Abstract
The simulated responses of the El Niño–Southern Oscillation (ENSO) to volcanic forcings are controversial, and some mechanisms of these responses are not clear. We investigate the impacts of volcanic forcing on the ENSO using a long-term simulation covering 1400–1999 as simulated by the Bergen Climate Model (BCM) and a group of simulations performed with the Community Atmosphere Model version 4.0 (CAM4) and the BCM’s ocean component Miami Isopycanic Coordinated Ocean Model (MICOM). The analysis of the long-term BCM simulation indicates that ENSO has a negative-positive-negative response to strong tropical volcanic eruptions (SVEs), which corresponds to the different stages of volcanic forcing. In the initial forcing stage, a brief and weak La Niña-like response is caused by the cooling along the west coast of the South American continent and associated enhancement of the trade winds. In the peak forcing stage, westerly wind anomalies are excited by both reduced east–west sea level pressure gradients and weakened and equatorward shifted tropical convergence zones. These westerly wind anomalies extend to the equatorial eastern Pacific, leading to an El Niño-like response. At the same time, easterly wind anomalies west of 120°E and strong cooling effects can promote a discharged thermocline state and excite an upwelling Kelvin wave in the western Pacific. In the declining forcing stage, forced by the recovered trade winds, the upwelling Kelvin wave propagates eastward and reaches the equatorial eastern Pacific. Through the Bjerknes feedback, a strong and temporally extended La Niña-like response forms. Additional CAM4 simulations suggest a more important role of the surface cooling over the Maritime Continent and surrounding ocean in shaping the westerly wind anomalies over the equatorial central-eastern Pacific and the easterly wind anomalies west of 120° E, which are key to causing the El Niño-like responses and subsequent La Niña-like responses, respectively. The MICOM sensitivity simulations confirm that SVE-induced tropical atmospheric circulation anomalies play a dominant role in regulating post-eruption ENSO evolution in the observation, while the influences of anomalous buoyance forcing (heat and freshwater fluxes) are secondary. Therefore, SVEs play an important role in modulating the ENSO evolution. Compared with proxy data, the simulated El Niño-like responses and subsequent La Niña-like responses are consistent with the reconstructed ENSO responses to SVEs. However, the simulated initial brief La Niña-like response, which is reproduced by most models, is seen in only one proxy dataset and is absent in most of the reconstructed ENSOs and those observed. The reason for this model-data mismatch will require further investigation.