Relationship between Near-Field and Teleseismic Observations of Seismic Source Parameters

Abstract

StCUMITY CLASSIP)CATION Or THIS PA;";-,x .ASSFI the Grafenberg array (6 to 13 stations) were averaged for each event and compare( to the single station (Al) results. Additionally, transverse component data were available for three of the GRF stations. At MAIO only vertical component Lg data were available. For the propagation path from Semipalatinsk to GRF, the narrow-band filtering method and the noise-corrected power spectra of both vertical and transverse components give an average Q for Lg of approximately 1000 For the path from Azgir to GRF the average Q is around 1100. For the path from Semipalatinsk to MAIO the average Q is approximately 625. There appear to be no significant differences in the results among individual events for each path, indicating that the strong tectonic release effects in evidence at long periods for some of the Semipalatinsk events do not influence the Lg spectra of those events appreciably in the frequency band .3 to 1.2 Hz, and the best band to use to establish an m.vs mb relationship is centered at .6 to .7 Hz. To overcome og-period non-isotropic source effects for Shagan River event several very low tectonic release events are used to obtain an empirical "pure"l explosion" Rayleigh signature (shape) which then is incrementally scaled and sub tracted from the total Rayleigh wave signature of high tectonic release events until the best double-couple solution that also fits the Love wave signals is found (i.e. when the trace of the moment tensor is closest to zero); the scaled signal is then the appropriate explosion contribution. The remaining "pure earthquake"'Rayleigh and Love wave signals for each event are interpreted in .terms of a best-fitting equivalent earthquake model.To determine the approximate threshold for obtaining source parameters from long-period surface waves in an intraplate setting, the phase matched filtering method is used to analyze the 1982 New Brunswick aftershock sequence with the l*6 4166c0rded mainshock seismograms as the reference signals. The source excitation function of each aftershock is extracted from the observed signal spectrum by correcting for the path transfer function and the main shock source excitatation. Preliminary results using this approach on recordings at GDSN stations in North America indicate that source parameters (including depth) can be extracted for events as small as about mb 4 to 4.5. (This will be part of a M.S. thesis b B. Yan).The 1982 New Brunswick earthquake series, including the mb 5.7 mainshock and several high magnitude aftershocks were investigated to determine their source mechanisms and focal depths. Band-pass filters applied over a broad frequency range provided group velocity and spectral amplitude data for Rayleigh and Love modes for the path to SCP. Synthetic amplitude spectra were generated assuming structure parameters common to both the source area and the propagation path. Key aspects of the source mechanism could be inferred from the fundamental mode observations, particularly for the mainshock on January...

Relative Lg and P-Coda Magnitude Analysis of the Largest Shagan River Explosions

“…There have been a number of studies which have indicated that measurements made in the P-coda are at least as stable as network-averaged m b estimates and significantly better than single-station m b (Baumgardt, 1983;Bullitt and Cormier, 1984;Baumgardt, 1984;Gupta et al, 1984). Similar results have been obtained for Lg and Lg-coda wave measurements (Ringdal, 1983;Alexander, 1984).…”

“…In Table 3, the source parameters of the explosions in Table 2, including the tectonic components by Sykes and Cifuentes (1983) and Given et al (1983), are listed. As pointed out by Alexander (1984) Another important observation which can be drawn from these plots is that the slopes of the best fitting lines for the Lg measurements are quite small, ranging from 0.6 to 0.8. Interestingly, however, the P-coda measurements have higher slopes of 0.9 to 1.0.…”

“…The Lg phase has proven to be useful for yield estimation because it combines the advantages of surface waves along with the fact that Lg magnitudes appear to be insensitive to tectonic component (Alexander, 1984). Nuttli (1984) has developed an absolute Lg magnitude method in which Lg magnitude attenuation corrections are determined from Lg-coda Q estimates.…”