abstract

A frictional sliding model consisting of lumped masses and elastic springs has been constructed to simulate a seismic strike-slip fault. The model was capable of generating stick-slip events with some statistical features similar to those of natural earthquakes. The patterns of strain buildup and release associated with a long sequence of shocks are described in the hope that they may reflect to some extent the behavior of a real seismic fault: (1) Shocks of greatly different sizes were generated, although the friction inhomogeneity along the fault was small. (2) The recurrence relation for the shock sequence resembled that for earthquakes. (3) Large shocks occurred at irregular time intervals and released most of the strain energy of the sequence. (4) The logarithm of the product of average particle displacement and rupture length increased linearly with shock size (logarithm of strain-energy release). (5) Shocks of the same rupture length sometimes had considerably different particle displacements. (6) The shear stress drops increased with shock size from about 1 per cent of the preshock stress for small shocks to 40 per cent for the largest shock; the occurrence of small and moderate shocks did not greatly perturb the operating stress level. (7) The preshock stresses and energy densities along the ruptured segments did not increase much with the shock size; larger shocks occurred mainly when a longer segment of the fault was stressed simultaneously to near the breaking level. (8) Ruptures preferred to propagate in the forward direction owing to a built-in asymmetry of the model fault. (9) Particle displacements were approximately finite ramp functions of time with periods of acceleration at the beginning and deceleration at the end. (10) Particle velocities increased with particle displacements, ranging from 2 to nearly 30 cm sec−1. (11) Rupture propagation velocities scattered considerably but tended to increase slightly with rupture length. (12) Aftershocks were few but occasionally occurred at places where slippage was small during the main shocks. The lack of a significant number of aftershocks in the model probably results from the absence of such time-dependent elements as viscosity or moving pore fluids that operate along natural faults.

First Page Preview

First page PDF preview
You do not currently have access to this article.