Using surface marine wind and sea surface temperature data from the period 1950-1987, together with sea surface temperature and sea level pressure data from several stations in the Pacific, we have identified two dominant time scales of El Nino-Southern Oscillation (ENSO) variability. One is a biennial mode, with periods near 24 months, the other a lower frequency concentration of variance in periods of four to five years. A particularly well defined standing biennial component of ENSO variability, tightly phase-locked with the annual cycle, appears in the surface wind field of the equatorial eastern Indian Ocean. This is part of a larger scale biennial circulation over the low latitude eastern Indian Ocean-western Pacific sector. This biennial circulation is a fundamental element of ENSO variability. It exhibits a strong tendency for phase locking with the annual cycle and introduces a degree of regularity into the ENSO cycle. However, it varies in amplitude from cycle to cycle and sometimes changes phase. "Warm" and "cold" episodes are not unrelated events, but reflect opposite phases of the biennial cycle, modified by the low-frequency mode. Variations in ENSO swings can be formally described in terms of relative phasing and amplitude of the biennial and low frequency components. While the low-frequency mode typically exhibits about the same amplitude as the biennial fluctuations, it does not seem to exhibit the features associated with "ENSO episodes", which are captured by the biennial mode. The problem of understanding biennial variability seems to be intimately linked with an understanding of the mean annual cycle. The observational results point to the annual cycle as a fundamental pacemaker of the ENSO cycle. They strongly suggest that coupled mode theory must be developed in the context of the annual cycle. They also indicate that the ENSO cycle cannot be fully described and understood without consideration of the entire Indian Ocean Pacific Ocean sector and thus raise questions regarding the limitations of modeling the tropical Pacific as an isolated coupled ocean/atmospheric system.
