotated EOF analysis of global sea surface temperature variability with interannual and interdecadal scales

Abstract Spatiotemporal variability of preferred global-scale sea surface temperature anomaly patterns is documented, applying a varimax-rotated empirical orthogonal function (R-EOF) analysis to monthly mean SST anomalies. The present study focuses especially on the interdecadal variability of leading R-EOF modes. It is first found that temporal variability of R-EOF1 has a quasi periodicity of 2–5 years and coincides quite well with the occurrence of the ENSO event; hence this mode can be identified with the ENSO mode and distinguished from the other modes dominated by interdecadal variability. The authors find that R-EOF2 typically shows the dominance of interdecadal variability and signals of the ENSO phenomenon are removed. This mode is characterized by increasing Indian Ocean SST and decreasing central North Pacific SST around 40°–50°N in the recent ten or more years. A further indication is that both R-EOF3 and R-EOF4, which show the dominance of interdecadal variability, are fundamentally regarded a...