Page 366 - Caldera Volcanism Analysis, Modelling and Response

Magma-Chamber Geometry, Fluid Transport, Local Stresses and Rock Behaviour 341

associated with large caldera collapses. It follows that it is of fundamental importance to
understand the processes that lead to ring-fault formation and caldera subsidence.
The general mechanics of formation of the ring faults of collapse calderas is as yet
poorly understood. Nevertheless, certain aspects are clear. For example, it is known
that the ring faults are primarily shear fractures, so that their initiation and
development must depend on the state of stress in the host rock. The state of stress in
a volcano is partly controlled by the mechanical properties of its rock units and
structures such as existing contacts, faults and joints (Figure 4). Partly, however, the
volcano stresses are controlled by the loading conditions to which the volcano is
subject, in particular, the geometry of and magma pressure in the associated chamber
and the tectonic regime (external tension, doming etc.) within which the chamber
is located. The initiation of a ring fault is thus essentially a problem in rock physics,
whereas the commonly associated fluid transport out of the magma chamber during
the subsidence of the caldera floor is a problem in geological fluid dynamics.

7.1. The underpressure model

One mechanism of the formation of collapse calderas that has been widely discussed
for many decades is underpressure in the magma chamber into which the caldera
floor eventually subsides. This mechanism, also referred to as ‘withdrawal of
magmatic support’, represents the earliest attempt to account for caldera formation
(Anderson, 1936). In Anderson’s model, the underpressure is modelled as a centre-
of-compression strain nucleus. In the extreme version of the underpressure model,
it is assumed that following an eruption there is an empty cavity, a void, that forms a
part of or perhaps the entire magma chamber into which the chamber roof and,
therefore, the caldera subsides (Williams, 1941; Scandone, 1990; Branney, 1995;
Lavallee et al., 2006). The volume of the void, and thus of the subsequently formed
caldera, is then supposed to correspond to that of the erupted and intruded
materials during the caldera-forming eruption.
This model is, of course, very appealing in its simplicity and implies that caldera
collapses are analogous to many ground subsidences, such as sink holes and pit craters.
There is, of course, little doubt that if a magma chamber could suddenly be partly or
totally emptied so as to leave a cavity in the ground, there would be surface subsidence
and, depending on the size and shape of the chamber in relation to its depth and
associated stress field, some sort of collapse. However, sink holes and pit craters are
small structures that normally reach only very shallow depths. Sink holes, for example,
are on average about 50 m in diameter and 10 m deep and are clearly related to
collapse of the roof of underground cavities such as caves (Esterbrook, 1993).
Similarly, most pit craters are related to tension fractures and normal faults at shallow
depths where absolute tension occurs; these pit craters are not related to magma
chambers (Okubo and Martel, 2001). In contrast, the magma chambers to which
most collapse calderas are related occur at depths as great as many kilometres, that is,
too great depths for absolute tension to occur except next to the chamber itself.
Several difficulties with the underpressure model of ring-fault formation are
discussed by Gudmundsson and Nilsen (2006). One problem is already discussed,
namely, how the fluid in the chamber is supposed to be driven out if the excess