Is complex fault zone behaviour a reflection of rheological heterogeneity?

Å Fagereng,A Beall

doi:10.1098/rsta.2019.0421

Abstract

Fault slip speeds range from steady plate boundary creep through to earthquake slip. Geological descriptions of faults range from localized displacement on one or more discrete planes, through to distributed shearing flow in tabular zones of finite thickness, indicating a large range of possible strain rates in natural faults. We review geological observations and analyse numerical models of two-phase shear zones to discuss the degree and distribution of fault zone heterogeneity and effects on active fault slip style. There must be certain conditions that produce earthquakes, creep and slip at intermediate velocities. Because intermediate slip styles occur over large ranges in temperature, the controlling conditions must be effects of fault properties and/or other dynamic variables. We suggest that the ratio of bulk driving stress to frictional yield strength, and viscosity contrasts within the fault zone, are critical factors. While earthquake nucleation requires the frictional yield to be reached, steady viscous flow requires conditions far from the frictional yield. Intermediate slip speeds may arise when driving stress is sufficient to nucleate local frictional failure by stress amplification, or local frictional yield is lowered by fluid pressure, but such failure is spatially limited by surrounding shear zone stress heterogeneity.This article is part of a discussion meeting issue ‘Understanding earthquakes using the geological record’.

Highlights

Faults are classically thought to creep steadily or slip episodically in earthquakes, with more complex conceptual models involving seismogenic patches embedded within otherwise aseismic faults [1,2,3,4]
The upper, frictional regime is characterized by brittle fault rocks, including gouges, cataclasites and pseudotachylytes, where one or more discrete fault cores are surrounded by a fractured damage zone [2,10]
Potential instabilities will be dampened by surrounding viscous material and suppressed by a large nucleation length scale; unstable slip may occur if fault weakening by fluid pressurization overcomes velocity strengthening [136]—and this has been suggested previously as a mechanism for episodic tremor and slip’ (ETS) spatially separated from the frictional–viscous transition [102]

Summary

Introduction

Faults are classically thought to creep steadily or slip episodically in earthquakes, with more complex conceptual models involving seismogenic patches embedded within otherwise aseismic faults [1,2,3,4]. The geological and mechanical observations outlined above for a homogeneous fault zone relate to, but do not fully describe, the fault’s depth-dependent seismic behaviour This is further described by velocity-dependence of friction, denoted by the parameter (a − b), defined by τ = σn[μ∗ + (a − b) log(V/V∗)] in the steady-state form of the rate and state friction law [9,16,17,18,19,20] (figure 1c). Stable areas cannot nucleate earthquakes without a dynamic load, but earthquakes may propagate into a conditionally stable field if the dynamic velocity step is sufficient This framework infers that rocks in the deep viscous regime are velocitystrengthening because of the stable sliding observed there, and predicts that stable sliding prevails at very shallow depths where poorly lithified rocks accommodate displacement by granular flow involving dilatancy-hardening [30]. There is no distinct pressure (P) or T regime associated with slow earthquakes, but they do seem associated with conditional stability at or close to the seismic–aseismic transition [20,46,47]