Page 324 - Caldera Volcanism Analysis, Modelling and Response

Structural Development of Calderas 299

Regional tectonics may also affect the shape of the reservoir, forming elliptic
magma chambers (e.g. Bosworth et al., 2003). The role of the shape of the magma
chamber on caldera development was previously tested, with results consistent with
those previously described (Roche et al., 2000). Therefore, the shape of the
chamber (possibly) resulting from regional tectonics does not play an important role
in controlling the structure and development of experimental calderas.

4. Discussion

4.1. Consistency of the experiments
The most important similarities within this category lie in the structural development
of the calderas and the related final deformation pattern. In fact, a complete collapse
consistently gives two ring structures. The inner outward dipping reverse ring fault
develops first. The outer inward dipping to subvertical normal ring fault develops at
later stages, after a certain amount of slip of the inner ring. Each ring may be formed
by a single continuous fault, accommodating all the displacement or, more often, by
multiple, closely spaced systems of concentric ring faults or portions of them. The
amount of faults forming a ring structure may depend upon the reservoir shape.
With a flat shape, the strain accumulates at the tips of the reservoir, enhancing the
nucleation of a single pair (reverse+normal) of ring faults (Acocella et al., 2000;
Roche et al., 2000). With a domed shape, strain may accumulate also at the points of
maximum curvature of the reservoir (Mandl, 1988), generating additional ring faults
(Acocella et al., 2001). Moreover, if the contraction of the domed-shaped reservoir is
not homogeneous and migrates outwards, multiple sets of concentric reverse ring
faults may form (Marti et al., 1994; Walter and Troll, 2001; Kennedy et al., 2004).
Whether consisting of a single fault or a multiple set, the central volume delimited by
the reverse faults constitutes the sinking piston of the caldera. The inward dipping
normal faults form at the periphery of this volume, accommodating the gravitational
deformation. Similarly to the reverse faults, multiple sets of concentric ring faults
may also form (Walter and Troll, 2001; Kennedy et al., 2004). A consistent
deformation pattern, given by outward dipping reverse faults and inward dipping
normal faults, has been observed also in the experiments under a regional stress field,
with elliptical calderas (Acocella et al., 2004; Holohan et al., 2005); the only
difference lies in the variation of the dip of the reverse faults.
Another important similarity among the underpressure experiments is the
transition from diffuse to localised strain before and during the development of a
ring fault. Since the collapse is imposed at depth in the experiments, all the ring
faults nucleate at the top of the chamber analogue and propagate towards surface. In
their upward propagation, the diffuse deformation beyond the fault tip results in an
inward tilt at surface. Increased upward propagation of the fault localises the
deformation. Therefore, the presence of an inward tilt along the caldera sides
suggests a partially developed collapse, controlled by a limited subsidence.
Roche et al. (2000) suggest that the deformation pattern during collapse is
controlled by the roof aspect ratio. Lower ratios (type A; Section 3.2) are associated