Page 28 - Caldera Volcanism Analysis, Modelling and Response

Residence Times of Silicic Magmas Associated with Calderas 3

volcanism is the topographical depressions left after the eruption. These are thought
to be the result of either tremendous explosions that blew apart a pre-existing
volcanic cone or due to subsidence of the roof of the reservoir after or during
magma evacuation. High-level magma emplacement (typically o10 km depth)
seems to be required for caldera formation, but when combined with the apparent
large size of some reservoirs, questions arise as to the thermal and mechanical states
of the crust and the magma, the rates and mechanisms of vapour- and silica-rich
magma differentiation, and the timescales of transport and storage of huge
quantities of eruptible silicic magma (e.g., Smith, 1979; Hildreth, 1981; Shaw 1985;
Jellinek and DePaolo, 2003). These issues were addressed by Shaw (1985) who
noted that ‘‘the interaction of magma generation rates, stress domains and injection
rates leads to a spectrum of residence times which effectively determine the
types of intrusive and volcanic suites seen at high crustal levels and at the surface.’’
Almost 25 years later, progress in analytical techniques have enabled the
quantification of the time over which crystals and magma are stored before a
caldera-forming eruption. This allows analysing the relations between the volumes,
compositions, temperatures and depths of magma reservoirs below calderas from a
new perspective. The purpose of this manuscript is to describe the approaches used
to obtain the time scales of magmatic processes, to compile the data on residence
times of major caldera-related complexes, and to use this information for deriving
modes and rates of silica-rich magma production and storage in the Earth’s crust.

1.1. What is the residence time of a magma?
It can be defined as the time elapsed since the magma was formed and its eruption.
Uncertainties arise with the meaning of ‘when’ a magma is formed because what
is finally erupted is a mixture of phases that might have very different origins in
time and space (e.g., Bacon and Lowenstern, 2005). The most widespread use of
residence time involves pinpointing when a given mineral started to crystallise,
presumably during storage in a magma reservoir. This is different from the
definition used in oceanic geochemistry or in highly active volcanic systems where
it refers to the (mean) time that a given element or isotope spends in a reservoir
before being removed (e.g., Holland, 1978; Albare `de, 1993). In practice, one can
calculate the residence time as the difference between the eruption age as obtained
39
by K–Ar (or 40 Ar/ Ar), (U, Th)/He and 14 C methods (for prehistorical eruptions)
and the age provided by other radioactive clocks, such as Rb–Sr, and U–Th–Pb.
From this definition it is apparent that the residence time does not need to be a
single value, and might depend on the phases and radioactive isotopes that are used.
Multiple values of residence times may arise from different crystallisation ages of
different minerals, but also from the fact that the very definition of an age requires
knowledge of when the radioactive system became closed. This condition depends
on several factors but strongly on the diffusion rate of the daughter isotope, and
has been quantified with the use of a closure temperature (Dodson, 1973). This
explains the a priori paradoxical situation that, for example, a sanidine might
have two different ages and both could be correct: dated by the K–Ar system the
mineral gives the eruption age but using Rb–Sr clock it may give a much older