FENG

Volcanic effects from large eruptions on climate manifest for a few years both at earth’s surface as a cooling and in the stratosphere as a warming. Such impacts lie in principle among the potential predictable features of climate after a volcanic eruption took place. Previous works based on models and observations also suggest an increased probability for a positive phase of the North Atlantic Oscillation (NAO) during the first winter following stratospheric tropical eruptions. Large gaps remain however in our understanding of the climate’s response to volcanic eruptions, related to the paucity of observed volcanic events during instrumental era and the degree to which such response depends on the characteristics of the eruption (i.e., hemispheric loading, strength, season) or initial conditions of the climate system when the eruption occurs. Tackling all these factors is crucial to improve our understanding of the underlying physical mechanisms and eventually assess the potential risks associated with future large volcanic eruptions.

This PhD work aims at exploring the above mentioned issues with the IPSL-CM6A-LR model as part of framed PMIP4 and VolMIP standardized CMIP6 coupled model experiments designed to systematically tackle specific uncertainty factors. The first part of the thesis is devoted to characterising the simulated NAO signal in winters following stratospheric volcanic eruptions using three long transient simulations of the last millennium (500-1849 CE). The uncertainties related to the season, strength and the latitude of the eruptions were also explored. The results reveal that the model simulates increased warming in the stratosphere, a stronger polar vortex and a surface pattern similar to the positive phase of the NAO in the first winter following a summer or a winter tropical eruption, with polar night jet responses amplitudes that are linearly related to the eruption strength. No linear relationship with the eruption magnitude is identified for extra-tropical Northern Hemisphere events while a tendency towards a positive NAO phase is only significant during the same winter as the eruption occurrence.

To further assess the physical mechanisms and the robust responses related to sampling climate background conditions, the second part of the study uses a suite of 25-members ensemble simulations for the well observed Mt Pinatubo tropical eruption (Philippines, June 1991). Ensemble members all start from predefined initial conditions sampling the full ENSO cycle and use the same volcanic forcing dataset following the VolMIP protocol. Our experimental protocol also comprises sensitivity simulations designed to separate the influence of volcanically-induced surface cooling and stratospheric heating on key observed features during the first two winters following Pinatubo. Results indicate that our model experiments and observations both have a positive NAO-like pattern in the first winter after the eruptions. The model also simulates realistic levels of warming in the stratosphere while, both mean sea-level pressure and temperature anomalies display similar patterns as in observations in the first winter. Sensitivity experiments indicate that the surface positive NAO signature in our model experiments is primarily attributable to stratospheric heating in the lower tropical stratosphere which generates stronger subtropical zonal winds through the thermal wind balance and accelerates the polar vortex.