Explosion Source Phenomena Using Soviet, Test-Era, Waveform Data. Final Technical Report

摘要

During the nuclear testing era, the former Soviet Union carried out extensive observations of underground nuclear explosions, recording both their own shots and those of foreign nuclear states. Between 1961 and 1989, the Soviet Complex Seismological Expedition deployed seismometers at time-varying subsets of over 150 sites to record explosions at regional distances from the Semipalatinsk and Lop Nor test sites and from the shot points of peaceful nuclear explosions. This data set included recordings from broadband, multi-channel ChISS seismometers that produced a series of narrow band outputs, which could then be measured to perform spectral studies. (ChISS is the Russian abbreviation for multichannel spectral seismometer. In this instrument the signal from the seismometer is passed through a system of narrow bandpass filters and recorded on photo paper. ChISS instruments have from 8 to 16 channels in the frequency range from 100 sec to 40 Hz. We used data mostly from 7 channels, ranging from 0.08 to 5 Hz.) Quantitative, pre-digital era investigations of high-frequency source scaling relied on this type of data. To augment data sets of central Central Asia explosions, we have measured and compiled 537 ChISS coda envelopes for 124 events recorded at Talgar, Kazakhstan, at a distance of about 750 km from Semipalatinsk. Envelopes and calibration levels were measured manually from photo paper records for seven bands between 0.08 and 5 Hz. We obtained from 2 to 10 coda envelope measurements per event, depending on the event size and instrument magnification. Coda lengths varied from 250 to 1400 s. For small events, only bands between 0.6 and 2.5 Hz could be measured. Envelope levels were interpolated or extrapolated to 500 s and we have obtained the dependence of this quantity on magnitude. Coda Q was estimated and found to increase from 232 at 0.08 Hz to 1270 at 5 Hz. These relationships were used to construct an average scaling law of coda spectra for Semipalatinsk explosions. Significant differences from average scaling were observed and may result from variations in emplacement conditions. The ChISS envelope data have been integrated into coda processing at Los Alamos National Laboratory (LANL) by applying ChISS filter bands to modern, digital data from central Central Asia, for purposes of magnitude and yield calibration. In addition, we have compiled regional, direct phase measurements for ChISS recordings at Talgar, Garm, Zerenda, and Novosibirsk. The ChISS envelope data have been integrated into coda processing at Los Alamos National Laboratory (LANL) by applying ChISS filter bands to modern, digital data from central and east Asia, for purposes of yield calibration. The difference in manual versus digital measurement methods are captured in site terms that are higher by up to 0.5 log10 units for ChISS data, relative to modern Talgar data due to the measurement of peak, rather than mean envelopes. After correction for site and path effects, ChISS amplitudes compare well to measurements from the Borovoye archive for events in common. Direct wave measurements have been used to construct spectra for Semipalatinsk explosions, and can be used to explore the behavior of regional phase amplitudes with shot point and emplacement condition.