In the past few years, there have been renewed interest on the question of frictional strength, especially in the context of the coupling of tectonic deformation and fluid flow. The interest is primarily motivated by recent observations indicating that the pore pressure in many seismotectonic settings are anomalously high. Heat flow, in situ stress and seismological measurements (e.g. Zoback et al., 1987; Lachenbruch and Sass, 1992) suggest that the San Andreas fault is mechanically weak in an absolute sense as well as in a relative sense. These low shear stress value are in accordance with laboratory measurements of the frictional strength only if anomalously high pore pressure exists in a mature fault zone such as the San Andreas (Byerlee, 1990; Rice, 1992). In subduction zones and convergent margins, Ocean Drilling Project (ODP) data indicate that overpressure is common in and below a number of accretionary wedges (Bangs et al., 1990; Byrne and Fisher, 1992).
A number of mechanisms have been proposed for the generation and maintenance of pore pressure excess. Laboratory measurements by Blanpied et al. (1992) show that permanent porosity reduction induced by hydrothermal processes can significantly lower the permeability and maintain a pore pressure excess, resulting in an anomalously low frictional strength. Rice (1992) appealed to the continuous influx of fluid from the ``ductile root'' of the fault zone to generate a nearly lithostatic pore pressure. To maintain the pore pressure excess, the permeability is required to have a strong sensitivity to normal stress, comparable to what has been observed for clay gouge and fractures (David et al., 1994). A number of related studies will appear in a special JGR issue on ``Mechanical Involvement of Fluids in Faulting''.
In his analysis, Rice (1992) emphasized that the stress field inside the gouge zone might be significantly different from the tectonic stress field experienced by the surrounding country rock. The implication is that one cannot relate the applied stress field directly to the stress and deformation within the gouge zone unless the rheology of the gouge material is completely specified. A number of theoretical analyses have been made assuming Coulomb plasticity for the gouge material (Byerlee and Savage, 1992; Scott et al., 1994), and some attempts were made to relate microstructural observations to the development of shear localization (Logan et al., 1992; Gu and Wong, 1994) with theoretical analyses.
Friction constitutive relation continues to be a focus of investigation, especially in relation to velocity dependence and stick-slip instability. Most of the experimental work has been conducted within the framework of the Dieterich-Ruina type of rate and state dependent friction law. New data on serpentine (Reinen et al., 1992) highlight the complexity of the frictional sliding behavior, with velocity dependence which varies with conditions of velocities and normal stress. Several different frictional regimes may need to be specified for a given rock-gouge system. The frictional response to perturbations in normal stress were measured by Linker and Dieterich (1992). The observed response to a perturbation in normal stress mimics that to velocity perturbation, with the implication that an additional constitutive parameter should be specified and that the frictional stability behavior is also sensitive to normal stress history. New measurements of frictional constitutive parameters at elevated temperatures were obtained by Blanpied et al. (1991) and Chester (1994). Assuming that the micro mechanisms are thermally activated and follow an Arrhenius relationship (i.e., exponential dependence on temperature), Chester (1994) formulated a modified friction constitutive relation which captured some of the key attributes of observed temperature dependence.
A number of related studies on friction constitutive behavior will appear in a special volume of Pure and Applied Geophysics on ``Faulting, Friction and Earthquake Mechanics''(Marone and Blanoied, 1994). Overall the quasi-static data on frictional sliding behavior suggest that the friction constitutive relation for geologic material under crustal conditions of pressure and temperature can be very complicated. Not surprisingly, recent laboratory measurements of the transition from stable sliding and stick-slip instability also show complex dynamical behavior typical of nonlinear systems (Gu and Wong, 1994b). That such complexity is necessary for modeling the temporal evolution of seismicity during an earthquake cycle was illustrated by Dieterich (1994) who successfully used a rate and state friction relation to simulate the decay of aftershock activity. The interplay of nonlinear frictional behavior and spatial heterogeneity was analyzed by Rice (1993), who showed that whether a spatio-temporally complex slip pattern would develop or not is highly sensitive to how the continuum system is discretized and also how inertia is incorporated into the model.
To scale the laboratory measurements to crustal scaling faulting, one needs a fundamental understanding of the micromechanics of frictional sliding. Important technical advances were made by Dieterich and Kilgore (1994), who obtained direct microscopic measurements of the density and size distribution of frictional contacts during slip, which can potentially be related to phenomenological parameters in the constitutive relation. As frictional sliding develops, wear processes operate to generate a gouge zone which progressively widen with cumulative slip. The micromechanics of wear and its dependence on fault surface topography was systematically investigated by Wang and Scholz (1994), and the scaling of critical slip parameter with gouge thickness was analyzed by Marone and Kilgore (1993).