My work generally focuses on rock deformation both in the lab and through field investigations of faults. Below are some new and on-going projects for which I am actively seeking graduate students .
Examining earthquake slip in the rock record using temperature proxies
Faults are complex structures that consist of cracked wall rock (the damage zone), crushed up and altered rock (gouge or cataclasite) and thinner zones within the gouge where more localized slip has occurred (principal slip zones). None of these structures necessarily tell us where earthquake slip has occurred, because comminution and slip can localize at slow or fast speeds. However, only earthquakes generate frictional heat fast enough to create a localized temperature rise within the slipping layer. Sometimes the temperature gets high enough to melt the rock, which creates a fault rock known as pseuodotachylyte, but otherwise we need to look for proxies that can tell us what temperatures were achieved. Our group developed a proxy using the thermal maturity of organic matter (biomarkers) to find earthquakes in the rock record (Polissar et al., 2011; Savage et al., 2014), and have established the presence of earthquake slip in various faults such as the Japan trench and the creeping section of the San Andreas fault (Rabinowitz et al, 20202; Coffey et al., 2022). We continue to use the biomarker proxy to investigate new faults as well as calibrate the kinetics of various biomarker reactions by running rapid heating experiments. We are also beginning to use untargeted machine learning to find new biomarkers that might be effective thermal proxies. Furthermore, we are investigating clay dehydration and amorphization as a potential earthquake proxy (Krogh et al, AGU 2022).
From Savage and Polissar, 2019; Sampling grid at the Punchbowl fault, an ancient strand of the San Andreas fault.
Frictional properties of gouges at high temperatures
The strength and stability of faults is dictated by mechanical and chemical properties of fault gouge. However, the vast majority of friction experiments on gouge have been conducted at room temperature. My students, collaborators, and I are examining gouges – both natural gouges sampled from faults and synthetic gouges – to test how gouges behave at high temperatures.
One ongoing project is to test IODP drill cores and exhumed fault rocks from several subduction zones including the Hikurangi subduction zone off of New Zealand, Costa Rica, and Japan. This project is in collaboration with UT Austin, the Planetary Science Institute, and USC.
From Rabinowitz et al., 2018: The strength and velocity dependence of friction changes with increasing temperature and pressure for marls near the Hikurangi Subduction Zone.
At the extreme end of the temperature spectrum, we are working with collaborators at the University of Minnesota and Rice University to test gouges at temperatures high enough for sintering to occur, which is when grains weld together without melting. These experiments aim to understand frictional behavior higher temperature environments like solid-state extrusion from volcanoes.
Transition from creeping to locked fault behavior on the San Andreas fault at San Juan Bautista
Faults can either fail in earthquake slip, or at slower rates which we call creep. The San Andreas fault is famously broken up into a central creeping segment surrounded by locked segments that have been known to host large earthquakes. It is not known exactly what causes faults to creep, but the presence of weak mineral phases (like clay) in the fault gouge has been proposed as a potential mechanism. We are investigating the northern end of the creeping section of the San Andreas where the transitions from creeping to locked, near the town of San Juan Bautista. By mapping creeping and locked fault strands in the area and comparing their mineralogies and frictional properties, we are trying to determine what makes the San Andreas stop creeping in this region.
From Williams, 2022: En echelon road cracks delineate creeping fault strands
Creating long-term earthquake records with geo/thermochron techniques
To understand earthquake cycles, we need to look farther back than the historic earthquake record (i.e. earthquakes for which humans were present). Traditional paleoseismology can date earthquakes using proxies such as carbon dating, which can extend ~50 to 70 ky. Along with geochronology colleagues at Lamont-Doherty Earth Observatory, we are investigating other dating techniques for earthquakes over a range of ages, including K-Ar (Coffey et al., 2022). One aspect of this work involves determining the rates at which geochronometers can be reset during earthquake heating.