The future in the past: Risk of more frequent extreme El Niño supported by paleoclimate data and models

The occurrence of recent extreme El Niño events in the historical record raise the question of whether anthropogenic warming is making these events more frequent. Attributing trends in the El Niño/Southern Oscillation (ENSO) phenomenon is difficult, as Rajagopalan, Lall, and Cane warned us back in the 1990s. However, is it possible to know whether ENSO is sensitive to external forcings and could change in the future driven by greenhouse warming? Models are inconclusive, mainly because their predictions cannot be verified using the historical record. We addressed these questions focusing on the Last Glacial Maximum, a past climate interval with robust paleoclimatic evidence of much weaker ENSO variability. Using model simulations, we show that past and future ENSO changes are connected to global mean temperatures via a single mechanism governed by the strength of the Walker Circulation. This mechanism is consistent with the dynamics of observed extreme El Niño events, which develop when warm pool waters expand rapidly eastward due to strongly coupled ocean currents and winds. These coupled interactions weaken under glacial conditions because a deep mixed layer makes zonal currents less responsive to wind variability. Conversely, under greenhouse warming, a shallow mixed layer makes currents more responsive to winds favoring the onset of extreme El Niño. This mechanism together with paleoclimatic evidence of a weaker glacial ENSO increase our confidence in model predictions of more frequent extreme El Niño under future warming. The outstanding source of uncertainty remaining is the response of the mean state – cementing this question’s preeminence in our field.