Experimental Constraints on Fault Healing and Shallow Earthquake Rupture Propagation in Altered, Clay-rich Serpentinite Gouge, Bartlett Springs Fault, CA
Monica R. Barbery, Emma M. Armstrong, Ainsley MacDonald, Srisharan Shreedharan, Alexis K. Ault, & Greg HirthSubmitted September 8, 2024, SCEC Contribution #13861, 2024 SCEC Annual Meeting Poster #140
Creep influences the earthquake cycle and modulates stress on faults. The northern San Andreas fault system comprises creeping faults with altered serpentinite. An outstanding question is the extent to which instabilities can nucleate in and/or propagate through this material. We address this by investigating the frictional healing behavior of exhumed clay-rich serpentinite gouge along the Bartlett Springs fault (BSF). The BSF creeps at ~3.5 mm/yr, is capable of hosting a M7.5 earthquake, and recently had a M5 earthquake. We revisit the BSF locality north of Lake Pillsbury of Moore et al. (2018), where a serpentine-rich gouge zone cuts Pleistocene sediments. Similar gouge is exposed and likely present at depth along multiple faults in the northern San Andreas fault system that are also creeping, making this material a key rheology. Prior work demonstrated BSF gouge comprises clasts of antigorite, talc, chlorite, tremolite-actinolite, and Franciscan metamorphic rocks and a matrix derived from these materials plus quartz and saponite. Friction experiments indicate the coefficient of friction (µ) for BSF gouge is 0.14 – 0.3 at variable conditions and that this material is velocity strengthening at slow speeds.
We build on prior work and quantify the healing rate of BSF gouge with slide-hold-slide experiments on a direct shear apparatus at Utah State University. Experiments were conducted on an unconsolidated gouge sample adjacent to the contact with Pleistocene sediments with a median grain size of ~8 µm, and at 10 MPa normal stress, room temperature, and water saturated and room humidity (i.e., dry) conditions. The µ of the saturated and dry gouge are ~0.17 and ~0.36, respectively. Healing experiments with slow sliding velocities (0.001 mm/s) and variable hold times up to 10,000 s show healing rates are ~0.0007/decade for the saturated gouge and 0.0012/decade for dry gouge. Additional experiments performed using the Tullis rotary shear apparatus at Brown University yieldied similar µ values and healing behaviors. To investigate the role of fabric on frictional properties, experiments were repeated following a run-in displacement of 0.5 m. Following the run-in, the peak and steady-state µ dropped by 50% and 20%, respectively. Together, healing rates across a spectrum of slip rates and microstructural analysis will provide insight on how fault slip in altered serpentinite gouge influences either a transition to dynamic weakening or earthquake arrest.
Key Words
frictional healing, fault creep, Bartlett Springs fault
Citation
Barbery, M. R., Armstrong, E. M., MacDonald, A., Shreedharan, S., Ault, A. K., & Hirth, G. (2024, 09). Experimental Constraints on Fault Healing and Shallow Earthquake Rupture Propagation in Altered, Clay-rich Serpentinite Gouge, Bartlett Springs Fault, CA. Poster Presentation at 2024 SCEC Annual Meeting.
Related Projects & Working Groups
Fault and Rupture Mechanics (FARM)
