Rapid Regional Centroid Solutions

The 2008 Wells Nevada Earthquake was recorded by 164 broadband USArray stations within a distance of 550km (5 degrees) with all azimuths uniformly sampled. To establish the source parameters, we applied the Cut and Paste (CAP) code to all the stations to obtain a mechanism (strike/dip/rake=35/41/-85) at a depth of 9km and Mw=5.9. Surface wave shifts range from -8s to 8s which are in good agreement with ambient seismic noise (ASN) predictions. Here we use this data set to test the accuracy of the number of stations needed to obtain adequate solutions (position of the compressional and tension axis) for mechanism. The stations were chosen at random where combinations of Pnl and surface waves were used to establish mechanism and depth. If the event is bracketed by two stations, we obtain an accurate magnitude with good solutions about 80% of the trials. Complete solutions from four stations or Pnl from 10 stations prove reliable in nearly all situations. We also explore the use of this dataset in locating the event using a combination of surface wave travel times and/or the full waveform inversion (CAPloc) that uses the CAP shifts to refine locations. If the mechanism is known (fixed) only a few stations is needed to locate an event to within 5km if date is available at less than 150km. In contrast, surface wave travel times (calibrated to within one second) produce amazing accurate locations with only 6 stations reasonably distributed. It appears this approach is easily automated as suggested by Scrivner and Helmberger (1995) who discussed travel times of Pnl and surface waves and the evolving of source accuracy as the various phases arrive.