Several kinds of analysis are applied to sea surface temperature anomalies
(SSTA) (1856-1991) to determine the degree to which SSTA of opposite sign i
n the tropical North and South Atlantic occur. Antisymmetric ("dipole") con
figurations of SSTA on basin scales are not ubiquitous in the tropical Atla
ntic. Unless the data are stratified by both season and frequency, inherent
dipole behavior cannot be demonstrated. Upon removing the global El Nino-S
outhern Oscillation signal in SSTA (which is symmetric between the North an
d South Atlantic) from the data, the regions north or south of the Intertro
pical Convergence Zone have qualitatively different temporal variabilities
and are poorly correlated. Dipole configurations do occur infrequently (12-
15% of the time), but no more so than expected by chance for stochastically
independent variables. Nondipole configurations that imply significant mer
idional SSTA gradients occur much more frequently, nearly half of the time.
Cross-spectral analysis of seasonally averaged SSTA indices for the North
and South Atlantic show marginally significant coherence with antisymmetric
phase in two period bands: 8-12 years for the boreal winter-spring and 2.3
years for the boreal summer-fall. Antisymmetric coherence is optimal for a
small subregion west of Angola in the South Atlantic, with respect to SSTA
of basin scale in the tropical North Atlantic. Dipole variability, even wh
ere optimal, explains only a small fraction of the total variance in tropic
al Atlantic SSTA (< 7%).