Figure 2 - By analyzing the video image taken during the experiment, we concluded that the material cast, especially the impact of the cast material on the pit, took much longer time than the duration of the ground motion data such that we could not resolve its effects in the inversion. On the other hand, the spallation of the source region strata seems to be able to explain, at least partially, the observed source complexity. The source representation of the vertical strata spallation would have a larger M33 component and a peaked spectrum. The strata spallation can be significant. The figures below and to the left give such an example. On the left is the ground acceleration from a multi-shot mining blast recorded by an accelerometer 25-m behind the blast (see figure below). Well-defined vertical strata spall caused by the closest shot(s) in each of the five rows is observed, even though the first row was 69-m away from the receiver.
Figure 3 - Although a vertical strata spall source would produce a spectral peak as what was observed, synthetic calculation shows that to account for the amplitude of the peak, the required spalled mass is too large to be realistic. We then investigated the effect of the low-velocity surface layer on this peak by putting the source in the layer and conducted a trial inversion. The result (above figure) shows that the peak is largely reduced. It may indicate that the explosion was not ideally a point source in the lower high-velocity half-space. Rather, it channeled certain amount of energy into the surface layer because of the shot-hole configuration. This energy was then trapped in the layer and was not modeled by the Green's functions.
Figure 4 - Even though off-diagonal moment-tensor components are much smaller than the diagonal components, due to their high efficiency, they are important in seismic-wave generation. In the figure below, observed ground velocities recorded at one of the receivers are compared with synthetic seismograms calculated from different components of the source moment-tensor shown before. The off-diagonal moment-tensor components, whose amplitudes are less than 20 % of those of the diagonal components, generate ground motions with comparable amplitudes as ground motions generated by the diagonal components. It is the constructive and destructive interference of these two components of ground motions that results in the final synthetic seismograms. The characteristics of the off-diagonal moment-tensor components also indicate that they were generated by sources other than the primary explosion or the spall that might be responsible for the diagonal components.