The space and time features of global SST anomalies studied by complex principal component analysis

In this paper, the variability characteristics of the global field of sea s urface temperature (SST) anomaly are studied by complex principal component (c.p.c.) analysis, whose results are also compared with those of real p.c. analysis. The data consist of 40 years of global SST monthly averages over latitudes from 42.5 degrees S to 67.5 degrees N. In the spatial domain, it is found that the distribution of the first complex loading amplitude is c haracterized by three areas of large values: the first one in the eastern a nd central equatorial Pacific Ocean, the second one in the northern tropica l Indian Ocean and South China Sea, the third one in the northern Pacific O cean. As it will be explained, this pattern may be considered as representa tive of El Nino mode. The first complex loading phase pattern shows a stati onary wave in the Pacific (also revealed by real p.c. analysis) superimpose d to an oscillating disturbance, propagating from the Pacific to Indian or the opposite way. A subsequent correlation analysis among different spatial points allows revealing disturbances actually propagating westward from th e Pacific to the Indian Ocean, which could therefore represent reflected Ro ssby waves, i.e. the west phase of the signals that propagate disturbances of thermal structure in the tropical Pacific Ocean. In the time domain, a r elation between the trend of the first complex principal component and the ENSO cycle is also established.