A textural and geochemical study of quenched mafic inclusions from the Oruanui eruption, Taupo volcano, New Zealand

摘要

Interactions between mafic and felsic magma are common in magmatic systems. In silicic volcanic systems, mafic melts are the parental magmas for rhyolites, and also provide heat and volatiles that drive magma chamber processes and may trigger eruptions. Owing to their higher density relative to rhyolite, these mafic magmas rarely reach the surface. However, where they do so they provide information about immediately pre- and syn-eruptive magma chamber processes. These processes are investigated in this thesis by examination of textural relationships, and whole-rock, mineral and glass chemistries in quenched juvenile mafic clasts from the 25.4 ka Oruanui supereruption (~530 km³ magma) of Taupo volcano, New Zealand. The mafic component of the Oruanui eruption is significant because of its volume (3-5 km³), and because two distinct compositional groups, one tholeiitic and one calc-alkaline, were erupted simultaneously. Groundmass textures in the Oruanui juvenile mafic clasts show features indicative of rapid crystallisation (quenching), including micro-crysts with disequilibrium habits (acicular, swallowtail and hopper), and abundant interstitial residual glass. These features are inferred to result from chilling of the mafic magmas against the cooler Oruanui rhyolite upon injection into the silicic magma system. Quantitative textural data reveal a marked diversity in the groundmass textures of the mafic clasts, including crystal sizes, aspect ratios, area fractions, number densities, and the relative abundance of minerals. Any textural contrasts between the tholeiitic and calc-alkaline groups are overshadowed by more significant variations within each group, reflecting differences in the conditions under which crystallisation occurred as well as the compositional differences between the crystallising magmas. The fact that textures vary more significantly within than between the compositional groups suggests that bulk composition exerted only a second-order control on textural development. Crystal sizes and area number densities, in particular, show no correlation with whole-rock chemistry, implying that diverse cooling histories exerted an important control on the resultant textures. Diversity in the conditions of mafic clast crystallisation inferred from groundmass textures is reinforced by the chemistries of groundmass phases. Although the relative abundance of plagioclase and amphibole micro-crysts correlates with whole-rock chemistry to a first order, a range of compositions are observed, and compositional differences between clasts are not primarily dictated by bulk chemistry. Clustering of groundmass mineral compositions with respect to textural groupings suggests a fundamental link between compositions and textural development. This is inferred to result from a complex combination of factors, including the degree of undercooling, water content, cooling rate, bulk composition, and possibly intensive variables. Many macro-crystals display compositions overlapping with crystals from the Oruanui high- and low-silica rhyolites, indicating they are xenocrysts ingested from the rhyolite. However, residual glass compositions in the mafic clasts are chemically distinct from the Oruanui rhyolite glasses, and a mixing trend between them is not observed. The dominance of crystals derived from the low-silica rhyolite, combined with a scarcity of ingested rhyolite melt, suggests interaction of the mafic and felsic magmas occurred predominantly within a transition zone between the crystal mush and the overlying crystal-poor high-silica rhyolite. Diversity in the conditions of crystallisation inferred from the groundmass textures and chemistries of the mafic clasts are inconsistent with previously proposed models invoking crystallisation in a ponded mafic layer at the base of the melt-dominant body, prior to mechanical breakup and dispersal. A new model for generation of the mafic clasts is presented, whereby mafic dikes encountering a transition zone between the rigid crystal mush and the melt-dominant body disaggregate into discrete blebs of a range of sizes that quench against the rhyolite. Each bleb experiences its own unique cooling history, resulting in diversity in the resultant groundmass textures and chemistries. Ingestion of crystal-rich felsic material by the mafic blebs occurs, and transfer of heat to the semi-mobile transition zone magma induces convection, facilitating transfer of the quenched mafic blebs into the overlying melt-dominant body. Evidence for plastic deformation and the abundance of residual glass in the mafic clasts suggests their formation was a syn-eruptive processes. Spikes in the abundance of juvenile mafic material in phases 3 plus 4, 7 and 9 (Wilson, C.J.N. (2001). The 26.5 ka Oruanui eruption, New Zealand: an introduction and overview. Journal of Volcanology and Geothermal Research, 112, 133-174) are inferred to signify fresh injections of mafic magma into the silicic system during the eruption, likely a consequence of the regional-scale rifting inferred to have triggered and modulated the eruption (Allan, A.S.R., et al. (2012). The invisible hand: tectonic triggering and modulation of a rhyolitic supereruption. Geology, 40, 563-566).