Volcanoes are often associated with vertical collapse, due to deeper magma withdrawal. Calderas are the most notable type of vertical collapse, on the summit of volcanoes. Caldera collapse has been observed and monitored only at Miyakejima (Japan; 2000), Dolomieu (Reunion; 2007) and Fernandina (Galapagos; 1968), highlighting our limited knowledge on its kinematic behavior. Here we use experimental models to investigate the kinematic evolution of calderas and vertical collapses in general. We extract velocity and strain fields using the Particle Image Velocimetry (PIV) technique, generating time series. Experimental vertical collapses undergoing constant subsidence velocity show three main kinematic behaviors: (1) continuous collapse, whose velocity is similar to the source subsidence velocity; (2) incremental collapse, with episodic (stepwise) accelerations along pre‐existing ring structures; (3) sudden collapse, resulting from the upward migration of a cavity, only for T/D > 2 (T and D are the depth and width of the magma chamber, respectively) and without ring structures. The velocity in the collapsing column may increase up to four orders of magnitude with regard to the constant subsidence velocity of the source. Comparison to nature suggests that: (1) there are close kinematic similarities with monitored collapse calderas, explaining their incremental subsidence after the development of ring structures; (2) sudden pit crater formation is induced by the upward propagation of cavities, due to magma removal at depth and in absence of ring structures; (3) all these types of vertical collapses have a consistent mechanism of formation and kinematic behavior, function of T/D and the presence/absence of ring structures.
The 2014 and 2015 eruptions of Kuchinoerabujima Volcano followed a ~15-year precursory activation of the hydrothermal system induced by a magma intrusion event. Continuous heat transfer from the degassing magma body heated the hydrothermal system and the increase of the fluid pressure in the hydrothermal system caused fracturing of the unstable edifice, inducing a phreatic explosion. The 2014 eruption occurred from two fissures that traced the eruption fissures formed in the 1931 eruption. The explosive eruption detonated the hydrothermally-altered materials and part of the intruding magma. The rise of fumarolic activities before the past two activities in 1931-35 and 1966-1980 also suggest activation of the hydrothermal system by magmatic intrusions prior to the eruption. The long-lasting precursory activities in Kuchinoerabujima suggest complex processes of the heat transfer from the magma to the hydrothermal system.
We estimated time scales of magma-mixing processes just prior to the 2011 sub-Plinian eruptions of Shinmoedake volcano to investigate the mechanisms of the triggering processes of these eruptions. The sequence of these eruptions serves as an ideal example to investigate eruption mechanisms because the available geophysical and petrological observations can be combined for interpretation of magmatic processes. The eruptive products were mainly phenocryst-rich (28 vol%) andesitic pumice (SiO2 57 wt%) with a small amount of more silicic pumice (SiO2 62–63 wt%) and banded pumice. These pumices were formed by mixing of low-temperature mushy silicic magma (dacite) and high-temperature mafic magma (basalt or basaltic andesite). We calculated the time scales on the basis of zoning analysis of magnetite phenocrysts and diffusion calculations, and we compared the derived time scales with those of volcanic inflation/deflation observations. The magnetite data revealed that a significant mixing process (mixing I) occurred 0.4 to 3 days before the eruptions (pre-eruptive mixing) and likely triggered the eruptions. This mixing process was not accompanied by significant crustal deformation, indicating that the process was not accompanied by a significant change in volume of the magma chamber. We propose magmatic overturn or melt accumulation within the magma chamber as a possible process. A subordinate mixing process (mixing II) also occurred only several hours before the eruptions, likely during magma ascent (syn-eruptive mixing). However, we interpret mafic injection to have begun more than several tens of days prior to mixing I, likely occurring with the beginning of the inflation (December 2009). The injection did not instantaneously cause an eruption but could have resulted in stable stratified magma layers to form a hybrid andesitic magma (mobile layer). This hybrid andesite then formed the main eruptive component of the 2011 eruptions of Shinmoedake.
