A method for disentangling El Nino-mean state interaction

The amplitude of the El Nino-Southern Oscillation (ENSO) is known to fluctuate in long records derived from observations and general circulation models (GCMs), even when driven by constant external forcings. This involves an interaction between the ENSO cycle and the background mean state, which affects the climatological precipitation over the eastern equatorial Pacific. The changes in climatological rainfall may be ascribed to several factors: changes in mean sea surface temperature (SST), changes in SST variability, and changes in the sensitivity of precipitation to SST. We propose a method to separate these effects in model ensembles. A case study with a single GCM demonstrates that the method works well, and suggests that each factor plays a role in changing mean precipitation. Applying the method to 16 pre-industrial control simulations archived in the Coupled Model Intercomparison Project phase 5 (CMIP5) reveals that the inter-model diversity in mean precipitation arises mostly from differences in the mean SST and atmospheric sensitivity to SST, rather than from differences in ENSO amplitude. Citation: Watanabe, M., and A. T. Wittenberg (2012), A method for disentangling El Nino-mean state interaction, Geophys. Res. Lett., 39, L14702, doi:10.1029/2012GL052013.