|The San Andreas fault|
In a paper*** in this weeks Science, Goldsby and Tullis point out that when two rock surfaces are brought together, they only touch at a few small contact points compared to the total surface area. These points have average sizes of tens of microns. Like the floor under a spiky woman's high heeled shoe, the stresses are concentrated on these points, typically having local stresses of 10 GPa for even modest average stresses. When these microscopic contacts are sheared in an earthquake, very high temperatures can result. If the shearing rate is slow, heat can diffuse away from these points and the temperature remains low. However, if the shearing rate is high, there is no time for diffusion and the temperature at the points increases, sometimes to melting temperatures.
The results are based on laboratory experiments on a number of different rock types. In these experiments rocks were sheared past each other at velocities up to 0.4 m/s over distances up to 45 mm. The results showed that the friction coefficient decreased dramatically when sliding velocity exceeded about 0.1 meters per second. Visual inspection of the samples after the sliding experiment showed that a thin layer of gouge (melted rock and crushed rock) had formed. The gouge layer was less than 30 microns thick.
The authors propose that flash heating is the dominant mechanism of weakening in small-slip, small-magnitude earthquakes, and that it is likely to be the dominant mechanism determining the strength of a fault in the early stages of larger earthquakes. During continued slip during large earthquakes other fault-weakening mechanisms may combine with or dominate over flash heating, such as melt lubrication, gel formation, or pore-fluid pressurization.
***Goldsby, D.L., and Tullis, T.E., Flash heating leads to low frictional strength of crustal rocks at earthquake slip rates, Science, 334, 216-218, 2011.