Abstract Spectral and cepstral analysis were performed on regional-array NORESS recordings of mining explosions in Scandinavia and Russia in order to detect the effects of delayed explosions or “ripple firing” commonly used in mine-blasting practice. NORESS array-stacked spectra, corrected for instrument and noise, were computed in windows on Pn , Pg , Sn , and Lg waves from the mine blasts by averaging individual channel spectra, frequency by frequency, across the array. For comparison, spectra were also computed for six western Norway earthquakes located near the Blasjo and Titania Norwegian mines. Also, stack spectra of P coda, Sn , and Lg were computed using NORSAR seismograms for presumed peaceful nuclear explosions (PNEs) located at near-regional distances (Δ > 10°) in western Russia. The NORSAR spectra were determined in the same manner as the NORESS spectra, except that an individual source correction was applied to the spectra. Comparison of western Norway mine-blast spectra and nearby earthquakes shows that the earthquake Pn spectra are peaked in the 8- to 16-Hz band, with little energy below 6 Hz, whereas the explosion spectra have flatter spectra from 3 to 16 Hz. This difference in the Pn spectra may be caused by the earthquakes occurring at greater depths than the explosions, differences in the near-source media, or source mechanism effects in the earthquake spectra. The main difference between earthquake and explosion spectra is that the explosion spectra exhibit marked scalloping or modulation patterns not observed in the earthquake spectra. The modulation patterns are identical in spectra of all phases, indicating that they are caused by multiple-shot sequences or ripple firing. Simulated multiple-source spectra, determined using the NORESS seismograms for one of the western Norway earthquakes, resulted in spectra that resembled many of the explosion modulation patterns. Based on the simulation studies and cepstral analysis of the spectra, the mine explosions appear to be composed of two to three separate explosions delayed by between 80 to 150 msec. Moreover, the NORSAR spectra of PNEs indicate that these signals are also produced by multiple explosions, although they have larger delay times of between 0.9 to 1.5 sec. These results show that spectral modulations in seismogram spectra of economic explosions, including mine explosions and nuclear explosions fired for peaceful, economic purposes, can be used to distinguish them from earthquakes and nuclear-explosion weapons tests.
[1]
T. C. Bache,et al.
Spectral characteristics of regional phases recorded at Noress
,
1988
.
[2]
D. Dickson.
Soviet union suspends plans to divert four rivers.
,
1986,
Science.
[3]
D. Baumgardt.
Comparative analysis of teleseismic P coda and Lg waves from underground nuclear explosions in Eurasia
,
1985
.
[4]
T. C. Bache,et al.
Q for teleseismic P waves from central Asia
,
1985
.
[5]
H. Bungum,et al.
Processing of regional seismic events using data from small-aperture arrays
,
1984
.
[6]
E. S. Husebye,et al.
Seismic array configuration optimization
,
1983
.
[7]
R. C Kemerait,et al.
A multidimensional approach to seismic event depth estimation
,
1982
.
[8]
T. J. Bennett,et al.
A discrimination analysis of short-period regional seismic data recorded at Tonto Forest Observatory
,
1982
.
[9]
M. Nordyke.
A review of Soviet data on the peaceful uses of nuclear explosions
,
1975
.
[10]
E. A. Flinn,et al.
Detection and analysis of multiple seismic events
,
1973,
Bulletin of the Seismological Society of America.
[11]
David von Seggern,et al.
Source Time Functions and Spectra for Underground Nuclear Explosions
,
1972
.
[12]
U. Langefors,et al.
The modern technique of rock blasting.
,
1968
.