The Randomness of Extreme El Niño Events

The mechanisms and predictability of extreme El Niño events, such as those in 1982, 1997 or 2015, remain unclear. Here we systematically investigate the combined role of ocean heat recharge and westerly wind bursts (WWBs) in the equatorial Pacific in the generation of extreme El Niño events. In particular, we investigate the role of early, middle and late season westerly wind bursts in El Niño development in comparison with the role of initial ocean heat recharge. To that end, we conducted multiple sets of 100-member ensemble simulations using a coupled model, NCAR CESM1.2. Each set was initialized with different upper ocean heat content, a key precursor of El Niño development. Despite conventional thinking, our results show the early-season WWBs do not provide a good indication of the El Niño state at the end of the same year, regardless of the initial upper ocean heat content. The occurrence of early-season WWBs is also not a good indicator for the subsequent WWBs that may appear as part of the Bjerknes feedback during the rest of the year. This is especially true for moderately charged ocean states. While a strongly charged ocean state is likely to develop into an El Niño, possibly extreme, its magnitude will depend on the early season WWBs only weakly. These model findings generally agree with reanalysis data, with the caveat of a much shorter time span of the latter. Thus, this study confirms that even after strong early-season WWBs with ocean heat recharge the occurrence of an extreme El Niño still has limited predictability.