Experimental strike-slip earthquakes (“ricequakes”)

During large strike-slip earthquakes, the surface displacement can be measured from correlation of satellite images acquired before and after the event. These measurements allow quantifying the total surface displacement and knowing how it distributes between on and off-fault deformation, which is important for seismic hazard assessment. These measurements are highly variable, partly due to the sparsity of natural examples. This research focuses on analogue modelling to study parameters affecting the surface displacement in a controlled environment. However, current analogue models do not address the issue of surface deformation associated with strike-slip earthquakes. Instead, models using granular materials, such as sand or clay, rather focus on the surface deformation associated with continuous deformation without earthquakes, while models using rigid materials, such as foam or gelatin, focus on the localization and frequency of seismic events without looking at surface deformation.

To analyze the long-term deformation of seismogenic strike-slip faults from analogue experiments, we used a box composed of two juxtaposed PVC plates simulating a vertical and linear basement fault. A strike-slip fault emerges from this discontinuity. The box is filled with rice and rubber pellets in order to produce both aseismic displacements and earthquakes along the evolving strike-slip fault. Dry rice is a stick-slip granular material already used in subduction experiments. We used a twice broken rice with peak, dynamic, and reactivation friction values of respectively 0.78, 0.67 and 0.68 at a constant shear velocity. Those values decrease when the shear velocity is increased (the parameter “a-b” = -0.012 in the rate friction law of the rice). Since rice is too rigid to produce a measurable elastic strain release, we added a basal layer of fine rubber pellets between the basal PVC plates and the rice layer in order to store elastic strain. Applying a constant displacement velocity and taking photos every 25 micrometers of displacement, we follow the surface displacements through image correlation.

We observe that the average displacement along a profile parallel to the fault (taken at a distance of the basal fault corresponding to half the rice layer thickness) only matches the applied displacement when averaged over the whole experiments. Indeed, during most of the experiment, the observed incremental displacement is lower than the applied one, but from time to time it catches up during discreet events that produce large displacements of up to four times the applied incremental displacement. We interpret these events as seismic events. Hence, the evolution of cumulative displacement with time exhibits some phases of creep, more or less at the same rate as the input rate, during the inter-seismic period, and phases of sudden displacements corresponding to sudden release of elastic strain, i.e., earthquakes. However, we never observe a complete blocking phase (sticking phase). These first results show that it is possible to build an experimental strike-slip fault system in a granular medium with a low normal stress, i.e. a free surface, that produces extended periods of partial stress loading during creep phases, alternating with period of sudden stress release during displacement phases.