Friction and seismic instability: scaling from lab to natural faults

Though the aim of laboratory experiments on friction is to improve our understanding of earthquakes, it is not straightforward to compare the small scale measurements obtained under controlled conditions, to the seismological and geodetical observations obtained at the natural scale. In fact, several frictional parameters are difficult to retrieve with seismological tools. One parameter which can be both measured in the lab and estimated from real earthquakes - at very different scales - is fracture energy G. This offers a common ground to investigate whether the frictional dissipation processes observed in the laboratory have much in common with those occurring during natural earthquakes. Surprisingly, both values do scale with slip in a compatible fashion for a large portion of the magnitude spectrum, but tend to diverge for earthquakes with slip > 10 cm. Promising interpretations based on the diffuse nature of damage in real earthquakes may help to reconcile these differences and explain the similarities.