How are large western hemisphere warm pools formed?

During the boreal summer the Western Hemisphere warm pool (WHWP) stretches from the eastern North Pacific to the tropical North Atlantic and is a key feature of the climate of the Americas and Africa. In the summers following nine El Niño events during 1950–2000, there have been five instances of extraordinarily large warm pools averaging about twice the climatological annual size. These large warm pools have induced a strengthened divergent circulation aloft and have been associated with rainfall anomalies throughout the western hemisphere tropics and subtropics and with more frequent hurricanes. However, following four other El Niño events large warm pools did not develop, such that the mere existence of El Niño during the boreal winter does not provide the basis for predicting an anomalously large warm pool the following summer.In this paper, we find consistency with the hypothesis that large warm pools result from an anomalous divergent circulation forced by sea surface temperature (SST) anomalies in the Pacific, the so-called atmospheric bridge. We also find significant explanations for why large warm pools do not always develop. If the El Niño event ends early in the eastern Pacific, the Pacific warm anomaly lacks the persistence needed to force the atmospheric bridge and the Atlantic portion of the warm pool remains normal. If SST anomalies in the eastern Pacific do not last much beyond February of the following year, then the eastern North Pacific portion of the warm pool remains normal. The overall strength of the Pacific El Niño does not appear to be a critical factor. We also find that when conditions favor a developing atmospheric bridge and the winter atmosphere over the North Atlantic conforms to a negative North Atlantic Oscillation (NAO) pattern (as in 1957–58 and 1968–69), the forcing is reinforced and the warm pool is stronger. On the other hand, if a positive NAO pattern develops the warm pool may remain normal even if other circumstances favor the atmospheric bridge, as in 1991–92. Finally, we could find little evidence that interactions internal to the tropical Atlantic are likely to mitigate for or against the formation of the largest warm pools, although they may affect smaller warm pool fluctuations or the warm pool persistence.