Glacial influence on caldera-forming eruptions

Abstract

It has been suggested that deglaciations have influenced volcanism in several areas around the world increasing productivity of mantle melting and eruptions from crustal magma chambers. However, the connection between glaciations and increased volcanism is not straightforward. Investigation of Ar–Ar, U–Pb, and 14C ages of caldera-forming eruptions for the past million years in the glaciated arc of Kamchatka has lead to the observation that the majority of large-volume ignimbrites, which are associated with the morphologically preserved calderas, correspond in time with “maximum glacial” conditions for the past several glacial cycles. In the field, the main proof is related to the fact that glaciated multi-caldera volcanoes hosted thick glacial ice caps. Additional evidence comes from clustering Kamchatka-derived marine ash layers with glacial moraines in DSDP cores. Here we present a set of new results from numerical modelling using the Finite Element Method that investigate how the glacial load dynamic may affect the conditions for ring-fault formation in such glaciated multi-caldera volcanoes. Different scenarios were simulated by varying: (1) the thickness and asymmetric distribution of the existing ice cap, (2) the depth and size of the magmatic reservoir responsible for the subsequent collapse event, (3) the thickness and mechanical properties of the roof rock due to the alteration by hydrothermal fluids, (4) the existence of a deeper and wider magmatic reservoir and (5) possible gravitational failure triggered, in part, by subglacial rock mass build up and hydrothermal alteration. The results obtained indicate that: (1) Any ice cap plays against ring fault formation; (2) Asymmetric distribution of ice may favour the initiation of trap-door type collapse calderas; (3) Glacial erosion of part of volcanic edifice or interglacial edifice failure may facilitate subsequent ring fault formation; (4) hydrothermal system under an ice cap may lead to a quite effective hydrothermal rotting of the intracaldera roof rocks and hence to variations of their mechanical properties and inhibit/deflect ring fracture propagation; and (5) rock surface topography/load influenced by glacial erosion and ice volume change during the interstadials. Although, the analysis of the stress field may inform us about the possibility of ring-fracture initiation, it does not ensure its complete propagation. Parameters controlling this phenomenon are also discussed here. Overall, the maximal glacial time represent the most dynamic time in a multi-caldera volcano life (as compared to more quiet interglacial) promoting physical and chemical feedbacks. We consider that brief interstadial periods during maximal glacial creates most favourable conditions for initiation of caldera-forming eruption, largely through its influence on surface topography by glacial action, mass wasting, and influencing magma vesiculation/discharge as a function of rapidly changing overload.