Fault Surface Geometry, Wear Processes and Evolution: Implications for Earthquake Mechanics and Fault Rock Rheology
James Kirkpatrick, Colorado State University
Kate Shervais, Colorado State University
To investigate how fault geometry impacts fault surface stresses, strain localization and wear during deformation, we use digital mapping techniques to collect observations of faults in two orientations. Terrestrial laser scanner (TLS) point clouds that define slickensides show that 3D fault surfaces are non-planar at all scales of observation. We use the point clouds to measure slickenline orientations, which do not vary significantly over exposed fault surfaces. Assuming the slickenlines parallel the local maximum resolved shear stress direction, the results show that the shear stress direction does not vary spatially. This suggests geometric asperities at exposure length scales (up to a few m) cause relatively minor stress perturbations. We model the scale-dependent strain of asperities during slip and show that relatively small asperities fail inelastically and support little stress, while larger ones deform elastically. Using structure-from-motion to produce rectified images of faults in cross section, we map the internal structure of the Boyd fault, CA. Our observations show that the thickness of fault rock varies along strike. Gouge-filled injections provide evidence for coseismic pressurization, and show that earthquake slip is distributed across layers of fault rock that vary in thickness from 10-3 to 10-2 m thickness over 101 m distances. Cross cutting relationships between different generations of gouge show that fault gouge formation is a non steady-state process, with multiple discrete layers forming successively. The roughness of the edges of gouge layers decreases with displacement, consistent with wear models. However, we document evidence for failure of wall rock protrusions that are entrained into the fault gouge, a process that is episodic and serves to both increase gouge volume and increase fault roughness. Together, the two data sets show that fault strength and fault rock rheology are a function of fault core and slip zone geometry.