The evolution of lithosphere deformation due to infiltration of asthenosphere melt into a lithosphere with inherited weakness: insight from 2D numerical models of continental rifts

Melt formed in the asthenosphere affects lithosphere evolution through its accumulation and subsequent infiltration and heating of the lithosphere. Magmatic weakening of the lithosphere has received particular attention in the context of continental rifting because homogeneous continental lithosphere is too strong to rift under available tectonic forces without a weakening mechanism. But the observation that rifting generally initiates in heterogeneous continental mobile belts suggests that rift zones develop in lithosphere with some inherited compositional weakness, obscuring the importance of magmatic weakening. To test the relative roles of magmatic and compositional weakening, we construct a 2D numerical model that includes both effects. We treat the lithosphere and asthenosphere as separate, but coupled, domains. The lithosphere deforms via a composite brittle-ductile rheology with a strain rate that varies horizontally but is independent of depth. We apply an extensional force that is constant throughout the lithosphere, causing thinned or weakened regions of the lithosphere to extend at higher strain rate. We treat the asthenosphere as a viscous and partially molten and solve the 2D conservation equations for mass, momentum and energy for both the solid and melt phases. Melting and freezing are treated using a hydrated peridotite solidus. Solid velocities in the asthenosphere are calculated using the solid velocities from the lithosphere base as boundary conditions. The asthenosphere and lithosphere are coupled through magma infiltration and subsequent lithosphere heating. Melt fractions accumulating above an imposed critical melt fraction are extracted and emplaced within a few kms of the lithosphere-asthenosphere boundary, where it freezes, releasing latent heat. We test scenarios with a fixed critical melt fraction as well as a variable critical melt fraction determined by our previously published parametrization of dike propagation [Havlin et al., EPSL, 2013]. Initial results without magmatic infiltration indicate that even with compositionally weakened lithosphere, standard thickness continental lithosphere is too strong to rift. Specifically, due to the temperature dependence of diffusion and dislocation creep, ~150 km thick lithosphere requires unrealistically weak lithosphere to rift under reasonable tectonic force. In contrast, in simulations with thinner initial lithosphere (~70 km), the lithosphere does extend with modest pre-existing weaknesses and the asthenosphere upwells and melts in response. Melting in the asthenosphere is initially broadly distributed beneath the rift, with melting maxima beneath the rift flanks, consistent with observations of off-axis volcanism occuring early in rift zone development [e.g. Wolfenden et al., GSA Bulletin, 2005]. Further simulations will test how these lateral variations in melting and melt migration may lead to lateral variations in lithosphere weakening and strain rates.