The investigators propose to analyze experimentally the processes of initiation, acceleration and self-healing of seismic slip at a patch on the surface of a fault. The experiments will be conducted on a new rotary shear apparatus that allows testing frictional sliding along rock blocks for large slip (a few meters), slip-velocity of ~1 m/s, and normal stress up to 30 MPa. Further, the design of this apparatus allows application of a finite amount of stored energy on a fault surface (up to 107 J/m2) for a short period of time (up to 5 s). Their preliminary results show that these conditions can generate an Earthquake-Like lip Event (ELSE) along the rock sample with rise time <0.1 s, slip velocity up to 1 m/s, slip-distance up to 3 m, and self-healing (strength recovery and rupture arrest). They propose to use the unique capabilities of this apparatus for an extensive series of ELSE experiments that are analogous to earthquake processes at a fault patch under in-situ loading conditions. These experiments will allow analyzing fundamental earthquake parameters, such as rise-time, weakening, self-healing, slip velocity, slip distance, heat generation and energy dissipation under laboratory-controlled conditions.
The objectives of the project is to improve the apparatus capabilities in two main aspects: 1. Stiffening of the loading frame to eliminate (or significantly reduce) the unstable shattering at high velocities. 2. Develop and implement the torque control system that will allow us to simulate a wide range of earthquake scenarios.
A typical earthquake starts at a small nucleus and propagates as a fast moving rupture front along a fault surface. Every patch on the fault surface is at rest before the earthquake, it is accelerated to slip velocity of about 1 m/s by the rupture front and it stops slipping when the elastic energy that was stored prior to the earthquake has been dissipated. Thus, the patch experiences abrupt acceleration and deceleration over periods from a fraction of a second to a few seconds. During this period of intense acceleration/deceleration, the patch friction changes dramatically without steady-state velocity. On the other hand, typical experimental studies of fault friction are designed to determine the steady-state friction that probably does not realized during earthquakes.
The proposed research eliminates this fundamental experimental limitation by utilizing the unique capabilities of an apparatus that was built recently in University of Oklahoma. The apparatus is suitable to simulate earthquake-like events as it can load a laboratory rock patch by energy stored in a massive flywheel (225 kg). This unique, advanced design of the our experimental system allows simulating earthquake rupture processes under in-situ conditions of (1) high stress and high velocity; (2) finite energy supply; and (3) stress and velocity control. The proposed experiments will provide better links between experiments, theory and seismic concepts, and, by doing so, will significantly advance the understanding of earthquake rupture processes, earthquake energy balance, physics of fault weakening, and the scaling of slip rates, slip magnitude, and radiated energy.
Motivation and Approach A large earthquake initiates at a small nucleation area and grows as a propagating rupture front (Fig. 1). The propagating front activates a multitude of fault patches that undergo intense deformation. When the front encounters a patch, the patch fails and starts slipping while releasing the energy that was stored in the surrounding rocks before the earthquake. Typically, once slip initiates, the patch strength drops in a process called dynamic weakening. The patch first accelerates, then slips and eventually stops after a period up to a few seconds. This model is widely accepted, yet the (1) fault-strength evolution, (2) mechanisms of dynamic weakening, and (3) energy dissipation processes are poorly constrained. The present project focuses on two aspects of large earthquake slip: the effect of fast loading by the earthquake rupture front (this effect has never been tested experimentally), and the mechanism of dynamic weakening (which was studied in many previous investigations, but it is still poorly understood for earthquake conditions). We used a dedicated rotary shear apparatus in the University of Oklahoma to shear experimental faults at slip velocity and slip distances that approach seismic conditions. The effects of fast rupture loading during large earthquakes We simulated the complex process of fault-patch slip by abruptly loading of an experimental fault with a spinning flywheel. We refer to this approach as "Earthquake-Like-Slip-Event" or ELSE. The spontaneous evolution of strength, acceleration, and velocity in these experiments (Fig. 2) indicated that the new procedure of ELSE provides good proxy for fault-patch behavior during earthquakes of moment magnitude (Mw) = 4 to 8. What did we learn from these experiments about large earthquakes? First, the abrupt, impact loading by the earthquake front weakens the fault patch very quickly (a fraction of a second, Fig. 2), much faster than observed in previous, standard experiments. Second, when the stored energy dissipates, the experimental fault quickly strengthens and heals, in agreement with many seismological field observations. Third, we found that the law of dynamic weakening, which expresses the weakening as a function of slip velocity and slip distance, strongly depends on the loading style (Fig. 3). Thus, as this law is one of the central modeling parameters of seismic hazard evaluation, the use of weakening law that were derived from standard experimental loading may lead to inaccurate estimate of ground motion during large earthquakes. Mechanism of earthquake dynamic weakening Our project searched for dynamic weakening by running the experimental faults at constant velocities up to 1 m/s, and then examining the fault-zone with ultra-high resolution tools. The analyses utilized transmission electron microscopy (TEM), the scanning electron microscopy (SEM), and atomic force microscopy (AFM), that display rock structures down to the nanometer scale. These tools revealed that during seismic slip conditions, the fault spontaneously smoothens, and the slip localizes along these highly smooth, reflective surfaces; this smoothening led to significant strength reduction in the experiments (Fig. 4). Further, tiny powder rolls composed of nanometer-scale broken rock powder, about 1 micron in diameter, spontaneously developed along the smooth surfaces (Fig. 5). The fault with these rolls acted as a roller-bearing, similarly to roller-bearing which are used in industrial machinery for lowering frictional resistance. We found that these processes are active along experimental faults of wide range of compositions, and at wide range of stresses, thus we envision that they are active along natural faults during large earthquakes. In the near future, we will examine natural faults to find evidence of these mechanisms. Intellectual merit and broader impacts The understanding of earthquake physics is based on seismic observations, theoretical modeling, and experimental results. Relevant experimental data are essential for derivation of useful models and interpretation of seismic observations. We think that the unique, advanced design of our apparatus generated realistic experimental simulation of earthquake rupture processes under in-situ conditions. The results provide better links between experiments, theory and seismic concepts. Our current applications of these results to dynamic earthquake rupture, which is based on direct experimental observations reveals rupturing features that are in good agreement with seismic observations. Thus, the present results are likely to benefit society by advancing our understanding of the mechanics of earthquakes and their capability of producing high-frequency seismic radiation. Further, better understanding of earthquake rupture will improve the estimation of near-fault ground motion, which is of central importance in reducing seismic hazard for major facilities and residential dwellings.
