Abstract
A number of recent observational and theoretical studies indicate that under certain conditions tropical sea surface temperature anomalies may be capable of producing climatic anomalies at extratropical latitudes. According to the hypothesis put forth in these studies, perturbations in the climatological mean distribution of precipitation in the tropics can influence the extratropical circulation through the action of forced, quasi-stationary, two-dimensional, Rossby wavetrains, which may tend to excite the fastest growing normal mode associated with barotropic instability.
Results of previous GCM experiments designed to simulate the atmospheric response to SST anomalies are re-examined in light of this hypothesis and are found to be generally consistent with it.
A modeling investigation consisting of three separate GCM experiments was carried out using the GLAS climate model with January initial conditions based on observed data and an equatorial Pacific sea surface temperature anomaly based on the recent analysis of Rasmusson and Carpenter (1982).
The observed eastward shift of the belt of heavy convective precipitation in the western Pacific during the episodes of positive sea surface temperature anomalies was correctly simulated in all three experiments. Associated with this, the ascending branch of the Walker circulation also shifted eastward and the north-south overturning intensified in the central Pacific. The 300 mb height difference field, averaged for the three pairs of experiments, showed evidence of two-dimensional Rossby wave propagation along a great circle path, poleward over the North Pacific and eastward across the North Pacific, in agreement with observations and with the results of experiments with simpler models. However, there was considerably variability from experiment to experiment and there were unexpected features over Eurasia in the average difference pattern. As in previous experiments, the simulations with sea surface temperature anomaly produced an increase of low-level westerlies along the equator, slightly to the west of the enhanced rainfall. It was the moisture convergence, associated with the anomalous low-level circulation rather than the local changes in evaporation, which accounted for most of the simulated changes in precipitation. The simulated circulation changes at extratropical latitudes exhibited an equivalent barotropic vertical structure, in agreement with observations. The sea level pressure associated with warm sea surface temperature anomalies was lower over the eastern Pacific and higher over the western Pacific and the Indian Ocean in agreement with the observed Southern Oscillation.
