Abstract

Natural earthquakes are usually a sign of rocks breaking along faults, relieving stress in the outer part of the Earth. In this talk, I will discuss ideas about the links between earthquakes and the strength and frictional behaviour of these faults. The classical view is that earthquakes only contain information about the change in strength of a fault during the rupture, which could be large or small compared to the ambient deviatoric stress levels. However, it is now widely thought that earthquakes are due to dynamical weakening of the fault during the early stages of its failure and slip, with the strength dropping to negligible values at seismic slip velocities. In this case, the change in strength of the fault during rupture is also a measure of the total magnitude of the driving stresses. I show that measured stress drops during earthquakes are typically very low, with a log normal distribution mainly in the range 1 – 10 MPa. This distribution is essentially the same regardless of whether earthquakes are grouped by magnitude, focal mechanism, or depths, providing the basis for the well-established earthquake scaling laws. I argue that this creates a problem for dynamical weakening models, because it requires either very low ambient deviatoric stresses in the seismogenic parts of the crust, and hence weak rocks, or very prescriptive rupture behaviour in which normal stress has a confounding effect, and hence may not be sufficiently general. This raises the question of whether or not laboratory-scale experiments on dynamical friction are a good description of much larger scale rupture behaviour in natural earthquakes.