No abstract

We present a detailed velocity model across the 1.1 billion year old Midcontinent Rift System (MRS) in central Lake Superior. The model was derived primarily from onshore‐offshore large‐aperture seismic and gravity data. High velocities obtained within a highly reflective half‐graben that was imaged on coincident seismic reflection data demonstrate the dominantly mafic composition of the graben fill and constrain its total thickness to be at least 30km. Strong wide‐angle reflections are observed from the lower crust and Moho, indicating that the crust is thickest (55–60km) beneath the axis of the graben. The total crustal thickness decreases rapidly to about 40 km beneath the south shore of the lake and decreases more gradually to the north. Above the Moho is a high‐velocity lower crust interpreted to result from syn‐rift basaltic intrusion into and/or underplating beneath the Archean lower crust. The lower crust is thickest beneath the axis of the main rift half‐graben. A second region of thick lower crust is found approximately 100km north of the axis of the rift beneath a smaller half graben that is interpreted to reflect an earlier stage of rifting. The crustal model presented here resembles recent models of some passive continental margins and is in marked contrast to many models of both active and extinct Phanerozoic continental rift zones. It demonstrates that the Moho is a dynamic feature, since the pre‐rift Moho is probably within or above the high‐velocity lower crust, whereas the post‐rift Moho is defined as the base of this layer. In the absence of major tectonic activity, however, the Moho is very stable, since the large, abrupt variations in crustal thickness beneath the MRS have been preserved for at least a billion years.

No abstract

Three‐component seismograms from local earthquakes recorded by three University of Wisconsin stations in the south moat of Long Valley caldera in 1982 have been analyzed for shear wave splitting. The average polarization direction of the first‐arriving shear waves observed at the two stations south of the resurgent dome is NNW–SSE. Time delays between the fast and slow shear waves are proportional to hypocentral distances, and the maximum velocity anisotropy inferred is 6.4%. The average shear wave polarization observed at the station above the resurgent dome is north‐south. The time delays there are not a function of hypocentral distance, and the normalized time delays (ms/km) from events at the southwest boundary of the caldera are larger than those at other azimuths, indicating maximum 9.6% velocity anisotropy. The east‐west to ENE‐WSW variation of the tensional stress fields implied by the shear wave polarizations indicates the interaction between the inflating caldera and the Sierra Nevada range‐front faults under the generally WNW–ESE tension expected from the regional tectonics.

We report our investigation of the Lg attenuation characteristics in Eurasia using digital recordings from the Incorporated Research Institutions for Seismology/Chinese Digital Seismic Network broadband stations. To maximize the utilization of the current station and source distribution in the region, we have derived a new technique known as the source pair/receiver pair (SPRP) method for Lg attenuation measurement. The procedure eliminates seismic source and site effects, leaving behind decontaminated amplitude data from which regional wave attenuation is estimated. For Lg attenuation studies the SPRP method is more flexible and of greater general applicability than its predecessor, the reversed two‐station method (Chun et al., 1987). The application of the SPRP method to our time domain analysis of 33 regional and far‐regional seismic events has provided stable regional average for Lg attenuation in periods of 1–6 s. Our time domain analysis, aided by a damped least squares inversion algorithm, has led to robust measurements of Lg attenuation (Q) and its frequency dependence (η) in nine subregions of tectonic significance. Regions of active tectonics and seismicity generally have lower Q0 than that in the stable continental interior. Subregional Q and η are found compatible with the published results in areas of relatively uniform geology; η values are comparatively poorly determined, suggesting that they are susceptible to the presence of crustal heterogeneities along the wave paths.