Assessing Hypocentral Accuracy and Lower Magnitude Completeness in the Pacific Northwest Using Seismic Refraction Detonations and Cumulative Frequency-magnitude Relationships

Abstract

Absolute hypocentral location errors are traditionally estimated from the errors in the locations of quarry blasts. Catalog completeness has been evaluated by examining cumulative frequency-magnitude relationships ( e.g., Rydelek and Sacks, 1989). In the Pacific Northwest, however, the number of quarries with well defined blasting schedules is relatively sparse ( e.g., Benson et al., 1992). Seismic refraction detonations provide an independent assessment of actual location errors of surface events in the Pacific Northwest and elsewhere. Because the detonation yields are also known, their reported magnitudes can be used to investigate the detection and location thresholds for low-magnitude earthquakes. As the number and distribution of seismic stations in the Pacific Northwest has expanded (Figure 1), the location accuracy of the networks and the completeness of their catalogs has improved with time. Because the station coverage is not uniform geographically, these network properties vary with location. In this note we use 72 refraction detonations listed in the Advanced National Seismic System (ANSS) and Pacific Northwest Seismic Network (PNSN) catalogs for the Pacific Northwest to investigate the hypocentral accuracy and completeness of the earthquake catalog as a function of time and location since 1984. These reported detonations had an average charge size of 939 kg and yielded an average coda magnitude of 1.65. Applying magnitude-versus-charge-size relationships to 64 detonations not listed in the ANSS and PNSN catalogs permits us to extend our analysis back to 1965. Brocher (2003b) noted several reasons why the location errors for surface detonations may not be fully representative of location errors for tectonic earthquakes. Nonetheless, quarry blasts and refraction detonations represent seismic sources whose origin times and locations are precisely known, and they provide an independent measure of the quality of location solutions by the network. Furthermore, the regional variation in the ability of a seismic network to locate surficial …