Interpreting array records: explosion and earthquake P wavetrains which have traversed the deep mantle

Abstract

Before 1946 four explosions with yields of a kiloton or more were recorded by seismographs up to ranges of 10°. After the first motion, the seismograms were not obviously different from those of earthquakes at similar distances. In 1946 Bikini Baker was recorded out to distances of 80° with amplitudes equivalent to an earthquake of magnitude m 51/2. Since then numerous explosions have been recorded at teleseismic ranges along nine different transmission paths with signal/noise ratios large enough to enable details of the P wavetrain to be analysed. Between 30 and 90° these explosion seismograms are characterized by a large, relatively simple onset pulse followed by 20 to 30 s of low amplitude oscillations. Variations of waveform across a large array of seismometers are shown to be due to signal-generated noise. The time-shifted and array-sum records of a given explosion are almost identical. Differences between explosion seismograms are a function of focal depth and local geology. A good working hypothesis is that the deep mantle is very homogeneous. Waveforms of the majority of shallow earthquakes recorded under these experimental conditions are dependent upon azimuth and range. Signals related to focal depth may account for only a small proportion of the observed P energy. The conversion of SV waves qualitatively explains the seismograms. The observations are consistent enough to make it worth while taking stock of the capacity of the international network to study earthquake mechanism.