How and why is deformation localised in the continental crust?

Our ability to accurately assess the seismic hazard posed by crustal-scale faults relies on a robust understanding of how the crust, in which such faults occur, deforms. However, currently, we cannot confidently assign specific rheological properties to actively deforming regions in any location. This thesis combines geological field, microstructural and chemical analyses to identify processes of localised and transient strain in the dominantly viscous mid-to-lower crust, with numerical modelling to test the influence of such processes on the characteristics of geodetically observable surface velocities, before and after crustal-scale earthquakes. First, I identify geological signatures of localised, syntectonic fluid infiltration in the mid-crust, which facilitates localised hydration reactions and viscous deformation dominated by relatively weak dissolution–precipitation creep over comparatively strong dislocation creep. I propose a conceptual model for the exhumed mid-crustal Badcall shear zone of northwest Scotland, where localised brittle failure of anhydrous crust in the dislocation creep regime introduces sufficient grain-boundary fluid for deformation to favour dissolution–precipitation creep, localising strain. Next, I compare and contrast the distinguishing microstructural signatures of dissolution–precipitation creep and dislocation creep in amphibole-feldspar-bearing rocks to support the identification of these different creep mechanisms. Then, I present a rheological model for quasi-Newtonian solution creep (which includes dissolution–precipitation creep), for comparison with power-law dislocation creep. I test the impact of the likely distribution of these two creep mechanisms in deforming regions by simulating sequences of seismic cycles on a frictional-viscous strike-slip fault in 2D with spatially heterogeneous rheological properties. My results show that coseismic, postseismic, and interseismic deformation are significantly altered both spatially and temporally, according to fault-localised rheological properties in the mid-to-lower crust. This suggests that transient seismic cycle deformation patterns have the potential to inform us of the rheological behaviour of a fault at depth, and therefore provide insights into potential future seismicity.