Local Explosion Monitoring using Rg
Abstract
Rg is the high-frequency fundamental-mode Rayleigh wave, which is only excited by near-surface events. As such, an Rg detection indicates that a seismic source is shallow, generally less than a few km depending on the velocity structure, and so likely man-made. Conversely, the absence of Rg can indicate that the source is deeper and so likely naturally occurring. We have developed a new automated method of detecting Rg arrivals from various explosion sources at local distances, and a process for estimating the likelihood that a source is not shallow when no Rg is detected. Our Rg detection method scans the spectrogram of a seismic signal for a characteristic frequency peak. We test this on the Bighorn Arch Seismic Experiment data, which includes earthquakes, active source explosions in boreholes, and mining explosions recorded on a dense network that spans the Bighorn Mountains and Powder River Basin. The Rg passbands used were 0.4-0.8 Hz for mining blasts and 0.8-1.2 Hz for borehole shots. We successfully detect Rg across the full network for most mining blasts. The lower-yield shots are detectable out to 50 km. We achieve <1% false-positive rate for the small-magnitude earthquakes in the region. Rg detections on known non-shallow earthquake seismograms indicates they are largely due to windowing leakage at very close distances or occasionally to cultural noise. We compare our results to existing methods that use cross-correlation to detect retrograde motion of the surface waves. Our method shows more complete detection across the network, especially in the Powder River Basin where Rg exhibits prograde motion that does not trigger the existing detector. We also estimate the likelihood that Rg would have been detected from a surface source, based on the measured P amplitude. For example, an event with a large P wave and no detectable Rg would have a high probability of being a deeper event, whereas we cannot confidently determine whether an event with a small P wave and no Rg detection is shallow or not. These results allow us to detect Rg arrivals, which indicate a shallow source, and to use the absence of Rg to estimate the likelihood that a source in a calibrated region is not shallow enough to be man-made.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFM.S53D..08O
- Keywords:
-
- 7219 Seismic monitoring and test-ban treaty verification;
- SEISMOLOGYDE: 7290 Computational seismology;
- SEISMOLOGYDE: 7299 General or miscellaneous;
- SEISMOLOGY