HR: 10:25h
AN: S41G-01 INVITED [PDF]
TI: Earthquake Fracture Energies and Weakening of Faults by Thermal Pressurization of Pore Fluid
AU: * Rice, J R
EM: rice@esag.harvard.edu
AF: Dept. of Earth and Planet. Sci. and Div. of Engin. and Appl. Sci., Harvard Univ., 224 Pierce Hall,
Cambridge, MA 02138
AB:
Seismic inferences of fracture energy $G$ constrain how fault strength degrades during slip and allow testing of candidate
physical mechanisms. Recently $G$ has been estimated by interpreting parameters from seismic slip inversions within a
self-healing rupture model (Rice, Sammis and Parsons, 2003), and by studying the scaling of radiated energy and stress drop
with earthquake size (Abercrombie and Rice, 2003). Those and earlier studies suggest that for larger events (slip $> 0.1$ m),
$G$ ranges from 0.1 to 10 MJ/m$^2$ with average of 2-4 MJ/m$^2$. There is a clear trend for $G$ to increase with slip over
the broad range from mm to m slip.
Sibson-Lachenbruch thermal pressurization of pore water is examined as a possible general fault weakening mechanism for large
crustal events. For adiabatic and undrained conditions, with strength given by the effective stress law with a constant
friction coefficient $f$, the thermal properties of water in this context (Lachenbruch, 1980; Mase and Smith, 1988) lead to
$G = 1.7 (\sigma_n - p_o) (1 + r) h$. Here $h$ is shearing zone thickness, $\sigma_n$ is normal stress, assumed constant
during slip, $p_o$ is ambient pore pressure, and $r$ is the ratio fractional volume change of pore space per unit pore
pressure increase divided by the compressibility of the pore fluid. Dilatancy is neglected; if confined to only the early
phases of slip, it decreases $p_o$ from ambient and so increases $G$, but the effect may be modest. The model predicts
exponential decay of strength with slip, with e-folding slip distance $1.7 (1 + r) h / f$. The total temperature rise in K is
$\approx 0.6 (1 + r) (\sigma_n - p_o)$ where the latter factor is in MPa.
Estimating $r =$ 1-2 and evaluating $\sigma_n - p_o$ as overburden minus hydrostatic pore pressure at 7 km as a
representative centroidal depth for large crustal events, we obtain $G \approx$ 1-6 MJ/m$^2$ for $h =$ 2 to 10 mm. Shear
zone thicknesses towards the lower end of such a range are suggested by recent field studies (Chester and Chester, 1998),
which identify a narrow principal failure zone within a 10s of mm thick ultracataclastic core. Temperature rises are 150-230
K and preclude melting, although rises at 14 km may be twice larger and temperature there could reach the onset of melting.
That predicted $G$ compares favorably with the above seismically inferred range. For $f = 0.6$, the slip needed to fully
experience the weakening (as twice the e-foldng slip) is of order 0.02-0.17 m, so many smaller slip events will not
experience full weakening. $G$ values estimated seismically for them should be smaller, as seems to be the case (Abercrombie
and Rice, 2003). Those smaller $G$ values may be influenced by more modest weakening mechanisms like studied in rate and
state friction, which are likely to dominate at sub-mm slip.
Corrections to the thermal pressurization estimate of $G$ for large events arise because the shearing zone may not retain its
water during failure, so that strength is not fully reduced, and from the feature that fast-moving ruptures fronts generate
enough stress off the main fault plane to cause secondary failure within the bordering damage zone (Poliakov et al., 2002;
Rice et al., 2003). The latter effect becomes particularly strong as rupture speed approaches its subsonic limit (Rayleigh
speed for mode II); it consumes energy and increases $G$, but also leaves freshly failed material immediately adjacent to the
shearing zone. That will promote drainage, for which an elementary model will be described. Complete strength loss then
cannot be achieved by the pore fluid mechanism and temperature continues to rise so that ultimately melting occurs if slip
becomes large enough.
DE: 5114 Permeability and porosity
DE: 7209 Earthquake dynamics and mechanics
DE: 7260 Theory and modeling
DE: 8020 Mechanics
DE: 8164 Stresses--crust and lithosphere
SC: Seismology [S]
MN: 2003 Fall Meeting