HR: 1330h
AN: S52D-0158    [PDF]
TI: Seismic background noise and signal detection near Yucca Mountain, Nevada
AU: * VonSeggern, D H
EM: vonseg@seismo.unr.edu
AF: Nevada Seismological Laboratory, U. Nevada, MS 174 1664 N. Virginia St., Reno, NV 89557 United States
AB: The seismic monitoring network for Yucca Mountain, Nevada, a potential high-level nuclear waste repository, consists of thirty 24-bit, 3-component digital stations located within a ring of approximately 50 km radius around Yucca Mountain. This network has operated since October 1, 1995; and a catalog of nearly 20,000 earthquakes has been compiled thus far. The network has an extremely low threshold, with apparently complete detection to ML = 0.5 for earthquakes within the network. Within the denser part of the network at less than 25 km from Yucca Mountain, the complete detection threshold is ML = -0.2. The major reason for this low threshold is the low seismic background noise in the region. To characterize this noise, a sampling of all channels was made at hourly intervals for a period of one year. This provided data to quantify the background noise behavior and its statistics. Most stations have average rms noise levels of 10 nm/s or less. Day-night differences in average noise levels are present; some stations show up to a factor of four while most stations show much less than a factor of two. The distribution of noise samples at a given station is roughly log-normal, but usually with a tail on the upper side. The correlation matrix of noise rms shows significant values for almost all station pairs in the network, largely due to the day-night variations. The correlations are weakly dependent on inter-station distances. Noise samples taken regularly over one month were used to compute power spectra. These spectra confirm that noise levels are among the lowest in North America. The magnitude bias at each station (station correction) was computed from the large database of station ML's. A model of signal detection was coded by using the noise statistics derived in this study for mean, standard deviation, and inter-station correlation, by incorporating the station magnitude corrections, and by simulating the STA/LTA detection algorithm. A synthetic catalog of earthquakes was created from this model to compare predicted detection thresholds with those of the actual catalog. The thresholds derived from the synthetic catalog agree well with the actual ones seen for the network. A slight upward bias in estimated magnitude exists for small magnitudes where only a few stations detect; this is due to the fact that stations with positive magnitude bias added to a positive random component detect events preferentially.
DE: 7230 Seismicity and seismotectonics
DE: 7294 Instruments and techniques
DE: 7299 General or miscellaneous
SC: Seismology [S]
MN: 2003 Fall Meeting