A metric for quantifying El Nino pattern diversity with implications for ENSO-mean state interaction

摘要

Recent research on the El Nino-Southern Oscillation (ENSO) phenomenon increasingly reveals the highly complex and diverse nature of ENSO variability. A method of quantifying ENSO spatial pattern uniqueness and diversity is presented, which enables (1) formally distinguishing between unique and canonical El Nino events, (2) testing whether historical model simulations aptly capture ENSO diversity by comparing with instrumental observations, (3) projecting future ENSO diversity using future model simulations, (4) understanding the dynamics that give rise to ENSO diversity, and (5) analyzing the associated diversity of ENSO-related atmospheric teleconnection patterns. Here we develop a framework for measuring El Nino spatial SST pattern uniqueness and diversity for a given set of El Nino events using two indices, the El Nino Pattern Uniqueness (EPU) index and El Nino Pattern Diversity (EPD) index, respectively. By applying this framework to instrumental records, we independently confirm a recent regime shift in El Nino pattern diversity with an increase in unique El Nino event sea surface temperature patterns. However, the same regime shift is not observed in historical CMIP5 model simulations; moreover, a comparison between historical and future CMIP5 model scenarios shows no robust change in future ENSO diversity. Finally, we support recent work that asserts a link between the background cooling of the eastern tropical Pacific and changes in ENSO diversity. This robust link between an eastern Pacific cooling mode and ENSO diversity is observed not only in instrumental reconstructions and reanalysis, but also in historical and future CMIP5 model simulations.