Abstract
Melt extraction from the partially molten mantle is among the fundamental processes shaping the solid Earth today and over geological time. A diversity of properties and mechanisms contribute to the physics of melt extraction. We review progress of the past ∼25 years of research in this area, with a focus on understanding the speed and style of buoyancy-driven melt extraction. Observations of U-series disequilibria in young lavas and the surge of deglacial volcanism in Iceland suggest this speed is rapid compared to that predicted by the null hypothesis of diffuse porous flow. The discrepancy indicates that the style of extraction is channelized. We discuss how channelization is sensitive to mechanical and thermochemical properties and feedbacks, and to asthenospheric heterogeneity. We review the grain-scale physics that underpins these properties and hence determines the physical behavior at much larger scales. We then discuss how the speed of melt extraction is crucial to predicting the magmatic response to glacial and sea-level variations. Finally, we assess the frontier of current research and identify areas where significant advances are expected over the next 25 years. In particular, we highlight the coupling of melt extraction with more realistic models of mantle thermochemistry and rheological properties. This coupling will be crucial in understanding complex settings such as subduction zones. ▪ Mantle melt extraction shapes Earth today and over geological time. ▪ Observations, lab experiments, and theory indicate that melt ascends through the mantle at speeds ∼30 m/year by reactively channelized porous flow. ▪ Variations in sea level and glacial ice loading can cause significant changes in melt supply to submarine and subaerial volcanoes. ▪ Fluid-driven fracture is important in the lithosphere and, perhaps, in the mantle wedge of subduction zones, but remains a challenge to model.
Highlights
The viscous drag force associated with Darcian porous flow leads to a characteristic melt speed of k∆ρg w0 =
While the percolation velocity was noted by earlier workers, McKenzie (1984) recognised a natural length scale of magma– mantle interaction
Our focus in this review is on the physical mechanisms that account for the discrepancy in melt-transport speed between diffuse porous flow and inferences from observational proxies
Summary
Mantle partial melting is dispersed at the grain scale, where juxtaposition of different minerals leads to eutectic-like solidus lowering. In laboratory experiments textural equilibrium is reached in a matter of hours, but such experiments have much finer grain sizes than the mantle. The geometry is illustrated in the bottom right with melt network in yellow, grain in pink, with a porosity of 5%; ii) rock-analogue experimental measurements on synthetic quartzites from Wark & Watson (1998); iii) Stokes-flow calculations of permeability using X-ray microtomography of the pore space of olivine + basalt samples from Miller et al (2015). Scaled permeability shown assumes an average grain size of 35−+2155 μm for all experiments (similar to those quoted in Miller et al (2014)); iv) high-pressure, high-temperature centrifuge experiments on olivine + basalt samples by Connolly et al (2009); v) Short black lines show the expected slopes on this plot for k ∝ d2φ2 and k ∝ d2φ3. This section reviews laboratory experiments and how the relevant physics has been quantitatively upscaled
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
