Abstract
Recent advances in computer power and climate modeling capability have provided the opportunity for several modeling groups to undertake extended integrations with global coupled ocean-atmosphere climate models that allow the study of coupled processes thought to be important in producing Southern Oscillation and El Niño phenomena. Results are shown here from such a coupled model, developed at the National Center for Atmospheric Research (NCAR), that consists of a global, spectral (R15) atmospheric general circulation model (GCM) coupled to a global, 5° latitude-longitude, four-layer, ocean GCM. In spite of limitations of the coarse model grid, Southern Oscillation-type interannual variability of the ocean-atmosphere system is inherent in the coupled model. One of the mysteries of the Southern Oscillation cycle is how the system makes the transition from cold to warm phase and back again in the tropical Pacific in northern spring. Evidence is shown from the NCAR coupled model that a modulation of the mean seasonal cycle in the eastern Pacific drives the initiation and decay of warm and cold episodes in that region. The mechanism of this forcing in the model involves coupled seasonal anomalies of sea-level pressure (SLP), sea-surface temperature, surface wind stress, ocean upwelling, and convection-precipitation. These coupled anomalies form as a consequence of land-sea contrast in the eastern Pacific, in association with the evolution of seasonally low SLP over South America during northern winter and its movement with the seasonal cycle of solar forcing northwestward during northern spring. The anomalies become established farther west in the tropical Pacific as the year progresses and are associated with global patterns that resemble, in some ways, the phenomena observed with warm and cold events–the extremes of the Southern Oscillation. Similar sets of coupled processes exist in the observed long-term mean seasonal cycle, and the interannual events in the eastern tropical Pacific are manifested as a modulation of the mean seasonal cycle in the observations analogous to the coupled model. A reduction of coupling strength in the model (by reducing the strength of the wind-stress forcing from the atmosphere) eliminates both seasonal dependence and interannual anomaly signals Turning off the seasonal cycle of solar forcing in the model changes the nature and regular evolution of the warm and cold events. Since the model fails to simulate any of the observed phenomena in the western Pacific, it is likely that only one of several possible sets of mechanisms involved with the observed El Niño-Southern Oscillation is simulated in the present global coupled model.