Seismic beamforming to resolve a Mt St Helens event
I wrote a Matlab program that does a coherent "time delay of arrival" analysis on the seismic channels (currently only considering the Z direction). In examining the waveforms from the Mt. St. Helens event, the analysis gives a bearing to the source of 261.9 degrees. Computing a bearing using the published epicenter of the earthquake and the LHO location using the WGS84 earth model gives an actual bearing to the published epicenter of 263.9 degrees, a difference of 2 degrees. At the distance of Mt. St. Helens this corresponds to a distance of 7.5 km.
The figure at left shows the waveforms from the six seismometers at this time, after filtering with an elliptical bandpass filter with a pass band of 1-5 Hz, and decimation by a factor of ten. I estimated (using Google Maps) the location of the seismic vault in "LIGO coordinates" to be X=1040m, y=186m. The figure at right shows the "power" (not well-defined) associated with various wavevectors; the X axis gives "east slowness" and the Y axis gives "north slowness," in seconds per kilometer. The circle indicates a velocity of 5000 m/s, the approximate velocity of P-waves in rock. The seismometer locations are superimposed for directional reference only.
I am interested in whether this could be used to image local seismic noise.
Seismic beamforming to resolve a Mt St Helens event
I wrote a Matlab program that does a coherent "time delay of arrival" analysis on the seismic channels (currently only considering the Z direction). In examining the waveforms from the Mt. St. Helens event, the analysis gives a bearing to the source of 261.9 degrees. Computing a bearing using the published epicenter of the earthquake and the LHO location using the WGS84 earth model gives an actual bearing to the published epicenter of 263.9 degrees, a difference of 2 degrees. At the distance of Mt. St. Helens this corresponds to a distance of 7.5 km.
The figure at left shows the waveforms from the six seismometers at this time, after filtering with an elliptical bandpass filter with a pass band of 1-5 Hz, and decimation by a factor of ten. I estimated (using Google Maps) the location of the seismic vault in "LIGO coordinates" to be X=1040m, y=186m. The figure at right shows the "power" (not well-defined) associated with various wavevectors; the X axis gives "east slowness" and the Y axis gives "north slowness," in seconds per kilometer. The circle indicates a velocity of 5000 m/s, the approximate velocity of P-waves in rock. The seismometer locations are superimposed for directional reference only.
I am interested in whether this could be used to image local seismic noise.