We used a dense seismic network to test the P/S amplitude ratio discriminant for earthquakes and explosions at distances from 20 to 200 km in a complex tectonic setting. We used data from several hundred seismic stations in northern Wyoming to observe 22 earthquakes, 37 mining blasts, and 24 borehole shots and compared their P/S amplitude ratios. Frequencies of 6–8 Hz and above achieved the best discrimination results for single‐shot borehole explosions, whereas 6–8 and 8–10 Hz passbands were best for mining blasts. The P/S method appears valid for 50–200 km but fails at distances <50 km. Geologic heterogeneities and path effects have a large effect on amplitude measurements, even after magnitude and distance amplitude corrections are applied. Geologic boundaries between source and receiver that cause sharp lateral discontinuities in seismic velocity can alter the measured P/S ratio, to the point that an event could be misidentified. The source geology is a factor as well, with noticeable amplitude differences between shots in bedrock and shots in sediment. We tested the discrimination performance of the average log(P/S) values of a network of seismometers, which should reduce the discrimination errors caused by these heterogeneities relative to a single station, by bootstrapping the single‐station measurements over varying sample sizes. We found that averaging three or four stations achieved significantly better discrimination than just one or two stations. Finally, we corrected the signal amplitude to remove prephase noise to improve P/S discrimination results. This shows promise for improving event discrimination with only simple corrections.
Online Material: Table of event information and figure showing waveforms overlaid with amplitude windows.