The rift-to-drift transition at rifted margins is an area of active investigation due to unresolved issues of the ocean-continent transition (OCT). Deep structures that characterize modern OCTs are often difficult to identify by seismic observations, while terrestrial exposures are preserved in fragments separated by tectonic discontinuities. Numerical modeling is a powerful method for contextualizing observations within rifted margin evolution. In this article, we synthesize geological observations from fossil ocean-continent transitions preserved in ophiolites, a recent seismic experiment on the Ivorian Margin of West Africa, and GeoFLAC models to characterize mantle deformation and melt production for magma-poor margins. Across varied surface heat fluxes, mantle potential temperatures, and extension rates our model results show important homologies with geological observations. We propose that the development of large shear zones in the mantle, melt infiltration, grain size reduction, and anastomosing detachment faults control the structure of OCTs. We also infer through changes in fault orientation that upwelling, melt-rich asthenosphere is an important control on the local stress environment. During the exhumation phase of rifting, continentward-dipping shear zones couple with seaward-dipping detachment faults to exhume the subcontinental and formerly asthenospheric mantle. The mantle forms into core-complex-like domes of peridotite at or near the surface. The faults that exhume these peridotite bodies are largely anastomosing and exhibit magmatic accretion in their footwalls. A combination of magmatic accretion and volcanic activity derived from the shallow melt region constructs the oceanic lithosphere in the footwalls of the out-of-sequence, continentward-dipping detachment faults in the oceanic crust and subcontinental mantle.

Serpentinite subduction and associated dehydration vein formation are important for subduction zone dynamics and water cycling. Field observations suggest that en échelon olivine veins in serpentinite mylonites formed by dehydration during simultaneous shearing of serpentinite. Here, we test a hypothesis of shear-driven formation of dehydration veins with a two-dimensional hydro-mechanical-chemical numerical model. We consider the reaction antigorite + brucite = forsterite + water. Shearing is viscous and the shear viscosity decreases with increasing porosity. Total and fluid pressures are initially homogeneous and in the serpentinite stability field. Initial perturbations in porosity, and hence viscosity, cause fluid pressure perturbations during simple shearing. Dehydration nucleates where fluid pressure decreases locally below the thermodynamic pressure defining the reaction boundary. During shearing, dehydration veins grow in direction parallel to the maximum principal stress and serpentinite transforms into olivine inside the veins. Simulations show that the relation between compaction length and porosity as well as the ambient pressure have a strong impact on vein formation, while the orientation of the initial porosity perturbation and a pressure-insensitive yield stress have a minor impact. Porosity production associated with dehydration is controlled by three mechanisms: solid volumetric deformation, solid density variation and reactive mass transfer. Vein formation is self-limiting and slows down due to fluid flow decreasing fluid pressure gradients. We discuss applications to natural olivine veins as well as implications for slow slip and tremor, transient weakening, anisotropy generation and the formation of shear-driven high-porosity bands in the absence of a dehydration reaction.

Serpentinite subduction and the associated formation of dehydration veins is important for subduction zone dynamics and water cycling. Field observations suggest that en-échelon olivine veins in serpentinite mylonites formed by dehydration during simultaneous shearing of ductile serpentinite. Here, we test a hypothesis of shear-driven formation of dehydration veins with a two-dimensional hydro-mechanical-chemical numerical model. We consider the reaction antigorite + brucite = forsterite + water. Shearing is viscous and the shear viscosity decreases exponentially with porosity. The total and fluid pressures are initially homogeneous and in the antigorite stability field. Initial perturbations in porosity, and hence viscosity, cause fluid pressure perturbations. Dehydration nucleates where the fluid pressure decreases locally below the thermodynamic pressure defining the reaction boundary. Dehydration veins grow during progressive simple-shearing in a direction parallel to the maximum principal stress, without involving fracturing. The porosity evolution associated with dehydration reactions is controlled to approximately equal parts by three mechanisms: volumetric deformation, solid density variation and reactive mass transfer. The temporal evolution of dehydration veins is controlled by three characteristic time scales for shearing, mineral-reaction kinetics and fluid-pressure diffusion. The modelled vein formation is self-limiting and slows down due to fluid flow decreasing fluid pressure gradients. Mineral-reaction kinetics must be significantly faster than fluid-pressure diffusion to generate forsterite during vein formation. The self-limiting feature can explain the natural observation of many, small olivine veins and the absence of few, large veins. We further discuss implications for transient weakening during metamorphism and episodic tremor and slow-slip in subduction zones.