Analysis of earth motions and seismic sources by power spectral density

Abstract

abstractGround motion records from six high-explosive cratering events in northeastern Montana, ten contained nuclear explosive events at the Nevada Test Site, and motions of an earth-fill dam during the Gasbuggy underground nuclear explosion in New Mexico were analyzed for power spectral density, peak velocity and velocity spectra. The high-explosive events included four 20-ton single charges at depths of burst which varied between 42 to 57 feet, a 140-ton row charge consisting of three 20-ton and two 40-ton charges at optimum cratering depths of burst, and a 0.5-ton charge at the optimum depth of burst. It was found that at these depths and charge weights an increase in depth of burst resulted in an increase in peak velocities and power-spectral densities as measured at distant points (> 5 km). Power spectral density was found to be approximately proportional to the first power of yield. For this region it was determined that power spectral density varied inversely as radial distance to the 3.55 power. Three analysis techniques—peak velocity, velocity spectra and power spectral density—are compared, and it is shown that power spectral density is the most consistent method when comparing records from different measuring stations. An analysis of power-spectral density measured at one station for the ten events at the Nevada Test Site shows that a significant shift in the frequency of the energy in the seismogram occurs when the source location changes. For events in the Yucca Flat area the peak energy at Mercury was consistently at 1.0 Hz, while for events in the Pahute Mesa area this peak occurs at 2.5 Hz. A comparison of the power spectral densities on and near the Navajo Dam revealed that the natural frequencies and first harmonics of the dam are 1.4, 2.0 and 2.5 Hz in the mode where motion is parallel to the canyon axis. A simple model makes use of these frequencies to calculate a shear-wave velocity of 1130 ft/sec. A method of using power spectral density to measure earthquake magnitudes and measure the yield of underground explosions is proposed.