Double-difference relocations and spectral ratio analysis of volcanic seismic events in the Mount St. Helens crater using a 3D velocity model suggest slip events under the new dome with constant stress-drop scaling.

Abstract:

Shallow low frequency seismic events are common features associated with restless and erupting volcanoes. The physical mechanisms generating their characteristic low frequency, and often extended duration signals remain poorly understood. Here we present new double-difference relocations and spectral scaling of a group of ~400 shallow low-frequency seismic events occurring within the Mount St. Helens edifice during its 2004-2008 dome-building eruption, as recorded by a temporary seismic array for a month within the crater in 2006. Relocation results suggest that the majority of earthquakes occurred in the center of the crater close to the vent at depths < 500 m, with some events potentially locating under the new dome but ~200-300m southwest of the vent. Low-frequency events exhibit moment-corner frequency scaling roughly consistent with a constant static stress-drop, similar to tectonic earthquakes occurring elsewhere in shallow crustal faults. The scaling suggests that the ~400 events result from stick-slip behavior, and that the low frequency character of the waveforms may result from a combination of path effects and slow rupture speeds.

For relocation, we divide the 400 events into eight families based on waveform similarity, and use a subset of nearly 40 earthquakes with clear first arrivals ranging in moment magnitude from 0.4 – 1.8 to calculate hypocenters. We then relocate these events using a double-difference method with a three-dimensional velocity model of the edifice from Waite and Moran (2009). The relocated events are then used to estimate source parameters of the remaining earthquakes via a spectral ratio technique. Spectral corner frequency estimations based on this spectral ratio approach produce stress-drop values of ~1 MPa assuming a shear-wave velocity of 1500 m/s. The estimations also indicate a constant stress-drop scaling for all events, with two event families having lower estimated stress drops of ~0.1 MPa. While localized lithological differences could cause corner frequencies and thus stress drop values to be lower for two of the families, the variations may result from an erroneous assumption of a constant velocity value in the upper 500 m. Overestimations of shear-wave velocity of 30% near the two families with lower stress drops would explain the variation.