Page 299 - Caldera Volcanism Analysis, Modelling and Response

274 J. Martı ´ et al.

Analogue and scale models have verified that caldera-collapse formation is
influenced by multiple aspects such as regional tectonics, system geometry, magma
and host rock properties, pre-existing structural discontinuities, deformation
history, etc.
Differences among the existing analogue models lie on the applied experimental
setup, the host rock analogue materials (dry quartz sand, flour, etc.), and the magma
chamber analogue (water or air-filled balloons, silicone reservoirs, etc.). Although
the results obtained are generally similar, there are some marked differences. A
subsequent comparison of the experimental results with field data and theoretical
models may help to infer which discrepancies stem from experimental restrictions.
Experimental models are useful to offer a ‘semi-quantitative’ approach to the
understanding of caldera collapse. In some cases, experimental results are important to
crosscheck analytical results. In other cases, experimental results lay the foundations
for more elaborate future numerical analyses. Three types of semi-quantitative
analyses can be performed with the results from revised experimental models: (i) the
quantification of the erupted magma chamber volume fraction f required to achieve
each step of a collapse process (and the dependence f ( R)); (ii) the analysis of the
subsidence pattern and its spatial extent; and (iii) the study of the influence of the roof
aspect ratio on the extent of the collapsed zone at both surface and depth.
In contrast to inferences from scaled models and mining subsidence, there is
a significant difference between these analogues and natural calderas. Caldera
processes, to some extent, are often related to overpressure in an associated magma
chamber, which induces a ground deformation pattern that may later be reused
during collapse. In addition, caldera collapse appears to occur despite a still
significant pressurisation of the chamber, a phenomenon absent during mining
subsidence where collapse occurs into an empty cavity. This apparent discrepancy
raises an interesting observation as to the possibility of caldera subsidence in some
cases taking place at the end of an eruption, when most magma has already been
evacuated from the chamber, rather than during the eruption. In the latter case,
mostly applicable to large caldera eruptions, a great proportion of the caldera
products would be deposited into the depression. In the former case, caldera-
forming products would not be deposited inside the ensued depression.
It is also important to mention that experimental results reveal that the apparent
diversity of caldera morphologies and collapse mechanisms inferred from field
studies (Lipman, 1997, 2000) may just result from the effect of different combina-
tions of magma chamber geometries, magma chamber roof aspect ratios, and
regional faults, or may simply correspond to different stages of a single collapse
process. Therefore, based on the results from scaled experimental models, it may be
prudent to revise the classification of collapse calderas based on their morphology
or their inferred collapse mechanism. This classification may just result or
correspond to an artefact due to different degrees of exposure of natural examples.
Finally, experimental models also enable the examination of the effect of
regional faulting and/or tumescence (due to local uplift produced by the inflation
of a shallow magma chamber or regional doming resulting from underplating or
compressional tectonics) on the formation of collapse calderas.