Additional studies have been conducted to help verify the findings of previous investigations (cf. Murphy and Bennett, 1982) which indicated that spectral differences in short-period regional phases may be diagnostic of source type for Nevada Test Site (NTS) explosions and nearby earthquakes recorded at Tonto Forest Observatory in Arizona. Regional phase signals from two additional stations, the Uinta Basin Observatory array in Utah and the Blue Mountains Observatory array in Oregon, and from supplemental events, including aftershocks of the 1966 Caliente, Nevada, earthquake, have been analyzed. These stations are at epicentral distances of 430 to 530 km for Tonto Forest Observatory, 520 to 670 km for Uinta Basin Observatory, and 870 to 880 km for Blue Mountains Observatory, with the near ranges representing the distances to the Caliente source area and the far ranges to the average NTS source/receiver distances. This enhanced data base included 50 earthquakes within about 150 km of NTS and 35 NTS explosions recorded at one or more stations. The events cover a magnitude range from about 2.8 to 5.2 (mb), with the majority of earthquakes concentrated in the lower half of that range.
The current investigation essentially corroborated our previous findings that comparison of simple, peak-amplitude measurements of regional P and Lg phases did not consistently discriminate between earthquakes and explosions. This discriminant breakdown appears to be related to variability in excitation of regional phases from similar sources which has been dramatically illustrated by observations from the 1966 Caliente aftershocks; these events produced large variations in the relative amplitudes of Pg and Lg signals even though the events were spatially restricted to a small zone. In contrast, the Lg spectral ratio discriminant measure, defined in the previous study, continued to provide reliable distinction between the earthquake and explosion sources for data observed at all three stations. The discrimination capability of the spectral ratio prevails in spite of the evident mechanism variability between earthquakes and shifts in the discrimination threshold between stations apparently related to attenuation differences along the propagation paths.