Local seismic phenomena in the first three seconds after underground nuclear explosions

Accelerograms obtained at distances under 100 km from four underground nuclear explosions in Nevada (Benham, Greeley, Knickerbocker, and Faultless) were studied to determine the possibility of identifying secondary seismic events immediately following the explosion. The purpose of the investigation was (1) to determine the minimum amount and quality of data needed to perform such an analysis and (2) to try to identify any such secondary events, such as those associated with ground faulting, if possible. The records were digitized at 40 points/sec for 3 sec and then analyzed by various filtering and array-processing techniques to try to pick out significant arrivals from events subsequent to the explosions. The ten usable accelerograms available from each of the Nevada Test Site events Greeley and Benham proved inadequate for any definite conclusions, although there is a strong suggestion that two events were identifiable in the case of Benham. The first was a possible normal faulting event less than 0.2 sec after the explosion and near to it, and the second is associated with long-period arrivals appearing to originate a few kilometers to the southwest of Benham, which are likely to be associated with one of the major aftershock zones observed for this explosion. Of the 24 records available for the Faultless explosion, more than half were considered to be of good quality for this analysis. A technique using all the accelerograms to search a 4- by 4-km area around the explosion for events with particular assumed focal mechanisms was developed, and in at least one case it appears to have demonstrated an earthquake-type event that occurred about 1 km from the explosion and about 1.5 sec after it. Also, the surface reflection of the compressional wave was identified with some certainty, although the possibility exists (but is considered less likely) that a normal-faulting type of event very near the shot point could be mistaken for this surface reflection. The conclusion was that a minimum of perhaps fifteen such accelerograms of ideal quality, with station distributions chosen carefully within a 100-km radius, could be used to identify with certainty secondary seismic events with amplitudes as small as one-quarter that of the largest event and that such techniques might also be applicable to a natural earthquake as well as to explosions, although the earthquake would present certain additional problems.