Structure and fluid flow constraints on earthquake slip weakening mechanisms

Christopher A.J. Wibberley1

Dept. Geology and Mineralogy, Graduate School of Science, Kyoto University, Japan

1 From Jan 2003: CASP, Dept. of Earth Sciences, University of Cambridge, U.K.

cwibber@ip.media.kyoto-u.ac.jp

The long-standing idea that earthquake slip zones must dynamically weaken during slip to allow continued motion and dynamic runaway has lead to an equally long-running debate as to how such dynamic weakening may happen. Of the several mechanisms proposed, this contribution focuses on testing slip weakening mechanisms involving the role of pore water in the slip zone material. A combined field and laboratory study using an excellent exposure of the Median Tectonic Line fault zone in Japan has revealed a central principal displacement zone (approximately 10 cm wide) of very fine-grained low permeability clay gouge (k = 10-20 to 10-21 m2 parallel to foliation at effective pressures of 80 – 180 MPa). This zone is interpreted as the most recent seismogenic slip zone, and is laterally continuous across most of the outcrop. Evidence of previous slip zones such as oblique narrow shear zones either bifurcating from the principal displacement zone or being truncated attest to a complexity in previous rupture behaviour. A range of fault rocks is affected by these oblique zones, such as high-permeability cemented cataclasites and coarse gouges.

A model of slip weakening by thermal pressurization is tested whereby frictional heating of pore water in the slip zone gouge allows fluid pressure to either build-up (in a low permeability case) or dissipate by fluid escape (in a high permeability case). The modelling assumes a narrow slip zone of finite hydraulic diffusivity (constrained by the laboratory data), surrounded by a fracture-dilatant material of infinite hydraulic diffusivity. In this process, a trade-off exists between narrower slip zones causing higher rates of frictional heating (higher shear strain rates) than wider zones, and therefore faster fluid pressure build-up, against the fluid pressure dissipation being faster in a narrower zone from which fluid escape is easier. Modelling this thermal pressurization process using data from the clay gouge slip zone suggests that fluid pressure rises to lithostatic may occur in about one second. Using the data from other fault rocks such as adjacent coarse gouges input into the modelling, then only partial pressurization can occur, and only in relatively wide zones of 3 cm – 10 cm. The modelling suggests therefore that drastic slip weakening may occur on the timescales suggested from seismological evidence of large earthquakes, provided that the new rupture remains confined to the low-permeability clay gouge.