Seismic Swarm Associated With the 2008 Eruption of Kasatochi Volcano, Alaska

On August 6, 2008, a vigorous seismic swarm was detected by the automatic earthquake processing system at the Alaska Earthquake Information Center. It quickly became apparent that this swarm was occurring in close proximity to the little known Kasatochi Volcano in the central Aleutian arc. Starting at 15:00 UTC on August 6, the intensity and frequency of the earthquakes rapidly increased resulting in ~1,100 events located over the course of 48 hours during the most energetic part of the swarm. The largest earthquake, magnitude 5.8, occurred about 27 hours into the sequence and about 3.5 hours before the first aerial ash discharge was detected by the Alaska Volcano Observatory through satellite data. This discharge was followed by two more explosions, also detected via satellite monitoring.The 3rd explosion consisted of nearly continuous ash emissions and declined in frequency and intensity through ~22:00 UTC on August 8. The seismic swarm intensity started to decline at ~18:00 UTC on August 8, and continued to decline through the end of August. Overall, there is a strong causative relation between behavior of the seismic swarm and the eruption episodes. Due to seismic network limitations in the region (all stations are in a linear array along the island arc), determining accurate locations of the Kasatochi seismicity became a challenge. The volcano is located ~40 km NE of the nearest seismic network and only a single station in the vicinity was working to the east of the volcano (~80 km away). The locations have much larger uncertainties in the N-S direction than in the E-W direction, and overall are poorly constrained. Analysis of the travel-time differences between recording stations suggests an easterly migration of seismicity that started after the M5.8 earthquake. Moment tensor inversions of the 12 largest events resulted in solutions with large non-double-couple components. The rupture time history of the M5.8 earthquake, as determined from P-wave analysis, suggests that the non- double-couple component of the moment tensor does not reflect fluid flow directly. Rather, it may result from the movement of gas or a phase change as fluids exsolve from ascending magma.The consistent orientation of the moment tensors, combined with rupture lengths of a typical magnitude 5.8 earthquake and knowledge of the likely regional stress field, suggests intrusion of a dike system. This dike system may extend several kilometers from the volcano in a possible orientation ranging from NNW-SSE to E-W. To constrain the dike model better, we will present results of the earthquake relocations, waveform correlation, and source-time function analysis of the M5.8 earthquake.