Initial Results on Rugged Low Power Compact Silicon MEMS Sensors for Use in Nuclear Explosion Monitoring; Conference paper

摘要

The Frozen Rock Experiment (FRE) was conducted in central Alaska in August 2006 to provide empirical data on seismically-estimated yield from explosions in frozen rock. Laboratory studies have demonstrated that frozen rock is significantly stronger than unfrozen rock, and it has been hypothesized that this increased strength, due to ice in the pores and cracks, can alter seismic yield. Central Alaska has abrupt lateral boundaries in discontinuous permafrost, and we detonated 3 shots in frozen, saturated rock and 3 shots nearby in unfrozen, dry rock ranging in size from 200 to 350 lbs. Approximately 125 accelerometers and seismometers were deployed specifically for this experiment at distances of 10 m to over 20 km. During the past year, we have conducted various studies (moment tensors, magnitudes, etc.) to characterize the explosions in frozen and unfrozen rock. In order to conduct moment tensor inversions, a velocity and an attenuation model are required. We initially developed a 1-D model for central Alaska but found that the model could not account for azimuthally dependent velocities and did not explain observed surface-wave arrival times. We are currently developing a 3-D 'block' velocity model using full waveform modeling in order to improve the preliminary moment tensor results. To develop an attenuation model for the upper crust of central Alaska, spectral amplitudes of Rg were extracted from complex waveforms using phase match filtering. The Rg amplitudes were then corrected for geometric spreading and regressed against distance to determine the attenuation coefficients at each period, which were then converted to QRg using the group velocity dispersion curves. We developed the relationship QRg(f) = 12.5f - 0.39. When the attenuation coefficient data were inverted. Q values of less than 20 were required in the upper kilometer of the central Alaska crust.