Page 319 - Caldera Volcanism Analysis, Modelling and Response

294 Valerio Acocella

Figure 6 Experiments fromWalter andTroll, 2001. (a) Adopted apparatus; (b) undeformed
stage; (c) ¢nal stage of collapse (modi¢ed afterWalter andTroll, 2001).

Walter and Troll (2001) use a deflating sill-like shaped balloon, containing air or
water (magma chamber analogue) within a sand or flour box (upper crust analogue)

5
(Figure 6a). The length ratio between model and nature is L B10 . The role of a
conical volcanic edifice, added to the top of the sand pack, is also tested. The
evolution of the collapse (pure evacuation) experiments is consistent with the one
of Roche et al. (2000) and Acocella et al. (2000). This is given by the development
of one or more outward dipping reverse faults and, subsequently, by peripheric
inward dipping normal faults (Figure 6c). In some cases, the normal faults do not
nucleate directly from the top of the chamber, but in a shallower position. In
general, increasing the depth to the balloon increases the subsidence rate of the ring
faults and decreases their diameter. With a pre-existing volcanic edifice, the
distribution of the reverse and normal faults is influenced by the initial morphology
of the edifice: steep and irregular initial flanks result in a tilted and more structurally
complex caldera floor. Nevertheless, the overall deformation pattern is totally
consistent with the one observed without volcanic edifices. The same experiments
also suggest that repeated inflation and deflation cycles may explain the
development of piecemeal collapses, formed by radial and concentric structures,
as observed at Gran Canaria (Troll et al., 2002).