Abstract
This study uses a hybrid coupled model (referred to as the general‐circulation model (GCM)/Zebiak/Cane (ZC) model), which consists of the Goddard Institute for Space Studies' (GISS) Atmospheric general‐circulation model (AGCM) coupled to the oceanic component of the ZC intermediate model to assess the impact of global warming on El Niño behavior, with and without the influence of heat introduced from the subtropical Pacific (via subtropical cell (STC) pathway). The baseline GCM/ZC model produces El Niño variability with a two year periodicity and an amplitude of approximately half the magnitude of observed El Niño. The GCM/ZC model also produces an appropriate atmospheric global response to El Niño/southern oscillation (ENSO) as shown by composites of 500 hPa heights, sea‐level pressure (SLP), 200 hPa wind, and precipitation during El Niño and La Niña periods. To evaluate the importance of global warming on ENSO variability, 2× CO 2 and 4× CO 2 transient simulations were done increasing the atmospheric CO 2 one percent per year, then extending the runs for an additional 70 years to obtain equilibrium climates for each run. An additional set of global‐warming simulations was run after including a STC parameterization generated by computing 5‐year running means of the sea‐surface temperature (SST) difference between a transient run and the 1× CO 2 GCM/ZC run at the anticipated subduction zones (160–130°W, 20–40°N and 20–44°S, 160–130°W) and adding it to the base of the equatorial mixed‐layer of the ZC model with a time lag of 15 years. This effectively alters the vertical temperature gradient of the ZC model, which affects SST via upwelling. Two features of the GCM/ZC response to global warming are emphasized. Firstly, the inclusion of the STC results in a major redistribution of heat across the equatorial Pacific, leading to an El Niño‐like response in the final equilibrium solution with less variability about the mean. The global warming aspect due to the El Niño‐like response results in a positive feedback on global warming, which causes a higher global surface‐air temperature (SAT) than identical transient simulations without inclusion of the STC. Secondly, including the STC effect produces a far greater magnitude of global ENSO‐like impact because of the reduction of, or even the reversal of, the equatorial Pacific longitudinal SST gradient. The implications of such an extreme climate scenario are discussed. Copyright © 2006 Royal Meteorological Society