Large‐aperture seismic array (LASA) subarray amplitude anomalies are investigated using 395 medium sized events, distributed in ten azimuthally divided sectors. Although LASA magnitudes, when averaged over all azimuths, are only slightly biased in relation to National Earthquake Information Services (NEIS) magnitudes, the amount of bias varies with azimuth and subarray, suggesting that a simple station correction for mb bias is not adequate. In addition, fluctuations among LASA subarrays are about 0.15 in standard deviation even when these magnitudes are calibrated in sectors. Details of such fluctuations are explained in part by local crustal and upper mantle heterogeneities under LASA. The amplitude anomalies are linearly related to the amount of travel time anomalies in each sector, implying that both effects are due to crustal focusing. Using a fixed effects model, the authors attempt to separate the cause of mb bias into sector (azimuth) effect, subarray effect, and subarray‐sector interaction. However, even this detailed modeling could not, with confidence, explain mb bias. Using the reciprocity principle, even with a well‐calibrated station, a factor of 2 in uncertainty results for predictions of a stations's amplitude for an event only 50 km away from a calibration event in a region as complex as LASA. The event magnitude uncertainty would probably be reduced by network averaging, or by a nonstatistical detailed crust and mantle structure which could be analyzed by ray tracing to remove receiver effects.

No abstract

Large-aperture seismic array (LASA) the station, a low-velocity channel in the upper subarray amplitude anomalies are investigated mantle, high heat flow, and upper mantle velousing 395 medium sized events, distributed city structure. More specifically, Der [1976] in ten azimuthally divided sectors. Although showed that P wave amplitudes of stations in the LASA magnitudes, when averaged over all azimuths, eastern United States (EUS) are generally larger are only slightly biased in relation to National than those from stations in the western United Earthquake Information Services (NEIS) magnitudes, States (WUS) by a factor of 3. Also, he indi-

To reduce the number of false alarms it is necessary to steer detection beams to local areas. By detecting local events on these beams and by using a higher S/N threshold in processing these signals, we can effectively reduce the number of false alarms from the original 57% to 41%. A new beam set has been developed and deployed uhich concentrates teleseismic beams in high seismicity areas instead of spacing them equidistantly apart. This arrangement reduced the average detection errors from 200 km to 50 km, there is also some indication of a lowered detection threshold on the order of 0.1 + 0.1 magnitude units.

Justific