HR: 14:20h
AN: T12E-03 [PDF]
TI: Relationship Between Mapped Fault Stepovers and Earthquake Fault Planes at Depth
AU: * Zoback, M
EM: zoback@usgs.gov
AF: U. S. Geological Survey, 345 Middlefield Road, MS. 977, Menlo Park, CA 94025 United States
AB:
The San Andreas fault system in the San Francisco Bay area is complex, consisting of several sub-parallel strands with
numerous stepovers and bends. Source characterization of future likely earthquakes in the Bay Area requires understanding of
the role these fault stepovers and bends play in fault segmentation. Absent aftershocks or microseismicity, it is difficult
to determine the subsurface fault geometry. To better understand the role of geometric complexity in controlling earthquake
ruptures, I have examined data from several recent major strike-slip earthquakes. The 1995 Mw6.9 Kobe earthquake originated
within a 5-km right (dilatational) step in a right-lateral fault and produced a bilateral rupture. Wald (J. Phys. Earth,
1996) showed that the hypocenter occurred at the intersection of the two well-constrained, offset fault planes that were
steeply dipping toward each other. A 3-km right step in the San Andreas offshore from the Golden Gate inferred from seismic
and potential field data consistently produces normal faulting microearthquakes. The 1906 earthquake with its bilateral
rupture is thought to have originated along the offshore segment of the San Andreas fault near the Golden Gate; by analogy
with the Kobe earthquake, we have suggested that the 1906 also nucleated within a stepover region.
Modeling dynamic rupture propagation constrained by near-fault ground motion records for the 1999 Izmit M7.4 earthquake, led
Aochi and Madariaga (BSSA, 2003) to conclude that this rupture was rapid and continuous on a smooth fault structure with a
bend of only a few degrees beneath a 5-km right (dilatational) stepover mapped at the surface in the vicinity of Sapanca
lake. Similarly, aftershocks and surface faulting of the1995 Mw7.2 Landers earthquake suggest continuous rupture across a
5-km dilatational jog, utilizing an oblique fault connecting the 2 offset fault segments; whereas, the rupture across a
second, 2-km dilatational jog appears more diffuse, with no continuous through-going structure (Felzer and Beroza, GRL,
1999). Relocated East Bay microseismicity using the double-difference technique indicates a continuous zone consisting of
straight, near-vertical fault planes connecting the Calaveras and Hayward faults across a 5-6 km left (restraining) step
(Waldhauser and Ellsworth, JGR, 2002; Ponce et al., EOS, 2003; Simpson et al., EOS, 2003). These near-vertical planes are
well-defined below 5 km depth, in contrast to a complex pattern of surface fault traces with no through-going, connecting
structure. The data suggest that, for at least some fault stopovers, the earthquake rupture surface at depth may be far
simpler and more continuous than surface fault traces suggest and that fault stepovers and bends mapped at the surface do not
necessarily represent segment boundaries or major energy barriers to rupture at depth. These observations raise intriguing
questions about how, absent microseismisty, to determine if a fault stepover may have a simple connection at depth and what
parameters, e.g. step size, total displacement, rock type, relative fault strength, etc. might control the depth variation in
structural style
DE: 7209 Earthquake dynamics and mechanics
DE: 7223 Seismic hazard assessment and prediction
DE: 8123 Dynamics, seismotectonics
SC: Tectonophysics [T]
MN: 2003 Fall Meeting