Melting and melt migration in heterogeneous mantle beneath mid-ocean ridges

Evidence for chemical heterogeneity in the mantle is widespread in oceanic basalts, yet its consequences for basalt petrogenesis are little understood. A significant unknown is the effect that heterogeneity has on the dynamics of magma flow in the mantle. Observations of oceanic crust and the upper mantle suggest that magma migrates to the surface through a network of high porosity channels. In this thesis, I use computational models of coupled magma/mantle dynamics beneath mid-ocean ridges to question whether a physical connection exists between channelized flow and mantle heterogeneity. The models are intialized with simple, hypothetical patterns of heterogeneity that cause the fusiblity of the model mantle to vary. The principal result is that channel- ized melt flow is a consequence of melting in a heterogeneous mantle. Magma from preferentially melted heterogeneities nucleates high porosity, high permeability channels that grow by a feedback between magma flux and dissolution. Using the models in various configurations, I explore the dynamics of channel formation and investigate how the topology of mantle heterogeneity affects melt segregation and focusing beneath ridge axes. Additionally, I use the models to predict the speed and time scale of melt migration. A simple model of equilibrium partitioning is used to cast the results in terms of 230Th disequilibria. Comparisons of the modelled geochemistry against global measurements indicate that the models presented here provide a reasonable, first-order description of the dynamics of magma flow beneath ridges. I also explore a systematic connection between plate kinematics and global patterns of mid-ocean ridge bathymetry with three dimensional models of solid mantle flow beneath transform faults. The results provide new constraints on the scale of melt focusing and melt redistribution at ridge axes, and pose questions for future 3D studies of melt migration beneath ridges.