An examination of the reliability of focal depth estimates for a seven station, 70 km aperture network in Virginia is presented. Eight microearthquakes featuring good P and S wave arrival-time data sets were selected for analysis, and hypocenters were calculated using a conventional iterative least squares algorithm, in conjunction with a three layer crustal velocity model. The earthquakes selected are representative in the sense that their epicenters are located at positions within, at the periphery, and outside the confines of the network. Their focal depths ranged from 6 to 16 km. Origin times for these earthquakes were reestimated from the Wadati diagram, under the assumption of a constant P to S wave velocity ratio. Also, focal depths were reestimated by assuming a constant, but unknown crustal velocity. Both of these methods are essentially independent of the velocity model used in the initial locations.
Comparisons of the origin times and focal depths derived from the reestimate calculations with those determined by the conventional iterative least squares algorithm show good agreement. With both P and S wave arrival-time data at all stations, differences in depths derived by the various methods are less than 2.5 km, for events situated within the network. Also, the statistical error measures calculated by the iterative least squares location program appear to be realistic in the sense that they enclose the reestimated values. Location capability degrades considerably for earthquakes situated outside the network and it is shown (again) that the availability of good S wave arrival-time data is a critical element in obtaining reliable focal depth estimates for less than optimum epicenter-station recording configurations.