Rethinking the Ocean’s Role in the Southern Oscillation

The Southern Oscillation (SO) is usually described as the atmospheric component of the dynamically coupled El Niño–Southern Oscillation phenomenon. The contention in this work, however, is that dynamical coupling is not required to produce the SO. Simulations with atmospheric general circulation models that have varying degrees of coupling to the ocean are used to show that the SO emerges as a dominant mode of variability if the atmosphere and ocean are coupled only through heat and moisture fluxes. Herein this mode of variability is called the thermally coupled Walker (TCW) mode. It is a robust feature of simulations with atmospheric general circulation models (GCMs) coupled to simple ocean mixed layers. Despite the absence of interactive ocean dynamics in these simulations, the spatial patterns of sea level pressure, surface temperature, and precipitation variability associated with the TCW are remarkably realistic. This mode has a red spectrum indicating persistence on interannual to decadal time scales that appears to arise through an offequatorial trade wind–evaporation–surface temperature feedback and cloud shortwave radiative effects in the central Pacific. When dynamically coupled to the ocean (in fully coupled ocean–atmosphere GCMs), the main change to this mode is increased interannual variability in the eastern equatorial Pacific sea surface temperature and teleconnections in the North Pacific and equatorial Atlantic, though not all coupled GCMs simulate this effect. Despite the oversimplification due to the lack of interactive ocean dynamics, the physical mechanisms leading to the TCW should be active in the actual climate system. Moreover, the robustness and realism of the spatial patterns of this mode suggest that the physics of the TCW can explain some of the primary features of observed interannual and decadal variability in the Pacific and the associated global teleconnections.