Cohesion versus friction in controlling the long‐term strength of a self‐healing experimental fault

Abstract

The Coulomb's failure criterion, which postulates that failure occurs along a fault plane when the applied shear stress overcomes a resistance made of two parts of different nature, cohesion, and friction, remains the standard conceptual framework of faulting mechanics. More recently, rate‐and‐state friction laws became the main modeling tool of the seismic cycle. These laws implicitly assume that only frictional resistance sets the fault strength and its evolution. We therefore raise the question of the role of cohesion and related healing/sealing mechanisms on fault mechanics. We designed an original laboratory experiment based on a model material, an ice thin layer sitting on top of a water tank and mechanically deformed at various rates with a circular Couette‐like geometry. This allowed sliding along the fault plane over arbitrarily large slip distances, and analyzing the competition between faulting and cohesion‐healing rates, whereas frictional resistance, resulting from the fault roughness, is limited in magnitude. We show that cohesion‐healing plays an essential role in the time evolution of the interface strength under constant loading rate. In particular, the magnitudes of shear stress fluctuations are observed to be temperature and velocity weakening. The present experiment shares several properties similar to that of active faults during seismic cycles, suggesting that cohesive healing could control some of these features: scale invariance properties, Gutenberg‐Richter law of rupture event size distribution, Omori's law of energy dissipation, at least after major ruptures, time asymmetry in energy dissipation, and periods of creep under high shear stress alternating with major earthquake‐like ruptures.

Earthquake Stress Drop in the Charlevoix Seismic Zone, Eastern Canada

“…Lay and Wallace (1995) attributed higher stress drop values for intraplate events to shorter source duration relative to fault length and/or larger moment release per fault length. Fault frictional strength can be enhanced by cohesion healing typical of intraplate tectonic settings (Liu & Stein, 2016;Weiss et al, 2016). If we consider slip on a preexisting fault under the assumption of strict adherence to Coulomb failure criteria, it would imply that seismogenic intraplate faults may have higher coefficients of static friction.…”

“…If we consider slip on a preexisting fault under the assumption of strict adherence to Coulomb failure criteria, it would imply that seismogenic intraplate faults may have higher coefficients of static friction. Fault frictional strength can be enhanced by cohesion healing typical of intraplate tectonic settings (Liu & Stein, 2016;Weiss et al, 2016). The hypothesis is that intraplate seismogenic faults typically have slow loading rates, leading to less frequent slip, or, equivalently, longer contact time for developing stronger asperities, hence stronger faults.…”

Micro‐Seismic Monitoring of a Shear Fault Within a Floating Ice Plate

“…In this set‐up, cohesion / healing refers to processes at the micro‐scale which cause the (pressure‐independent) shear resistance to increase with time, e.g., asperity welding and crack sealing. These processes were shown by Weiss et al () to be thermal weakening in the case of ice, as fault healing is controlled by refreezing. Because of this healing, the shear resistance can be significant even in presence of a low normal stress.…”

“…The only observable available to Weiss et al () was a global (i.e., at the scale of the whole fault) torque measurement so that the local, spatially variable, complexity of stress release could not be investigated. Also, the temporal and amplitude resolutions did not allow to distinguish small ruptures: the torque increment measured between two successive time steps results from possible micro‐ruptures at different positions, as well as strengthening through healing, without the possibility to distinguish between those processes.…”

“…The focus of our study is thus to further develop the observations of Weiss et al (), by adding a seismic monitoring system (accelerometers). This allows isolating the damage (micro‐ruptures) and healing processes, and characterizing their distributions in time, space and size.…”