Meter-Scale Laboratory Earthquakes Triggered by Fluid Injection and the Role of Background Stress in Mediating Runaway Ruptures

ABSTRACT: I present large-scale laboratory experiments where water is injected directly into a stressed fault that is a saw-cut interface in a slab of granite (3 m by 1. 5 m and 300 mm thick) in an attempt to examine the earthquake energy budget. The laboratory fault is subjected to 2-10 MPa stress levels, and exhibits slow and aseismic slip within a localized "nucleation zone" prior to rapid dynamic fault rupture. Slow slip associated with the nucleation of dynamic rupture is observed both for cases where water is injected onto the fault and where slip initiates spontaneously due to the external application of shear stress. The 3-m-long sample is long enough that ruptures sometimes arrest within the sample. So, using arrays of sensors, we study the conditions required for rupture arrest, a crucial element in determining whether an earthquake rupture will grow into a large, devastating event, or will remain small and harmless. The rupture can be modeled as a shear crack, and we employ fracture mechanics to explain rupture arrest. Our experiments demonstrate that the initial shear stress plays an important role in an earthquake's size. Another key unknown in earthquake mechanics is the shear fracture energy required to advance rupture. In the lab, we can measure values in the range of 1-10 J/m², but it is unclear if shear fracture energy, or, similarly, breakdown work, increases with the rupture size as it propagates into an increasingly rough fault surface. Answering such a scaling question is key to understanding the earthquake energy budget, a major impediment in our ability to accurately model and predict earthquakes.