Modelling the effects of ice-sheet activity on CO2 outgassing by Icelandic volcanoes

Glacial cycles may play a significant role in mediating the flux of magmatic CO2 between the Earth's mantle and atmosphere. In Iceland, it is thought that late-Pleistocene deglaciation led to a significant volcanic pulse, evidenced by increased post-glacial lava volumes and changes in melt chemistry consistent with depressurization. Investigating the extent to which glacial activity may have affected volcanic CO2 emissions from Iceland, and crucially over what timescale, requires detailed knowledge of how the magma system responded to the growth and collapse of the ice-sheet before and after the LGM. To investigate this, we coupled a model of magma generation and transport with a history of ice-sheet activity. Our results show that the emplacement and removal of the LGM ice-sheet likely led to two significant pulses of magmatic CO2. The first, and most significant of these, is associated with ice-sheet growth and occurs as the magma system recovers from glacial loading. This recovery happens from the base of the melting region upwards, producing a pulse of CO2 rich magma that is predicted to reach the surface around 20 ka after the loading event, close in time to the LGM. The second peak in CO2 output occurs abruptly following deglaciation as a consequence of increased rates of melt generation and transport in the shallow mantle. Although these post-glacial melts are relatively depleted in CO2, the increase in magma flux leads to a short-lived period of elevated CO2 emissions. Our results therefore suggest a negative feedback, whereby ice-sheet growth produces a delayed pulse of magmatic CO2, which, in addition to increased geothermal heat flux, may contribute towards driving deglaciation, which itself then causes further magmatism and CO2 outgassing. This model is consistent with the seismic structure of the asthenosphere below Iceland, and the established compositional and volumetric trends for sub- and post-glacial volcanism in Iceland. These trends show that the earliest subglacial events involved small volumes of enriched melts, while eruptions that were synchronous with or immediately followed deglaciation involved larger volumes of more depleted melts.