In three parallel experiments, an atmospheric general circulation model has been subjected to observed, monthly varying sea surface temperature (SSI) conditions in each of the following domains: near-global ocean (GOGA run), tropical Pacific (TOGA run), and midlatitude North Pacific (MOGA run). Four independent realizations were obtained for the model response to the sequence of SST anomalies during the 1946–88 period in each of the above regions.
The principal modes of coupling between the imposed SST forcing and the simulated Northern Hemisphere wintertime 5 1 5-mb height field in various experiments have been identified using a singular value decomposition (SVD) procedure. The leading SVD mode for the GOGA experiment is Qualitatively similar to that based on observational data, although the amplitudes of the simulated height anomalies are notably lower than the observed value. The SST pattern of this mode resembles that associated with El Niño events. The accompanying 5 1 5-mb height anomaly is dominated by a wavelike pattern in the North Pacific/North American sector. The TOGA experiment reproduces many of the atmosphere-ocean relationships discerned from the GOGA output. Conversely, the MOGA run yields a much weaker and less reproducible response. The contrast between the TOGA and MOGA runs is indicative of the primacy of tropical Pacific SST anomalies in forcing the midlatitude atmospheric circulation.
In the TOGA experiment, the remote atmospheric responses to tropical Pacific SST anomalies influence the energy exchange across the local air-sea interface. and could thereby perturb the SST field outside of the tropical Pacific. Through this “atmospheric bridge,” the tropical Pacific could set the pace for variability of the global ocean. Analysis of the TOGA output indicates that, over the North Pacific, changes in the surface energy fluxes are mainly determined by the surface wind speed and by the strength of temperature and moisture advection. Over the Indian Ocean, variations in the incident solar radiation due to changes in cloud cover also affect the surface fluxes. The worldwide SST tendencies inferred from the variations in surface fluxes simulated in the TOGA experiment are in good agreement with the local observed SST anomalies.