Why the twenty-first century tropical Pacific trend pattern cannot significantly influence ENSO amplitude?

Although the climate is highly expected to change due to global warming, it is unclear whether the El Nino-Southern Oscillation (ENSO) will be more or less active in the future. One may argue that this uncertainty is due to the intrinsic uncertainties in current climate models or the strong natural long-term modulation of ENSO. Here, we propose that the global warming trend cannot significantly modify ENSO amplitude due to weak feedback between the global warming induced tropical climate change and ENSO. By analyzing Coupled Model Intercomparison Project Phase 5 and observation data, we found that the zonal dipole pattern of sea surface temperature [SST; warming in the eastern Pacific and cooling in the western Pacific or vice versa; 'Pacific zonal mode' (PZM)] is highly correlated to change in ENSO amplitude. Additionally, this PZM is commonly identified in control experiments (pre-industrial conditions), twentieth century observations, and twenty-first century scenario experiments [representative concentration pathways 4.5 and 8.5 W m(-2) (RCP 4.5, 8.5)]. PZM provides favorable conditions for the intensification of ENSO by strengthening air-sea coupling and modifying ENSO pattern. On the other hand, the twenty-first century SST trend pattern, which is different from PZM, is not favorable towards changing ENSO amplitude. Furthermore, we performed an intermediate ocean-atmosphere coupled model simulations, in which the SST trend pattern and PZM are imposed as an external anomalous heat flux or prescribed as a basic state. It was concluded that the SST trend pattern forcing insignificantly changes ENSO amplitude, and the PZM forcing intensifies ENSO amplitude.