Photo: Atacama Desert, Antofagasta, Chile

Research interests

My research to date has been primarily concerned with two main areas;

  • the experimental deformation of rocks under simulated geological conditions, in order to help interpret natural processes such as faulting and earthquake mechanics and
  • detailed field studies on the structure and properties of strike-slip fault zones over a range of scales to further understand fault growth processes, subsequent mechanics, and bulk hydraulic and seismological properties of a fault zone

I have always been very conscious of the need to relate laboratory work with nature.  To this end, in addition to laboratory based experimentation, through my primary training as a geologist, I have carried out a lot of field-based research, often being able to bring an experimentalist‘s approach to illuminate essentially field-geological problems. 

The following sections briefly summarize the main areas in which I have carried out investigations to date, as well as ongoing research.


1. Stress rotation around fault zones

Slip on unfavourably oriented faults with respect to a remotely applied stress is well documented and implies that faults such as the San Andreas fault and low-angle normal faults are weak when compared to laboratory-measured frictional strength. If high pore pressure within fault zones is the cause of such weakness, then stress reorientation within or close to a fault is necessary to allow sufficient fault weakening without the occurrence of hydrofracture. From my field observations of a major tectonic fault, combined with uniaxial cyclic damage experiments and numerical modellin, we modelled and showed that microfracture variations can play an important role in stress rotation surrounding faults (Faulkner, Mitchell, Healy and Heap. Nature, 2006) and that stress rotation occurs within the fractured damage zone surrounding faults. In particular, we showed that stress rotation is considerable for unfavourably oriented ‘weak’ faults. In the ‘weak’ fault case, the damage-induced change in elastic properties provides the necessary stress rotation to allow high pore pressure faulting without inducing hydrofracture.



2. Structure, damage scaling and fluid flow in fault damage zones

In nature, permeability is enhanced in the damage zone of faults in crystalline rocks, where fracturing occurs on a wide range of scales. Understanding this permeability structure is paramount for the understanding of a wide range of geological processes, including earthquake generation and crustal strength, and the recovery of natural resources. My PhD research was aimed at improving our understanding of the hydraulic transport properties of large fault zones by presenting a large dataset of detailed field and microstructural observations (Mitchell & Faulkner, JSG, 2009) and results from a suite of laboratory experiments (Mitchell & Faulkner, JGR, 2008) to provide a basis for studying the distribution, and fluid flow properties, of damage surrounding large natural fault zones. In our studies we combined quantitative field measurements with laboratory measurements of permeability to predict microfracture damage zone permeability in low porosity crystalline rocks as a function of distance from the fault core and displacement. In the field we analyzed the scaling relationships of microfracture densities surrounding strike-slip faults with displacements ranging over 4 orders of magnitude (~0.12 m – 5000 m). This allowed the variation of microfracture damage with increasing distance from faults to be determined empirically as a function of displacement. We showed that microfracture damage can be reproduced in the laboratory in a suite of triaxial deformation experiments by inducing cyclic damage in initially intact samples while continuously measuring permeability. Combining field and laboratory datasets through the fracture density parameter, allows permeability changes to be predicted as a function of fault displacement. By integrating these new lab derived data with the field data, we have built a novel predictive model for the distribution and fluid flow properties of damage surrounding large fault zones as a function of fault maturity/displacement, which should make a significant contribution to our understanding distribution, and fluid flow properties, of damage surrounding large natural fault zones (Mitchell &Faulkner, EPSL, 2012).



3. Internal fault zone structures and high velocity / coseismic friction experiments

Most earthquakes in Earth’s crust are caused by fast slip on pre-existing faults. Friction decreases as a result of rapid, localized heating produced by high slip rates. These results, combined with classic work of Byerlee have led to the slip weakening model, in which friction was assumed to decrease as slip on the fault increased. Mechanical studies of rocks in the late 1970s provided the first experimental evidence that steady-state friction indeed decreased logarithmically with slip rate. Recent advances in high-velocity friction experiments now allow such processes to be tested in the laboratory at coseismic slip velocities of metres per second. Such apparatus allows for the reproduction of internal fault zone structures and high velocity frictional experiments on various materials that are found in natural fault cores, in an attempt to reproduce slip conditions and structures similar to that seen during earthquake rupture events, along with detailed measurements of frictional properties and thermally activated processes, such as some of the first laboratory examples of thermal pressurization (De Paola, Hirose, Mitchell, Di Toro, Viti, & Shimamoto, Geology, 2011). Many theoretical and more recently experimental studies havesuggested that thermally controlled/activated processes (e.g. thermal pressurization, flash heating, melt lubrication) may dramatically reduce the shear strength within the slip zone.

Currently, I am conducting experiments on various gouge materials that have been found in natural fault cores. These include materials such as serpentinite minerals, gouge from the SAFOD drill hole on the San Andreas Fault, and various phyllosilicate rich fault gouges to explore the variations of high velocity frictional properties at various conditions. 



4. Earthquake damage and pulverized rocks

Pulverized fault zone rocks (PFZR) have previously been investigated in various studies along the San Andreas Fault describing their distribution and unique physical characteristics. Unlike typical fault zone rocks, such as cataclasites and breccias that reveal accommodation of deformation by localized shear, grain rolling and grain size reduction, PFZR are characterized by intense micro-scale damage due to extensive fragmentation of grains down to the micron-range while still maintaining the original grain boundaries and rock fabric. Indications for shear strain are generally lacking and the fragmentation appears associated with a high density of opening mode microfractures. Such structures have previously been interpreted as a product of dynamic stress fluctuations, hence the occurrence of PFZR has been associated with earthquake rupture.

We identified and described a newly recognized outcrop of pulverized rock on the Arima‐Takatsuki Tectonic Line, Japan. We show that the observed spatial distribution and properties of the pulverized rocks are consistent with damage generated during successive earthquake events by dynamic reduction of normal stress and/or high slip rates on the bimaterial interface, supported by new laboratory measurements of P wave velocities on intact samples collected from around the fault zone (Mitchell, Ben-Zion & Shimamoto, EPSL, 2011).



5. Stuck in the mud? Earthquake nucleation and propagation through accretionary forearcs

We have identified thermal pressurisation as a process to explain the propagation of tsunami-generating earthquakes through clay-rich accretionary wedge material at subduction zones which are typically velocity strengthening thus inhibiting earthquake nucleation (Faulkner, Mitchell, Benson, Hirose & Shimamoto, GRL, 2011). High velocity fault slip induces rapid thermal pressurization of pore fluid within clay gouge leading to immediate weakening and negligible critical slip weakening and fracture energy, which can explain how a large rupture propagating from depth may not be arrested by typically velocity-strengthening clays.



6. Coseismic damage and softening of fault rocks at seismogenic depths

Elastic stiffness, a critical property for stress-orientation, propagation of earthquake ruptures and associated seismic waves, and the capability of crustal rocks to store strain energy, is expected to be highly variable throughout the seismic cycle due to complex sequences of damage and healing. Post-seismic healing and exhumation-related alteration render it impossible to assess how well rock stiffness as measured in the laboratory on samples collected from fault zones represents in situ, coseismic rock stiffness at seismogenic depths. Here we estimate the in situ, coseismic stiffness of fault rocks from the pseudotachylyte-bearing Gole Larghe Fault Zone (Italian Southern Alps), using aspect ratio measurements of pseudotachylyte injection veins and numerical simulations. We show coseismic stiffness can be 5 to 50 times smaller than the stiffness obtained from laboratory measurements on present day exhumed rocks, suggesting that post-seismic healing completely alters rock mechanical properties (Griffith, Mitchell, Renner & Di Toro, EPSL, 2012).



7. Damage structure, anisotropy and seismic velocity of pulverized rocks on the San Andreas Fault.

Identification of unique damage structure, anisotropy and seismic velocity of pulverized rocks near the San Andreas Fault, using a combination of seismic refraction tomography, laboratory ultrasonic velocity measurements and microstructural observations (Rempe, Mitchell, Renner, Nippress, Ben-Zion & Rockwell, JGR, 2015).





8. Thermal decomposition and dynamic weakening at seismic slip rates in giant carbonate landslides

The 3400 km2 Heart Mountain landslide of northwestern Wyoming and southwestern Montana is the largest subaerial landslide known. This Eocene age slide slid ∼50 km on a carbonate rich basal layer ranging in thickness from a few tens of centimeters to several meters, along a shallow 2° slope, posing a long-standing question regarding its emplacement mechanism. It has recently been suggested that such large displacement was aided by strong dynamic weakening mechanism, thermal pressurization due to shear heating and thermal decomposition in the basal layer slip zone, with theoretical simulations suggesting slip velocities ranging between tens of meters per second to more than 100 m s−1. In this study, we are working on a suite of high velocity friction experiments on initially intact carbonates collected from the Heart Mountain region, in attempt to reproduce conditions experienced in the slip zone of the basal section during emplacement of the landslide.



9. Ultra-high permeability induced by seismic waves

During an earthquake, fracture damage can be imparted into the wallrock surrounding faults due to stress pulses induced by the radiation of seismic waves. High strain rates restricted to small strains in fast propagating earthquakes can induce pervasive microscopic fracturing which, if fast enough, can pulverize the rock. Therefore, large earthquake ruptures can impart significant changes on the strength, elastic and fluid flow properties of a fault zone immediately after rupture propagation, and prior to fault slip. Variations in hydraulic properties can therefore exert both positive and negative feedbacks on processes controlling fault slip depending on the type of damage imparted at the rupture tip. However, to date the extent of such coseismic high-strain rate damage and the effect on hydraulic properties is not fully understood. Here we show that damage due to rupture-induced high strain rate pervasive microfracturing can enhance bulk wallrock permeability by up to 7 orders of magnitude in a single high strain-rate event.

Laboratory measurements of permeability on granitic samples deformed at high strain rates compare well with measurements of permeability in natural samples collected from the damage zone of San Jacinto fault. Calculations of strain rates expected surrounding a dynamic shear fracture propagating with a velocity approaching the limiting Rayleigh wave speed show that high strain rates attained in experiments are attainable in nature.



10. Fault welding by pseudotachylite generation