Anomalous surface wave observations from underground explosions have been widely attributed to the phenomenon of spall, i.e., the detachment and subsequent slapdown of near-surface layers in response to explosion-induced tensile stresses. We argue, on the basis of a simple equivalent-force model, that, while spall can enhance explosion surface waves at very short periods, it cannot contribute significantly to the teleseismic surface waves radiated by underground explosions at periods exceeding about 10 sec. Previous theoretical results to the contrary are in error, having been based on source models which do not conserve momentum.
A nonlinear, two-dimensional (axisymmetric) finite difference simulation of a buried explosion in granite further supports the conclusion that spall cannot contribute to long-period surface waves. The simulation exhibits extensive spall; nonetheless, predicted fundamental mode Rayleigh wave spectra are nearly identical to those obtained from a spherically symmetric simulation (which does not include spall), at periods exceeding about 10 sec. At shorter periods, the two-dimensional simulation predicts some Rayleigh wave enhancement, compared to the one-dimensional simulation. The maximum enhancement, about a factor of 2, occurs at a period of approximately 2.5 sec. Synthetic long-period Rayleigh wave seismograms, at 3000 km range, show no perceptible phase or amplitude anomalies due to spall. These results are in excellent agreement with the predictions of our equivalent-force model.