Page 393 - Caldera Volcanism Analysis, Modelling and Response
P. 393

368                                                         Thomas R. Walter


             Surface eruptions circumscribing a caldera center have occurred at several
          locations, for instance, on Deception Island (1967–1969) and Niuafoou (1853,
          1867) (Newhall and Dzurisin, 1988). While some historical caldera activity may
          have been related entirely to ring-dikes, others have been due to a dike that was
          partially reactivated and intruded a pre-existing fault. At Rabaul, a caldera system
          with a strongly elliptical outline, eruptive activity simultaneously occurred on
          opposite sides in 1878 and 1937 (Mori and McKee, 1987; Nairn et al., 1995;
          Saunders, 2004). At some caldera systems, ring-dike intrusions and eruptive activity
          occurred at two or more ring vents, as for example at the 20   30 km wide
          Tondano caldera in 1952 and 1971 (Lecuyer et al., 1997).
             Several examples of calderas and their ring-dike structures are shown in
          Figure 10. The first image shows a downsag caldera, lacking prominent ring-
          fractures and ring-dikes. The Gross Brukkaros system in Namibia (Figure 10a) has
          not formed zones of major concentric structural weakness. The 10 km wide
          structure was first subject to shallow magma inflation at about 4 km depth (Komuro
          et al., 1984) and intense surface doming, bending the surface stratigraphy
          (Stachel et al., 1994). The doming phase was followed by depletion of up to
              3
          5km , but no major ring-dike or ring-fault structure was formed. This example
          illustrates that large magma evacuation and downsag caldera subsidence is possible
          without major structural fracturing and dike injection visible at the surface.
             Intrusion around a caldera basin is the most comprehensible representation of a
          ring-dike. Some of the ring-dike intrusions occur along a single ring-fault, while
          others occur along several subparallel faults that form a near-concentric pattern at
          the surface. Ring-dikes are described in the Peninsular Range Batholiths of Baja
          California, Mexico, and southern California (Johnson et al., 2002). The formation
          of the Ramona ring-dike (Figure 10b) is thought to have been associated with a
          phase of subsidence of the caldera floor (Mirriam, 1941). Another elliptical ring
          structure is shown, from the El Pinal complex (Figure 10c), which is interpreted as
          a ring-dike that causes a well-developed contact zone (Duffield, 1968).
             Ring-dike structures that develop in a region tectonically predisposed or subject
          to a tectonic stress field may develop very complex shapes. An intruding ring-dike
          may follow pre-existing faults or discontinuities and thus abruptly stop or change its
          direction, conditioned by various parameters such as material property, fault friction
          coefficient, pore pressure, and state of stress. A simple scenario was simulated here
          to illustrate the significance of pre-existing discontinuities. The numerical model
          showed that caldera subsidence close to a linear fault is highly asymmetrical, being
          focused at one side of the fault structure. The Erongo complex is one of the largest
          of the Damaraland complexes in Namibia, and shows a central 30 km wide massif
          with silicic intrusions, partly encircled by a tholeiitic ring-dike with a 50 km
          diameter (Figure 10d). The ring-dike half formed only on the north-western side of
          a SW–NE oriented regional fault that was thought to have been reactivated during
          intrusion (Wigand et al., 2004). It was suggested that the ring-dike formed after a
          caldera collapse (Wigand et al., 2004). The absence of a ring-dike on the segment
          opposing the fault implies that circular dike intrusion was hindered. The models
          shown in this study suggest that a reactivated tectonic fault may act as a barrier
          to deformation; so a change in magma chamber pressure leads to significant
   388   389   390   391   392   393   394   395   396   397   398