The El Niño-Southern Oscillation (ENSO) contributes to interannual variability, and an important question is whether changes would occur in the intensity or frequency of ENSO events as a result of global warming [ Trenberth, 1990]. El Niño is roughly defined as warming of equatorial surface waters of the eastern Pacific [ Tziperman et al., 1994]. Although the two are not always coupled [ Trenberth and Shea, 1987], the Southern Oscillation (SO) is El Niño's atmospheric counterpart [ Jin et al., 1994]. The ENSO phenomenon is thus a good example of the close interaction between atmospheric and oceanic dynamics [ Philander, 1990]. While the pattern of El Niño is essentially cyclic, it evidences much irregularity [ Jin et al., 1994]. The SO exhibits quasi-periodicity as well, with cycles typically varying from three to six years since 1950, although this portion of the record may not be representative of long-term behavior [ Trenberth and Shea, 1987]. The past record of ENSO events reveals that the intensity and frequency of ENSO events has changed, with strong events 1880-1920, 1939-1942, 1954-1955, and 1982-1983 (among others) [ Trenberth, 1990].
Because ENSO events have a strong influence on regional and
global temperature and rainfall [ Cole et al., 1993;
Trenberth and Hurrell, 1994; Keeling and Whorf, in press]
and may influence atmospheric CO
concentrations [ Winguth
et al., 1994], the reproduction of ENSO-like events is viewed as
a critical test of general circulation model (GCM) performance
[ Meehl, 1990]. Until GCMs are capable of simulating such
an influential feature of global climate, their ability to
project future climates is limited. With the advent of
synchronously coupled atmosphere-ocean GCMs in the late 1980s,
reproduction of some features associated with ENSO has been
possible [ Meehl, 1990; Campbell et al., in press].