Source processes and proximal cause of the 2006 seismic swarm activity at Mt. Martin, Katmai Volcanic Cluster, Alaska

In January 2006, a noneruptive swarm of over 1,000 volcano-tectonic (VT) earthquakes was recorded by the Alaska Volcano Observatory’s (AVO) seismic network. These VT earthquakes were located at approximately 3-6 km depth beneath the summit of Mt. Martin, a volcano located within the Katmai Volcanic Cluster. In an effort to determine the cause of the swarm, we calculated and analyzed double-couple fault plane solutions (FPS) for a pre-swarm period (Jan-Dec 2005), the swarm period (January 2006) and a post-swarm period (Feb - Dec 2006) using FPFIT. Only well-constrained FPS (those with locations which had an RMS < 0.2, location errors <= 1.5 km, azimuthal gap < 180°, and which had at least six clear first motion polarities, STDR > 0.4, and a unique solution or multiple similar solutions) were used for final analysis. A total of 56 well-constrained FPS were maintained for pre-swarm, 77 for swarm, and 61 for post-swarm. For all three periods, a mix of strike-slip, normal and reverse faulting was present, with strike-slip and normal faulting dominating. Reverse faulting increased from ~9% pre-swarm to ~25% during the swarm and decreased back to ~7% post-swarm for all well-located events. An examination of Rose diagrams and stereonets of FPS P- and T-axis orientations indicates that, for the pre-swarm period, P-axes (a loose proxy for maximum compression) for strike-slip and reverse-fault events were oriented NNW, roughly parallel to regional stress along this segment of the Aleutian arc. During the swarm, these P-axes rotated by approximately 60° to a NNE orientation. Post-swarm results show that the strike-slip and reverse FPS had not returned to their pre-swarm orientation as of the end of 2006. T-axes of normal-faulting earthquakes show a mix of NNW and NNE orientations during the pre-swarm period. During the swarm, T-axis orientations became homogenous and oriented N30°E. Post-swarm, normal faulting T-axes rotated even further to ~N70°E. The ~60° change in strike-slip/reverse fault P-axis azimuth and rotation of normal fault T-axis may be indicative of a shallow (~3-6 km below sea level) dike intrusion in 2006. Ongoing analyses of both FPS and precisely relocated hypocenters will help us ascertain whether or not there exists any spatial clustering of different FPS types that may provide further evidence of a mid-crustal magmatic intrusion.