How Do Tropical, Northern Hemispheric, and Southern Hemispheric Volcanic Eruptions Affect ENSO Under Different Initial Ocean Conditions?

Current understanding of volcanic effects on El Nino-Southern Oscillation in terms of eruption type and initial ocean condition (IOC) remains elusive. We use last-millennium proxy reconstructions to show how volcanic impacts depend on eruption type and IOC. When the IOC is not a strong El Nino, it is likely that an El Nino will mature in the second winter following 79% (p < 0.01) of Northern Hemispheric eruptions and in the first winter following 81% (p < 0.01) of tropical and 69% of Southern Hemispheric eruptions. For a strong El Nino-IOC, no significant El Nino will occur in the first winter after any type of eruption. The eruptions need to be large enough to cause these diverse effects. Our last-millennium simulation confirms the IOC effect, except that a La Nina occurs in the first winter following most tropical eruptions due to overestimated volcanic cooling in the model. Plain Language Summary Investigation of the volcanic effect on the El Nino-Southern Oscillation is important to understanding the climate variability excited by natural forcing. For the first time, we explore the combined effects of volcanic type and initial ocean condition, which have been found crucial to understanding the El Nino-Southern Oscillation evolution solely. Based on the reconstruction analysis in this work, we have enough confidence to predict an El Nino in the second winter following Northern Hemisphere eruptions and the first winter following tropical eruptions when the initial ocean condition does not involve a strong El Nino. Otherwise a La Nina should be predicted. This reconstruction analysis provides a good reference for the evaluation of models' performance with respect to the volcano-El Nino-Southern Oscillation relationship. The inconsistence of first-winter responses after tropical eruptions between the reconstructions and simulation found in this work calls for further investigation and model improvement. Our conclusions here are based on one model, but they may still provide some guidance to the Model Intercomparison Project on the climatic response to Volcanic forcing.