Michael Brown
Crustal melting and melt extraction, ascent and emplacement in orogens; mechanisms and consequences
Journal of the Geological Society of London (July 2007), 164(4):709-730

Abstract:
Significant volume of wet melting requires an influx of H ₂ O-rich volatile phase. In hydrate-breakdown melting, initial melt accumulation is diffusion-controlled and melt accumulates around peritectic phases in low-pressure sites. As the melt-bearing rock weakens, it becomes porous at a few volume per cent melt, initiating an advective flow regime; as melt volume reaches the melt connectivity transition, melt may be lost from the system in the first of several melt-build-up-melt-loss events. Using mineral equilibria modelling, major and accessory phase controls on melt chemistry are evaluated. In residual migmatites and granulites, microstructures indicate the former presence of melt whereas leucosome networks record melt extraction pathways. Sites of initial melting nucleate shear instabilities; strain and anisotropy of permeability control the form of millimetre- to centimetre-scale inferred melt channels and strong anisotropy promotes high fluid focusing. Focused melt flow occurs by dilatant shear failure of low melt volume rocks, which leads to the formation of melt flow networks allowing accumulation and storage of melt, and forming the link for melt flow from grain boundaries to ascent conduits. Melt ascent is via ductile fractures, which may propagate from dilation or shear bands. Fractures are characterized by blunt tips; also they may exhibit zigzag geometry close to the tips and petrographic continuity between leucosome in the host and granite in the dyke. Horizontal tabular and wedge-shaped intrusions commonly are associated with the brittle-ductile transition zone. Vertical lozenge-shaped intrusions represent congelation of magma in ascent conduits and blobby plutons record lateral expansion localized by instability; these intrusion types characterize emplacement at deeper levels in the crust.