S44A-07

Rupture Velocities of Intermediate-Depth and Deep-Focus Earthquakes

Thursday, 17 December 2015: 17:30
305 (Moscone South)
Linda M Warren, Saint Louis University Main Campus, Saint Louis, MO, United States
Abstract:
The rupture velocities of intermediate-depth and deep-focus earthquakes --- how they vary between subduction zones, how they vary with depth, and what their maximum values are --- may help constrain the mechanism(s) that generate the earthquakes. I select earthquakes >70 km depth with MW ≥ 5.7 since 1990 in Middle America, South America, Tonga-Kermadec, Izu-Bonin-Marianas, Japan-Kuril-Kamchatka, and Hindu Kush and analyze the rupture directivity of the P waves using teleseismic data from global and regional seismic networks. I assume a constant rupture vector and estimate the rupture velocity relative to the local P-wave speed (vr/α). Since the same method is used for all earthquakes, the results can be readily compared across study areas. The study areas include some of the warmest and coldest subduction zones: subducting plate ages range from 9-150 Myr and descent rates range from 1-13 cm/yr. Across all subduction zones and depth ranges, for the 226 earthquakes with observable directivity and well-constrained rupture vectors, most earthquakes rupture on the more horizontal of the two possible nodal planes. However, the rupture vectors appear to be randomly-oriented relative to the slip vector, so the earthquakes span the continuum from Mode II (i.e., parallel slip and rupture vectors) to Mode III rupture (i.e., perpendicular slip and rupture vectors). Estimated rupture velocities span a wide range of values, from zero (for earthquakes with no observable teleseismic directivity) to approaching the local P-wave speed. The range in estimated rupture velocities is similar between all subduction zones and all depths, suggesting that local slab heterogeneity plays a more important role than temperature in determining rupture velocity. The supershear ruptures are associated with earthquakes closer to Mode II than Mode III faulting. This observation is consistent with theoretical calculations, which limit the rupture velocity to the S-wave speed for Mode III rupture but the P-wave speed for Mode II rupture.