Is the Pacific's response to global warming La Nina-like or El Nino-like?

El Niño-Southern Oscillation (ENSO) is a dominant mode of natural oscillation of the equatorial coupled atmosphere-ocean system in the Pacific. The question of whether the equatorial Pacific responds to radiative heating in a La Niña-like (cold-ENSO) pattern or an El Niño-like (warm-ENSO) pattern is under debate in the context of global warming (see Vecchi et al. [2008] ). It has been argued that because of the tight coupling of the atmospheric Walker circulation with the thermocline depth in the eastern equatorial Pacific Ocean, the response to a larger radiative heating may not necessarily be a warmer sea-surface temperature (SST). There are currently two competing theories, differing in the degree with which the atmosphere is coupled to the ocean. Clement et al. [1996] presumed that the eastern Pacific SST is controlled by ocean cold water upwelling and therefore a basin-wide heating increases only the SST in the western Pacific. The resulting east-west temperature gradient strengthens the atmospheric Walker circulation, whose easterly flow near the surface induces stronger ocean upwelling in the eastern Pacific, thus a cold-ENSO-like response. On the other hand, Held and Soden [2006] and Vecchi and Soden [2007] suggested that tropical circulations, especially zonal overturning circulations (such as the Walker Circulation) would weaken in a warmer climate. The weakened surface easterlies lead to an El Niño (warm-ENSO)-like SST, of a warm tongue in the eastern Pacific. Held and Soden [2006] pointed out that this is a robust response of the current crop of coupled atmosphere-ocean general circulation models: As the SST warms, convection actually decreases, because the lower tropospheric water vapor increases faster than the global mean precipitation. Xie et al. [2010] suggested that the warming pattern should be less El Niño-like because of the strengthened southeasterlies south of the equator, due to hemispheric asymmetry in land-sea area (Liu et al. [2005] ). Observational evidence is ambiguous. Vecchi et al. [2008] showed that the trend in 1880-2005 has a cold-ENSO-like pattern in one dataset (HadISST) but a warm-ENSO-like pattern with asymmetry in another (ERSST). Karnauskas et al. [2009] found the zonal SST gradient strengthened in boreal fall but weakened in spring. We hope to reconcile these disparate results in the present study.