Consequences of mantle heterogeneity for melt extraction at mid-ocean ridges

Evidence for chemical heterogeneity of the mantle is found at most mid-ocean ridges, yet we are only beginning to develop a theoretical understanding of how this heterogeneity affects basalt petrogenesis. In particular, the consequences of heterogeneity for magma transport through the mantle are not known. We address this question with computational models of coupled magma/mantle dynamics and thermochemistry beneath ridges. The models are initialised with simple, hypothetical patterns of mantle heterogeneity; we investigate the consequences of this heterogeneity for the predicted pathways of melt transport. Our results show that preferentially melted heterogeneities can nucleate magmatic channels that transport their melts. A channel generated in this way is cooler than the surrounding mantle; the temperature gradient drives an inward diffusive flux of heat, powering melting within and suppressing melting on its flanks. The channels become high-porosity, high-permeability pathways for rapid magmatic ascent toward the surface. Channelised flow has implications for the extent of mixing between melts that are derived from distinct sources. Melt segregation from the heterogeneous source in our models creates pools of trapped melt at the base of the lithospheric boundary layer, away from the ridge axis; these pools freeze and introduce new heterogeneity into the mantle. They may also deliver melt via dikes to off-axis magma chambers in the crust.

► We model melting and melt segregation from the mantle beneath a mid-ocean ridge.
► We impose chemical heterogeneity onto the mantle source.
► Magma from fertile heterogeneities can nucleate high-permeability channels.
► Channels are relatively cool and hence heat diffuses into them, powering melting.
► Melt pools beneath the lithosphere, freezes, and creates new mantle heterogeneity.