A Seismic Location Study of Station Anomalies, Network Effects, and Regional Bias at the Nevada Test Site

A location study is made of 28 large underground explosions detonated in the northern area of the Nevada Test Site NTS. Recording networks were comprised of between 9 and 49 teleseismic stations having two- to three-quadrant distributions. Errors of locations obtained without applying travel-time anomalies relative residuals, and with depths restrained to the known values, average about 7 km but are as large as 20 km. With anomalies, the errors are consistently 2.5-3.0 km. The size of the area at NTS across which the anomalies are valid is at least 70 km by 25 km. If a constant network is used to locate a set of events, the bias is a consistent and unique function of that network. The lack of a common bias among networks with similar azimuthal distributions definitely eliminates the source region as the principal cause of the anomalies. Alternatively, the anomalies may be attributed to slight lateral and vertical inhomogenities within the mantle between the source and receiver, the effects of which are integrated along the entire path. Furthermore, when constant networks are used, the relative locations obtained without anomalies are identical to those obtained with anomalies, except for a bias translation appropriate for that network.