Melt migration in the lower crust by porous melt flow

Migration of melt transfers heat, mass, and volatiles from depth towards the surface, driving differentiation of the Earth and playing a major role in generating its compositionally layered structure. Additionally, stress dissipation is strongly accommodated by strain partitioning into inherently weak melt-bearing rocks and particularly into the melt itself. Therefore, knowledge of the mechanisms of melt migration and their spatial distribution is fundamental to understanding the chemical and rheological evolution of the continental crust over time. The current understanding of crustal melt migration mechanisms is biased towards mechanisms that preserve large segregations of leucosome in outcrop, such as dyking or zones of stromatic migmatite. Therefore, existing knowledge is lacking about other melt migration mechanisms that preserve little to no melt along their pathways of migration, such as porous melt flow. Porous melt flow occurs at the grain scale and is defined as the migration of melt through an interconnected network of grain boundaries and triple junctions within a solid rock. This thesis aims to fill this knowledge gap by identifying zones fluxed by porous melt flow and assessing its potential as a significant melt migration mechanism by studying relationships in the Pembroke Granulite, a low-strain outcrop of magmatic arc lower crust, Fiordland, New Zealand. Identification of zones fluxed by porous melt flow is based on i) field relationships, ii) microstructures, and iii) composition of the whole rock and mineral assemblage. This multidisciplinary approach provides insight into the chemical and rheological evolution of the lower continental crust during porous melt flow. Four distinct episodes of partial to complete hydration of the host two-pyroxene-pargasite gneiss are identified in the Pembroke Granulite, each resulting in spatially and geologically distinct rocks. The development of these distinct rocks is largely due to different magnitudes of melt-rock reaction during porous melt flow. In this thesis, the difference in magnitude of melt-rock interaction ranges from a nearly isochemical reaction involving the addition of volatiles to the host rock, to the transformation of a gabbroic protolith to hornblendite. The spatial distribution of melt is inferred to localise strain, as high-strain zones are associated with melt-rich channels. Within the high-strain zones, a lack of deformation microstructures implies the dissipation of stress was accommodated by the fluxing melt. A lack of deformation microstructures indicates that deformation ceased when the melt crystallised, suggesting the timing and duration of deformation is controlled by the volume and spatial distribution of melt. The presence of microstructures indicative of the former presence of melt within partially to completely hydrated zones of the Pembroke Granulite indicate hydration occurred because of melt-rock interaction. In both low- and high-strain zones, interaction between the Pembroke Granulite and fluxing melt replaced the pre-existing, deformed assemblage with an undeformed, hydrated assemblage. Minerals in the hydrated assemblage have homogeneous compositions at an outcrop scale, and igneous-like rare earth element patterns. Four distinct episodes of porous melt flow are identified within the 0.15 km2 outcrop of Pembroke Granulite. The identification of multiple episodes within such a small, representative outcrop of magmatic arc lower crust strongly suggests that this melt migration mechanism may be more common in the lower crust than previously recognised. In general, recognition of porous melt flow in the lower crust is likely hampered by a combination of lack of exposure, overprinting during exhumation, and a lack of microstructural and geochemical tools to aid identification. This thesis characterises key microstructures indicative of the former presence of melt and igneous-like signatures within new, hydrated assemblages, which are proposed as tools for recognising rocks formed by melt-rock interaction during porous melt flow.