Origins of extreme climate states during the 1982-83 ENSO winter

The origin of extreme climate states during the exceptional 1982-83 El Nino event has continued to be a source of controversy and debate. On the one hand, empirical analyses of extratropical climate patterns during past El Nino events suggests the observed anomalies during 1982-83 were consistent with tropical forcing. On the other hand, the large amplitude of those anomalies have not been replicated in atmospheric general circulation mod (AGCM) simulations for that period performed as part of the Atmospheric Model Intercomparison Project (AMIP). It has recently become apparent, however, that the sea surface boundary conditions used to drive the multitude of AMIP simulations were deficient, in that at least 30% of available tropical Pacific SST observations were discarded in the analysis cycle due to excessive trimming constraints. It is shown from a reanalysis of the sea surface temperatures that the observed east equatorial Pacific waters were 1.5 degrees C warmer than original estimates. In order to address the extent to which simulations of the extratropical climate of 1982-83 are sensitive to different SST analyses of that period, a parallel suite of AGCM simulations using two SST prescriptions is performed. One set is based on the blended satellite-in situ data used also in the AMIP runs, whereas the other is based on the optimum interpolation (OI) reanalysis. A nine-member ensemble of such simulations is performed, and this is compared with observations. The model response using the original blended SSTs is shown to severely underestimate the tropical rainfall anomalies, and this contributes to the simulation of a weak extratropical response as reported earlier in the AMIP experiments. A larger, more realistic response during 1982-83 is shown to occur in an identical set of runs based on che OI SST boundary conditions, and most aspects of the observed pattern and strength of the Pacific-North American climate anomalies during that winter are reproduced in the model's ensemble mean response. Further analysis of the models' intersample variability are performed to ascertain the extent to which the observed anomalies may have been influenced by atmospheric initial conditions. It is shown from the OI runs that the observed tropical Pacific rainfall anomalies and the Southern Oscillation index were phenomena causally determined by the El Nino. Even over the Pacific-North American region, the spatial pattern of the anomalies in individual runs was highly reproducible, and several members of the OI runs achieved climate anomalies exceeding in amplitude those observed. The findings strongly indicate the important role of El Nino in determining the climate state over the Pacific-North American region during 1882-83, and various competing hypotheses are critiqued in light of these new model results.