Interannual variability in the tropical pacific as simulated in coupled ocean-atmosphere models

The space time structure of interannual sea level variability simulated with two simplified coupled ocean atmosphere models is investigated by means of Principal Oscillation Pattern (POP) analysis. Both coupled models consists of an Oceanic General Circulation Model (OGCM) of the tropical Pacific and linear atmospheric feedback. The first coupled model uses an empirical atmospheric feedback derived from data. It simulates low frequency oscillations with periods comparable to the ENSO period of a few years when driven with white noise. In the second coupled model we use a linear steady state atmosphere model. This coupled model shows quasiperodic oscillations with periods of about 16 months within a certain parameter range. It is shown that the coupling of ocean and atmosphere is an important contribution for the generation of interannual variability. In both coupled models the interannual variability appears to be linked to the propagation of equatorial waves. The results are compared to a run with the uncoupled OGCM driven with observed winds. The resulting variability patterns are similar to those in the coupled experiments. The implications of the results for ENSO prediction are discussed.