El Niño Southern Oscillation (ENSO) is the dominant mode of global climate variability at interannual timescales. Given its irregular occurrence at intervals of 2-7 years, few El Niño events have been directly observed. Each El Niño event has unique, physical characteristics and individual events often show substantial spatio-temporal differences in ocean and atmosphere conditions. The range of these varying characteristics and the limited number of observed events hampers generalisations about ENSO properties and limits our ability to sufficiently constrain multi-year variability in climate models. El Niño complexity can be simplified into two distinct warming patterns differentiated by the location of strongest SST warming in either the central or eastern tropical Pacific. The recognition of the diverse characteristics of ENSO events has led to a shift in the traditional view of tropical coupled ocean-atmosphere systems and presents a challenge in light of recent changes and possible future ENSO behaviour.
We investigate a number of climate models taking part in the Coupled Model Intercomparison Project Phase 5 (CMIP5) and identify models that simulate spatially distinct El Niño behaviour. Identification of El Niño events reveals a lack of model agreement about projected changes of El Niño diversity. The probability of infrequent El Niño characteristics such as a prolonged El Niño episode of 2014-2016 is evaluated. This recent episode of El Niño like conditions raised many questions about the predictability of El Niño events, underlying low-frequency variability and our general knowledge about the ENSO system and diversity within it. We use CMIP5 simulations to estimate the probability of these rare conditions. Our analysis points towards an under-representation of central Pacific events that are followed by eastern Pacific events Comparisons with future simulations indicate several possible changes in El Niño transitions.