Behavior of Mingled Magmas during Explosive Eruptions

Enclaves, chilled bodies of magma with compositions contrasting with those of their hosts, have long been recognized in felsic plutonic rocks (e.g. Didier, 1973; Vernon, 1984; Didier and Barbarin, 1991).  Enclaves are distinguished from xenoliths based on textural evidence that they were not solid within their host magma.  This evidence includes: enclave shapes similar to those of liquid droplets, the presence of fine-grained chilled margins on enclaves, and the inclusion and partial dissolution in enclaves, of crystals from host rocks.  The genetic significance of enclaves in plutonic rocks remains controversial (e.g., Anderson et al., 1998; Silva et al., 2000; Perugini and Poli, 2000).  Direct connection to mafic partial mantle melts that may generate felsic magmas by crustal melting has remained elusive, although empirical observation suggests that basaltic enclaves are common in A-type felsic magmas associated with crustal extension (Wiebe, 1995). This observation has led to the suggestion that basaltic enclaves may be evolved remnants of the partial mantle melt heat source that caused crustal melting (Reid et al., 1983; Furman and Spera, 1985; Reid and Hamilton, 1987; Dorais et al., 1990). The problem addressed in this work is the significance of droplet-shaped basaltic inclusions in pyroclastic volcanic rocks and the contrasting implications of survival or fragmentation of magmatic inclusions during pyroclastic eruptions.  Volcanic rocks bearing evidence for magma mingling prior to eruption are most commonly lava flows, rather than pyroclastic rocks. In lava flows, two typical features are taken to indicate that the flow represents an extrusion of magma involving mingling of contrasting compositions.  First, enclaves have been identified and described in non-explosive lava flows (e.g. Bacon and Metz, 1984; Davidson et al., 1989).  Second, flow bands in some lava flows are interpreted to be chilled magmas of contrasting composition (e.g., Seaman et al., 1995), suggesting that mingling of magmas occurred as the magma made its way to the surface, and as it flowed out on the surface.  Banded pumice, which preserves contrasting magma compositions as incompletely fragmented glass (Koyaguchi, 1979) has been documented in pyroclastic flows, and is strong evidence for pre-eruption magma mingling.  Xenocrysts (Clynne, 1999) and compositionally primitive cores of phenocrysts (Davidson et al., 1989; Tepley and Davidson, 1999) in pyroclastic rocks are more subtle records of magma interaction.   

Exposures in coastal Maine document the occurrence of coherent basaltic enclaves in rhyolitic pyroclastic rocks, and the occurrence of andesitic ignimbrite that bears evidence of origin by fragmentation of basaltic enclaves in a rhyolitic host.   In this case, rather than being the hallmark of subduction-related volcanism, these andesitic ignimbrite sheets are a consequence of the eruption of the interface of compositional layers in a bimodal magma chamber.  The tectonic setting of such chambers is probably more likely to be an extension-related, within-plate or back-arc setting than a setting in the heart of an island arc or a magmatic arc.