Earthquake cycle simulations with creep compaction and dilatancy
Yuyun Yang,  Stanford University

The goal of this study is to examine how porosity changes from dilation and compaction of the fault influence the occurrences of SSEs and earthquakes. We introduce porosity evolution encapsulating elastic, plastic and viscous deformation of the porous space, fully coupled to fault slip and pore pressure changes in a unified earthquake sequence simulation model. This is done in 2D antiplane shear for a planar, permeable rate-and-state fault in a homogeneous elastic solid. For fluid transport, we consider a mature fault zone with a well-developed damage zone. The damage zone is much more permeable than the fault core, and the modeled fault is a line, a 1D idealization of the damage zone which has a width varying from decameters to kilometers. Fluids are confined to the fault with along-fault diffusion in the damage zone. There is a constant fluid source at the bottom of the domain to approximate fluids produced from dehydration reactions of hydrous minerals. Porosity and permeability evolve with slip, with permeability related to porosity via a power-law relation. The SSEs are driven by the interaction between pore compaction which raises fluid pressure and weakens the fault, as well as pore dilation which decreases fluid pressure and limits the slip instability. We propose that the dynamics of pore compaction and dilatancy and the associated fluid pressure cycling, as well as velocity weakening friction likely have important implications for slow slip events across different tectonic settings.

Seismic cycles and earthquake statistics on heterogeneous faults
Camilia Cattania, MIT

Earthquake recurrence intervals are one of the most fundamental property of the seismic cycle, and a key ingredient of seismic hazard assessment. While back of the envelope calculations based on spring-slider models can provide an order of magnitude estimate, they do not capture the effect of realistic fault properties such as heterogeneous loading and fault complexity. Here I present numerical models and analytical insights into earthquake recurrence interval and its variability in the presence of realistic heterogeneity. I consider two aspects: 1) heterogeneity in stress due to loading from creeping fault sections adjacent to a locked segment; 2) structural heterogeneity induced by fault roughness and secondary faults in the damage zone.

I first consider asperities loaded by creep around them (analogous to small repeating earthquakes) or downdip of the seismogenic zone. The stress concentration at the boundary between locked and creeping segments induces localized creep within the asperity; the interevent time is controlled by the propagation of a creep front into the locked region, which occurs over a length scale comparable to the nucleation length. Analytical results match the behavior seen in quasi-dynamic earthquake cycle simulations of rate-state faults, and they reproduce the scaling between recurrence interval and seismic moment observed in nature for small repeating earthquakes. These models also predict a transition between quasi-periodic cycles for asperities with a dimension comparable to the nucleation length, to temporal clustering and a range of rupture dimensions for larger asperities.

I then consider the effect of fault roughness and off-fault damage, by modeling a rough fault surrounded by a collection of secondary fractures. Numerical simulations indicate that fault roughness and damage each lead to a significant reduction of the recurrence interval (30-50%), as well as abundant microseismicity between mainshocks. The effect of roughness is a consequence of normal stress perturbations induced by slip on a non-planar fault: regions with lower than average normal stress reach failure early in the cycle, and they start creeping at constant stress. This process redistributes stress on nearby locked patches, triggering a foreshock sequence culminating in the mainshock. Simple analytical models for this stress redistribution capture foreshock acceleration and mainshock recurrence intervals. Finally, I find that the presence of a damage zone further reduces the recurrence interval, which can be explained as follows. Early in the cycle, the mainshock fault is locked and at a shear stress close to its dynamic strength; in contrast, the shear stress may exceed static strength on nearby faults due to heterogeneous mainshock stress changes. Therefore, damage zone faults experience localized creep throughout the entire cycle which increases stressing rates on locked asperities. This process eventually initiates foreshock sequences in the damage zone, which in turn trigger the mainshock itself.

These applications demonstrate that analytical models accounting for heterogeneity in loading and fault strength are a powerful tool to quantify earthquake recurrence patterns in a continuum mechanics framework.

[register]