Oral Presentation

Observations suggest that geophysical activity in the volume of crust surrounding the main fault may change in the years, days and hours before a large earthquake. However, some large earthquakes do not appear to produce detectable signals, casting doubt on the feasibility of earthquake prediction. Two aspects of crustal fault system monitoring include 1) identifying the location relative to the main fault, or faults, that produce the strongest precursory activity, and 2) constraining the direction of ground motion that may produce the strongest precursors, such as horizontal motion parallel or perpendicular to the main fault strike, or uplift/subsidence. We use discrete element method models to constrain these two aspects of precursory deformation. Characteristics of a fault network that may influence the location and nature of precursory deformation include the fault geometry and the relative strength of the host rock to the faults. We vary the fault strike-perpendicular spacing of two underlapping parallel faults from releasing to restraining step over configurations embedded in a shear zone. We also vary the amount of diffuse preexisting damage spread throughout the model, thereby changing the host rock strength. One may expect that the strongest precursory deformation would occur in the region that ultimately hosts the faults. However, preceding the development of these parallel faults, the mean velocity of the particles within and outside these future fault zones are relatively similar to each other, and the highest magnitudes of velocity (>75% of the maximum at a model time) occur with similar frequency inside and outside these zones. Examining the components of the particle velocity vectors indicates that the fault strike parallel deformation generally provides the strongest contribution to the overall velocity magnitude, consistent with imposed shear loading parallel to the fault strike. However, in models with coplanar faults and step over gaps of 20% of a fault length, the fault strike perpendicular deformation provides a similar contribution to the overall magnitude as the fault strike parallel deformation. Both of the horizontal components of the deformation field tend to provide stronger precursory signals than the vertical component. This work thus highlights the importance of crustal monitoring of the volume of rock surrounding the main fault, and particularly the horizontal components of the ground deformation.