Page 350 - Caldera Volcanism Analysis, Modelling and Response

Magma-Chamber Geometry, Fluid Transport, Local Stresses and Rock Behaviour 325

periods, behave as elastic. Such a crustal magma chamber may thus be modelled as
a finite-size cavity or, for a two-dimensional model, a hole in an elastic host rock
when fluid, and as an elastic inclusion when solidified. These lead to different stress
concentrations since the fluid in the cavity may have zero Young’s modulus
(stiffness) whereas the inclusion stiffness is non-zero.
On solidification, the stiffness of the magma in the chamber increases until the
chamber rock and the host rock reach similar temperatures. If the chamber rock and the
host rock have similar composition, the stiffness of the solidified magma in the chamber
may then approach that of the host rock. The chamber (pluton) is then mechanically an
elastic inclusion (Eshelby, 1957). All cavities (Savin, 1961; Yu, 2000) and inclusions with
stiffnesses different from that of the host rock (Eshelby, 1957; Lekhnitskii, 1968)and
subject to loading concentrate stress and generate a local stress field (Gudmundsson,
2006). This local stress field controls the formation and slip of ring faults.
Magma chambers with volumes as great as hundreds or thousands of cubic
kilometres, and even larger for some of the volcanoes on Venus, Mars and Io, are
clearly formed over considerable periods of time. The exact mechanism by which
magma chambers form is not known, but some kind of magma traps must be
generated in order to arrest the magma and form the chamber. One possibility of
generating chambers is through stress barriers that lead to the formation of thick sills
that subsequently absorb the magma of the dykes that enter them and evolve into
chambers (Gudmundsson, 2006). The protochamber is obviously much smaller
than the mature chamber, so that it must somehow generate space for itself. The
problem of space for large plutons (or magma chambers) is an old one, but it is clear
that the space is partly generated by elastic–plastic expansion of the crust, partly by
partial melting of the crust, and partly by stoping.
During its growth, the magma chamber not only becomes larger but may also
change its shape. These two factors, that is, increase in size relative to depth below
the surface and change in shape, have strong effects on the local stress field around
the chamber and, hence, on the probability of generating collapse calderas
(Gudmundsson and Nilsen, 2006). Certain shapes of magma chambers, particularly
sill-like (oblate ellipsoidal), favour the formation of collapse calderas whereas other
shapes such as prolate ellipsoidal, do not (Figure 10).
During its evolution, and associated changes in shape and size, a particular
magma chamber may from time to time have a shape that is favourable to caldera
formation, whereas during the main part of its lifetime it may have shapes that are
unfavourable to caldera formation (Gudmundsson, 2006; Gudmundsson and
Nilsen, 2006). The evolution of a magma chamber partly explains, first, why
caldera formation or slip on an existing ring fault is such a rare event in evolution of
a volcano in comparison with the number of eruptions (Walker, 1984; Newhall and
Dzurisin, 1988) and sheet and dyke injections (Gudmundsson, 2002, 2006;
Gudmundsson and Nilsen, 2006). Second, why caldera collapse can occur
repeatedly at the same volcano while its magma chamber may change its location,
shape or both. In fact, multiple and nested calderas are very common on Earth
(Figures 11 and 12) as well as on Venus, Mars and Io (Scott and Wilson, 2000;
Mouginis-Mark and Rowland, 2001; Lopes and Gregg, 2004; Frankel, 2005), and
imply changes with time in magma chamber shape, size and, possibly, location.