SCEC Project Details
SCEC Award Number | 19218 | View PDF | |||||||
Proposal Category | Individual Proposal (Integration and Theory) | ||||||||
Proposal Title | Effect of pore pressure on rupture dynamics: Laboratory study of the off-fault plasticity and dilatant hardening coupling | ||||||||
Investigator(s) |
|
||||||||
Other Participants | |||||||||
SCEC Priorities | 2c, 3d, 3f | SCEC Groups | FARM, CXM, SDOT | ||||||
Report Due Date | 04/30/2020 | Date Report Submitted | 04/30/2020 |
Project Abstract |
Pore fluids play important roles on earthquake rupture dynamics. In particular, dilatant hardening, involving the coupling between near-fault plasticity and pore pressure changes, is thought to be a favored mechanism for generating the recently discovered slow earthquakes. However, despite the well-known concept, the precise mechanics of dilatant hardening – such as factors governing the drainage condition of failure - remains poorly understood. Here we conducted experiments on quartz sandstones with various bulk diffusivities at a constant effective pressure (produced with various combinations of confining and pore pressures). We found two stages of failure stabilization by dilatant hardening, in which the drainage condition of failure is strongly influenced by sample bulk diffusivities. Low diffusivity samples show prolonged stable failure, followed by eventual unstable failure that is also stabilized and aseismic. In contrast, high diffusivity samples fail instantaneously with seismicity. However, even these samples exhibit the stabilization of unstable failure at high confining and pore pressures. The undrained conditions inferred from the two stages of stabilization agree with those predicted based on sample bulk diffusivities and the two timescales of deformation associated with the imposed strain rate and dynamic rupture. Thus, the initial long-term stabilization is consistent with that expected from the classic model of dilatant hardening for rock strengths. Our results suggest that rupture dynamics models considering dilatant hardening also need to account for the subsequent, shot-term stabilization during dynamic rupture. |
Intellectual Merit | Our study addresses several of the scientific objectives of SCEC 5, including (Q2) Role of off-fault inelastic deformation on rupture dynamics and (Q3) Role of evolving fault zone structure and physical property on shear resistance to seismic and aseismic slip. In particular, assessing the role of dilatant hardening on fault stability is a major goal of (Q3). Our results show that the effect of dilatant hardening - the coupling between near-fault dilatancy and pore pressure changes – on rupture dynamics depends on the fluid drainage conditions governed by the bulk diffusivity of rocks. We found two stages of failure stabilization by dilatant hardening during the course of failure, associated with the two distinct timescales of the imposed strain rate and dynamic rupture. Our work provides key experimental constraints on the effect of pore pressure on the form of damage-zone and nonlinear fault rheology (P2c) and the role of evolving fault zone diffusivity on rupture dynamics (P3d). |
Broader Impacts |
Field evidences suggest that most plate boundaries, including the San Andreas Fault and the Cascadia megathrust, are permeated with pore fluids. This study – through improving fundamental understanding of fluid-infiltrated faulting - will lead to the accurate assessment of earthquake hazards to society. This project – supplementing PI Kanaya’s postdoctoral study – has provided a professional development opportunity to better prepare PI Kanaya for an academic career in Earth Sciences. |
Exemplary Figure | Figure 3. Timescale of bulk diffusion vs. that of deformation, showing the drainage conditions of faulting in Fontainebleau sandstone. See text for explanations for the drainage analysis. Bulk diffusion times are directly determined from pulse transient tests for >100 s and calculated using measured permeability and storage capacity for ≤100 s. Two distinct timescales of fault drainage were recognized: long-term stabilization (~100 s) during the entire failure and short-term stabilization (~1 s) during the unstable part of failure. If samples are undrained prior to shear localization during axial loading (as indicated by strength increases), failure also involves both long-term and short-term stabilization. In contrast, if samples are drained before localization, failure occurs instantaneously under long-term drained conditions. However, even in this case as in the 14% porosity samples, unstable failure becomes stabilized (short-term stabilization) at high confining and pore pressures. |