Structure and evolution of interannual variability of the tropical Pacific upper ocean temperature

Publisher Website
Google Scholar
Abstract
Yearly in situ temperature anomaly data for the period 1961–1990 are analyzed to reveal the dominant structure and evolution of interannual variability of the tropical Pacific upper ocean. We use multivariate empirical orthogonal function (EOF) analyses to detect the principal three‐dimensional structure related to El Niño. There are well‐defined subsurface thermal patterns, characterized by a prominent sea‐saw structure with opposite anomaly polarity in the equatorial and off‐equatorial tropical North Pacific regions. During an El Niño year a positive temperature anomaly is found in the eastern and central tropical upper ocean. This is accompanied by a corresponding negative anomaly at subsurface depths in the west, with a maximum at 100–150 m off the equator. Unlike sea surface temperature, whose variations are confined largely to the east with one dominant polarity, temperature variations at 50–200 m depths have a dipole pattern with out‐of‐phase oscillations in the central equatorial Pacific and in the western tropical North Pacific. A reverse pattern of these anomalies is observed during a La Niña year. Evolution between El Niño and La Niña involves a significant zonal transfer of anomaly phase across the equatorial basin as well as across the off‐equatorial tropical North Pacific, showing consistent and coherent variations from west to east, from subsurface to sea surface, and from on the equator to off the equator. This phase propagation is more evident at subsurface depths than that at the sea surface, suggesting a continual movement of anomaly pattern in succession, eastward along the equator and westward off the equator of the tropical North Pacific, with spatial inhomogeneities of zonal propagation in longitude. Our analyses present evidence of the manner in which temperature anomalies evolve at subsurface ocean depths in the tropical Pacific, thus providing an observational basis for evaluating theoretical studies and model simulations. The dynamical implication of these results is also discussed.