A refined lattice solid model to simulate earthquakes and localisation phenomena using parallel computers

Abstract

The particle-based lattice solid model developed to study the physics of rocks and the non-linear dynamics of earthquakes is refined by incorporating intrinsic friction between particles. The model provides a mean to study the causes of seismic wave attenuation, as well as frictional heat generation, fault zone evolution and localisation phenomena. A modified velocity-Verlet scheme is developed that allows friction to be precisely modelled. This is a difficult computational problem given that a discontinuity must be accurately simulated by the numerical approach (ie. the transition from static to dynamical frictional behaviour). An efficient numerical integration is achieved by using an adaptive adjustment of the time step increment. The total energy is calculated and verified to remain constant to a high precision during simulations.In order to allow large scale simulations to be performed, a high performance must be obtained when implementing the model. This is achieved by refining the model according to the computer architecture and by developing an algorithm to fully use the features and full potential of the given parallel super-computer.Field observations show that on the San Andreas fault, the heat generated during earthquakes is at least five times smaller than the theoretical prediction estimated by assuming the fault friction equals the rock coefficient of friction measured in the laboratory. So far, no unflawed explanation of this heat flow paradox has been proposed. By applying the model to study heat generation during earthquake events, numerical simulations show that a rolling type mechanism is responsible for low heat during numerical simulations when a model fault gouge is present.During numerical experiments when a weak and heterogeneous fault zone is specified, slip becomes localised in a narrow band and features are similar to those observed in recent laboratory experiments. This includes formation of Reidel shears and localisation of shear along narrow bands. As in the laboratory experiments, the decrease of the fault strength correlates with a decrease of the gouge thickness. After a large displacement, a re-organisation of the fault gouge is observed where the slip becomes localised along a very narrow shear zone. Localisation and re-organisation enhance the rolling type micro-physical mechanism that was responsible for low heat generated during the numerical experiments involving a pre-specified fault gouge.Numerical experiments show that the model can be applied to the study of earthquake dynamics, the stick-slip instability, heat generation, fault zone evolution and localisation phenomena. Such experiments may lead to a conclusive resolution of the heat flow paradox and an improved understanding of earthquake precursory phenomena and dynamics.