Discrimination of large chemical explosions (e.g., mining blasts) from possible clandestine nuclear tests (e.g., decoupled explosions) is a significant issue for seismic verification. Unless discrimination is possible, the numerous mining blasts within the USSR would give ample opportunity for concealing a clandestine test. Evernden et al. (1986) advocate high frequencies as a means to detect these clandestine tests; however, mining blasts must still give a unique signature for their discrimination. Specific conditions are also necessary for high-frequency verification: low attenuation and quiet seismic stations are among the most important. In this study we address these problems under conditions representing a best case scenerio.
During the summer of 1985, Lawrence Livermore National Laboratory (LLNL) deployed a regional array and high frequency seismic station near RSON (Red Lake, Ontario, Canada) to study regional seismic signals at a low-noise site with excellent propagation characteristics, the continental shield. One objective was to evaluate the performance of regional seismic arrays and high-frequency stations in this geological environment. Using high-frequency data from mining blasts within the Mesabi Iron Range in northern Minnesota (distance of 380 km), we also observed the high-frequency Pn phase of ripple-fired blasts, modeled their source properties, and evaluated a possible discriminant for chemical explosions.
This study suggests the existence of a distinctive signature for large chemical explosions: strong, high-frequency spectral peaks in the P spectra introduced by delay shooting or ripple firing. Its application depends, however, on a frequency band extending to at least 35 Hz at regional distances, reasonably uniform delay times in the blast pattern, predictably quiet seismic sites, and minimal path attenuation for high frequencies. At 380 km, we observe good propagation to 30 Hz; at 50 Hz the evidence suggests severe attenuation or scattering which is consistent with a total Q between 1500 and 2500. Attenuation estimates assuming a simple source behavior (e.g., f−2 or f−3) at high frequencies for mining explosions would give incorrect values of Q unless corrected for the blast pattern. These attenuations will limit the value of high frequencies for the detection of small decoupled explosions at regional distances.