Eastern equatorial Pacific SST variability:: ENSO and non-ENSO components and their climatic associations

Using an updated Kaplan et al. global SST anomaly (SSTA) dataset (1870-1999) a canonical representation of El Nino-Southern Oscillation (ENSO) is constructed. When this canonical ENSO is subtracted from the data, a residual (non-ENSO) dataset for SSTA is left that includes interseasonal to multidecadal variability. Over the eastern equatorial Pacific (Nino-3) the canonical ENSO accounts for about 79% of the total SSTA variability, and the residual, dominated by decadal timescales, accounts for the rest. In particular, about 40%-50% of the amplitudes of the strong 1982-83 and 1997-98 El Nino events were accounted for by the residual variability. The non-ENSO variability is characterized by the known shift from cold to warm eastern tropical Pacific in the mid-to late 1970s as well as by a (nonstationary) interannual variance increase during the 1980s and 1990s. Composite maps of surface (SST, sea level pressure, and winds) and tropospheric (divergent winds, velocity potential, and vertical velocity) variables are used to compare the spatial patterns characterizing the canonical ENSO and the residual components of the Nino-3 variability. The residual composites are found to only share large-amplitude fluctuations of SST anomalies in the equatorial Pacific east of the date line. When these composites are separated into decadal and interannual components, the decadal part closely resembles the structure of the Pacific decadal oscillation. The major patterns of tropospheric variability associated with the ENSO and decadal non-ENSO components are very different. At low latitudes, they imply nearly opposite impacts on farfield regional climates, based on their respective warming (or cooling) phases within the Nino-3 region. This unexpected result for low-latitude climate associations runs contrary to the naive expectation (recently shown to be true for North America) that a decadally warm tropical east Pacific will reinforce the climate effects associated with ENSO alone. This indicates that, in the Tropics, climate outlooks may be more accurate if based on separately analyzed relationships between these SSTA components and their associated climate fluctuations.