Linear Elastic Waveform Modeling in the Inelastic Region of Underground Nuclear Explosions

Abstract

In an attempt to characterize the way in which linear elastic theory breaks down within the inelastic region near underground nuclear explosions, we have compared the observed velocity waveforms from the inelastic region with predictions of ordinary elastic generalized ray theory. In earlier studies, models for events at both Pahute Mesa and Amchitka were developed using ground motion recordings from the 5 to 20 km distance range. The models consisted of a source time junction and a source strength, ψ∞ for a point source in a layered elastic crust. In this study the predictions of the models were evaluated at ranges from 0 to 5 km. Surprisingly, the elastic models predict the observed peak vertical velocity very accurately even in the spall zone. They also predict the rise times of the velocity pulses into surface zero. The elastic synthetic seismograms match the observed velocities up until the time when the spalled material begins to decelerate. The observed radial motions are much more poorly modeled. Nevertheless, the results suggest that only low-order corrections to elastic theory may be required to explain some of the phenomena within the inelastic region of contained nuclear explosions. Vertical velocity waveforms within the spall zones of Amchitka and Pahute Mesa events are well modeled by linear elastic theory during the compressive pulse prior to spall. We introduce the definition of the "compressive elastic radius."

Additional details