Effects of Fault Roughness on Coseismic Slip and Earthquake Locations

Abstract

Fault zone structure is well known to exert strong controls on earthquake properties including coseismic slip distribution, rupture propagation direction, and hypocenter location. It has also been well established that the principal slip surface, which accommodates the majority of earthquake displacement, exhibits roughness at all scales following self-affine fractal distributions.Here we explore the relationship between fault roughness and specific earthquake properties including coseismic slip distribution and hypocenter location based on long-term simulations of earthquake catalogs on fractally rough faults. We begin by using the von Kármán autocorrelation function to procedurally generate single faults with different fractal roughness properties, which we place in a homogeneous elastic solid and apply pure right-lateral shear at a constant back slip rate with the earthquake simulator RSQSim. Running the simulations for 10,000 years each, we generate millions of earthquakes including thousands of events with Mw > 6.0, which rupture the surface. We show that the patterns of surface rupture in these large events follow self-affine fractal distributions with consistent fractal dimension related distinctly to the fractal dimension of the fault. In addition, the hypocenters of these large events occur in very specific predictable locations where the longest wavelength structure produces a stress asperity (i.e., restraining bend). The resulting patterns can explain many features observed on real fault systems, including clustered hypocenter locations, spatially variable coseismic slip distributions, and characteristic slip recurrence behavior. These results demonstrate a quantitative link between a directly measurable fault property-roughness-and the properties of future earthquakes.

Bookshelf Kinematics and the Effect of Dilatation on Fault Zone Inelastic Deformation: Examples From Optical Image Correlation Measurements of the 2019 Ridgecrest Earthquake Sequence

“…The left‐lateral slip distribution of the foreshock rupture shows a simple asymmetric triangular shape, while the right‐lateral distribution of the mainshock is heterogeneous and multi‐peaked (Figure 2b). These along‐strike variations of slip at different length scales (from 1 to 10 km) are robust as indicated by the uncertainty in our measurements and may reflect variations due to the fault geometrical roughness and strength or applied stress (Allam et al., 2019; Bruhat et al., 2020; Dunham et al., 2011; Milliner et al., 2016; Shi & Day, 2013). These variations can provide useful constraints for the degree of aleatory variability of displacement along a rupture that inform scaling relations in probabilistic fault displacement hazard models (Lavrentiadis & Abrahamson, 2019).…”

Bookshelf Kinematics and the Effect of Dilatation on Fault Zone Inelastic Deformation: Examples from Optical Image Correlation Measurements of the 2019 Ridgecrest Earthquake Sequence

“…The left-lateral slip distribution of the foreshock rupture shows a simple asymmetric triangular shape, while the mainshock is right-lateral and has a heterogeneous multi-peaked distribution suggesting ( Figure 2b). These along-strike variations of slip at different length scales (from 1-10 km) are robust as indicated by the uncertainty in our measurements and may reflect variations due to the fault geometrical roughness and strength or applied stress (Dunham et al, 2011;Shi and Day, 2013;Milliner et al, 2016;Allam et al, 2019;Bruhart et al, 2020), and are an important source of information for scaling relations in probabilistic fault displacement hazard models (Lavrentiadis & Abrahamson, 2019). In addition, the second invariant strain maps clearly show changes of the total strain intensity, which correspond to variations of the fault geometry and orientation along the rupture.…”

Crack to pulse transition and magnitude statistics during earthquake cycles on a self-similar rough fault

“…Recently, Allam et al (2019) used the RSQsim cycle simulator to simulate seismicity on a self-affine fault over multiple cycles. They used backslip to relax stresses and thus achieve an efficient way to simulate long term fault behavior.…”

“…They used backslip to relax stresses and thus achieve an efficient way to simulate long term fault behavior. However, Allam et al (2019) used oversized dislocations and did not resolve the relevant length-scales that arise from elasticity and the assumed friction law. Such models generally produce complex behavior that becomes simpler with grid refinement (Rice, 1993;Ben-Zion and Rice, 1997).…”

“…4 to compute the elastic stress transfer. This approach differs from the RSQsim backslip implementation (Richards-Dinger andDieterich, 2012;Allam et al, 2019), since I do not have to slip the faults backward to determine the backslip stressing rate. I've simply…”