As a first step, we examined 2 deep-focus earthquakes, namely by the Bolivia (June 9, 1994 Mw8.2 623km) and Fiji earthquakes (March 9 1994 Mw7.6 563km). Kinematic models and other source parameters were determined by Antolik (1996).
To examine the slip weakening constitutive relationship for deep-focus earthquakes, at first, we calculated, or collected from the literature information, about the seismic moment, the radiated seismic energy, faulting area, average slip, and static stress drop. Second, following the method of Kanamori et al.(1998), the possible temperature rises for all of the events were estimated using the parameters collected. To compute the temperature rise, it was necessary to assume the width of the shear zone, which was inferred from the rise time (e.g. *Kanamori et al.*, 1998), or the scaling of anti-crack width to shear-slip from laboratory measurements (e.g. *Bouchon and Ihmle*, 1999). The radiated energy, scalar seismic moment and stress drop were used to determine the seismic efficiency, which we correlated to independent observations of the rupture velocity in these events. From this analysis the events can be categorized in terms of the degree of their dissipative character. Third, to determine the temporal displacement and stress fields, we used the finite-difference method with a kinematic representation of the source process as a boundary condition (e.g. *Ide and Takeo*, 1997). We used kinematic models which have been well determined (*Antolik*, 1996). The finite-difference method is used to solve the equation of motion with an isotropic, linear-elastic stress constitutive relationship to determine the velocity/stress wavefield due to the prescribed finite-source slip. We used the absorbing boundary condition on all surfaces because all of the events are at great depth. The stress drop models obtained from this study were used to estimate the frictional energy and the possible temperature rise during faulting to assess whether frictional melting could have played a role in the other large deep earthquakes. For this analysis, we followed the method introduced in Bouchon and Ihmle (1999).
The determined velocity/stress fields at the third step were use to examine the stress-slip constitutive relationship, the static stress drop, and the normal stress perturbation for each of the events. The magnitude of the dynamic shear and normal stress over the fault in advance of the rupture provides information on the level of deviatoric stress that triggers mechanisms by which ruptures continue to grow; in the case of the Bolivia earthquake possibly outside the region of the metastable slab core.

Berkeley Seismological Laboratory

215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760

Questions or comments? Send e-mail: www@seismo.berkeley.edu

© 2005, The Regents of the University of California