Search

Work in the following areas is summarized: Array and network evaluation; Signal detection methods; Signal estimation techniques; and Discrimination.

This final report summarizes the material covered in six previous semiannual technical reports, and discusses the conclusions obtained from these studies. The counterevasion studies were concerned with the evaluation and application of two techniques: complex cepstrum analysis and eigenspectrum analysis. These techniques were used to separate interfering seismic signals in order to distinguish underground nuclear explosions hidden in larger earthquakes, for the identification of sequential underground nuclear explosions, and for the detection of the pP depth phase. In addition, a new study on the effect of sampling rate on the detection of the pP depth phase is also presented. The main topic of the far-field research was directed toward the determination of seismic source parameters using teleseismic surface wave spectra. Techniques were developed to fit theoretical spectra generated using a double couple source in a layered half-space to observed spectra corrected to a near source reference point. Several procedures were used to correct the observed spectra for effective attenuation.

This first quarterly report summarizes progress under the VELA Network and Automatic Processing Research Program. Work in the following areas is summarized: -- array and network evaluation, signal detection methods, signal estimation techniques, and discrimination.

Examination of the seismic source from both the near- and farfields has been undertaken. Using a discrete Fourier transform (DFT) and a maximum entropy spectral estimator on near-field acceleration data produced corner frequency and low frequency level estimates. This latter estimator eliminates leakage, thereby obtaining more accurate values for the high frequency end, with better roll-off values and definition of side lobes. When the maximum entropy spectral estimator was applied to data from the Parkfield earthquake and the Bear Valley event of June 22, 1973, the lower frequency level was lower than that from the DFT in all cases. Using spectral fitting procedures on fundamental mode surface wave data from two central California events, generally close agreement was obtained with source mechanism solutions from bodywaves. For the Bear Valley earthquake of June 22, 1973, the seismic moment obtained from the surface wave data was an order of magnitude smaller than that obtained from acceleration data., See also ADA005786.

The solutions to eight earthquakes have been obtained from near-field strong-motion accelerograms using Haskell's moving dislocation source model. In particular, the solutions to the San Fernando main shock and three small aftershocks were re-examined using the corrected orientation of the Pacoima Dam instruments. A reasonable match to the observed waveforms for the main event could only be obtained by having a large strikeslip dislocation component on the deeper segments of the fault. Also, new locations and orientations were obtained for the three small aftershocks. Using both Haskell's model and spectral analysis, the Bear Valley earthquake of June 22, 1973 was examined. Both approaches yielded substantially the same results; a large dislocation on the order of 100 cm over a small fault area of less than 1.0 sq km, with a seismic moment (Mo) of 1.2 to 1.5 x 10 to the 25th power dyne-cm.

Both exhaustive and non-linear spectral fitting procedures are applied to far-field surface wave data for determining seismic source parameters, especially depth. These procedures were applied to a series of events with decreasing control on the source parameters. High quality Rayleigh wave data from the Southeastern Missouri Earthquake of October 21, 1965 was analyzed, and the parameter corroboration with independent estimates was good, with excellent spectral fits. Two low magnitude (mb = 4.1) Italian events were analyzed individually and by taking spectral ratios. Six events from Sinkiang were analyzed, again using both techniques. Initial efforts are directed towards determining the importance of the parameter variations; it was found that depth has the greatest effect on the shape of the spectra., See also AD0778956. Sponsored in part by DARPA.