HR: 16:15h
AN: G22E-02 INVITED [PDF]
TI: Elastic and Viscoelastic models of Crustal Deformation in Great Earthquake Cycles
AU: * Wang, K
EM: wang@pgc.nrcan.gc.ca
AF: Geological Survey of Canada, Pacific Geoscience Centre,
9860 W Saanich Rd, Sidney, BC, V8L 4B2
Canada
AU: * Wang, K
EM: wang@pgc.nrcan.gc.ca
AF: School of Earth and Ocean Sciences, University of Victoria, Victoria, BC, V8W 3P6
Canada
AU: He, J
EM: he@pgc.nrcan.gc.ca
AF: Geological Survey of Canada, Pacific Geoscience Centre,
9860 W Saanich Rd, Sidney, BC, V8L 4B2
Canada
AU: Hu, Y
EM: yhu@pgc.nrcan.gc.ca
AF: School of Earth and Ocean Sciences, University of Victoria, Victoria, BC, V8W 3P6
Canada
AB:
Ideally, a model of subduction zone earthquake cycles should include tectonic loading, fault friction, and viscoelastic
stress relaxation. Such a model is not yet available. The loading mechanism is rarely addressed. A model does not address
tectonic loading if some part of the fault is assigned a slip rate. Models based on rate and state dependent friction laws
are useful in demonstrating how seismic fault slips may occur and stop as a result of the interplay between fault frictional
behavior and system rigidity. For comparison with geodetic observations, however, the most widely used models treat the fault
motion in a purely kinematic fashion, that is, the fault slip (or state of locking) is estimated from surface observations
regardless of the loading mechanism and friction properties. These include the forward and inverse elastic dislocation and
viscoelastic models. Without addressing the loading mechanism, extra care should be taken to ensure that the assigned or
estimated fault motion is physically valid. It is often difficult to distinguish between contributions to surface deformation
from fault motion and from stress relaxation of the rock material. The same deformation observations can be explained by
different models, and any published model merely portrays one particular understanding of the processes being studied. It is
usually assumed that aseismic slip of the fault, such as "afterslip" and interseismic silent slip, is of the time scale of
days to years, but viscoelastic stress relaxation of the system has a time scale of decades to hundreds of years. On the
basis of the time scale argument, we modeled earthquake cycles at the Cascadia and Chile subduction zones using a 3-D
spherical finite element viscoelastic model with a mantle viscosity of about 10$^{19}$ Pa s. Contemporary crustal deformation
of the Cascadia forearc, 300 years after a great earthquake, can be explained by the viscoelastic model. The observation
that GPS sites 300-400 km landward of the rupture region of the 1960 great Chile earthquake are presently moving seaward,
opposite to the motion of the coastal sites, can be explained by stress relaxation. Elastic models can also fit most of the
observations by assuming that all deformation is due to fault motion. The fault motion thus determined effectively includes
contribution from stress relaxation.
DE: 1236 Rheology of the lithosphere and mantle (8160)
DE: 1242 Seismic deformations (7205)
DE: 3210 Modeling
DE: 7209 Earthquake dynamics and mechanics
DE: 8150 Plate boundary--general (3040)
SC: Geodesy [G]
MN: 2003 Fall Meeting