Thermo- and hydro-mechanical processes along faults during rapid slip

Abstract

Field observations of maturely slipped faults show a generally broad zone of damage by cracking and granulation. Nevertheless, large shear deformation, and therefore heat generation, in individual earthquakes takes place with extreme localization to a zone <1-5 mm wide within a finely granulated fault core. Relevant fault weakening processes during large crustal events are therefore likely to be thermal. Further, given the porosity of the damage zones, it seems reasonable to assume groundwater presence. It is suggested that the two primary dynamic weakening mechanisms during seismic slip, both of which are expected to be active in at least the early phases of nearly all crustal events, are then as follows: (1) Flash heating at highly stressed frictional micro-contacts, and (2) Thermal pressurization of fault-zone pore fluid. Both have characteristics which promote extreme localization of shear. Macroscopic fault melting will occur only in cases for which those processes, or others which may sometimes become active at large enough slip (e.g., thermal decomposition, silica gelation), have not sufficiently reduced heat generation and thus limited temperature rise. Spontaneous dynamic rupture modeling, using procedures that embody mechanisms (1) and (2), shows how faults can be statically strong yet dynamically weak, and operate under low overall driving stress, in a manner that generates negligible heat and meets major seismic constraints on slip, stress drop, and self-healing rupture mode.

Earthquake Sequence Calculations With Dynamic Weakening Mechanisms

“…Although the strength at each point on the fault at low slip rates is high ( e f 0 ~ 100 MPa), averaged shear stress on a fault is low because of low coseismic shear stress and frequent rupture nucleation, consistent with the results of [4], [5], [6]. Long-term shear stress distribution and averaged shear stress in the seismogenic region (-50 km < x < 50 km) is shown in Figures 2a and b, respectively, for cases with L = 10 mm.…”

“…These observations suggest that most fault points have low shear stress before large earthquakes, slip mainly at even lower shear stresses, and lock with the final stress only modestly lower than the initial value. [4], [5], and [6] demonstrated that even if a fault is strong at low slip rates, a dynamic rupture can propagate on a low-stressed fault (shear stress / normal stress ~ 0.3 or less) if the strength of a fault dramatically decreases coseismically, and discussed the operation of a fault at a low long-term shear stress.…”

Three‐dimensional earthquake sequence simulations with evolving temperature and pore pressure due to shear heating: Effect of heterogeneous hydraulic diffusivity

“…Previous studies [Perrin et al, 1995;Zheng and Rice, 1998;Lapusta and Rice, 2003;Noda et al, 2009;Rice et al, 2009] also focused on propagation of ruptures at low shear prestress compared to the static strength of rocks in the laboratory. Such low shear prestress can be inferred for natural faults, for example, from the measurements of the principal stress directions at boreholes around the San Andreas fault [Hickman and Zoback, 2004;Townend and Zoback, 2004].…”

Flash weakening of serpentinite at near-seismic slip rates

“…However, recent experiments employing near‐seismic slip velocities reveal dramatic frictional weakening in these materials, demonstrating the operation of dynamic weakening mechanisms, including flash heating of asperities, thermal pressurization of pore fluids, phase decomposition, frictional melting, and gel weakening [e.g., Rice et al , 2009; Tullis et al , 2007; Goldsby and Tullis , 2002; Di Toro et al , 2004, 2006; Hirose and Shimamoto , 2005, Han et al , 2007, 2010; Mizoguchi et al , 2007]. Dynamic fault weakening may resolve the apparent inconsistencies between estimates of static stress change and in situ crustal stress measurements [e.g., Noda et al , 2009; Rice et al , 2009]. …”