Peak Rates and Largest Magnitude Events in Earthquake Swarms From Different Tectonic Settings

Document Type


Publication Date



We present results of recent studies of earthquake swarms from local data at volcanoes, and teleseismic data from MOR and volcanic regions, focusing on identifying diagnostics. One clear pattern for local volcanic swarms is that peak rates often occur early, whereas the largest magnitude event occurs late. Using a dataset of 20 swarms from the literature, swarm durations, measured from swarm onset to eruption onset, ranged from 12 h to 180 d. Data were normalized to % duration. Peak rates occurred from 1-42% of the total duration (with 2 outliers), whereas the largest M event occurred from 32-99% of duration. Additional evidence from 4 cases suggests that the seismic source size grows systematically, especially for events with similar waveforms (families). This is revealed in plots of size versus time for event families. For comparison, 19 cases of mid-ocean ridge swarms and 67 cases of teleseismic volcanic swarms were analyzed. The MOR data show durations of 1-42 d, with peak rates occurring 1-24% of duration and largest M occurring 1-87% of duration. In 6 cases largest M occurs before or at the same time as peak rate. The teleseismic volcanic data show durations of <1 to 577 d, with peak rates occurring 1-100% of duration and largest M occurring 1-100 % of duration. Thus the patterns for MOR and teleseismic volcanic swarms are similar to each other but differ significantly from that for local volcanic swarms. Further work on volcanic swarms shows the distribution of seismicity before the peak rate differs from after, suggesting two dominant processes. The durations of post-peak portions are roughly proportional to the peak rates. This is similar to the behavior of aftershock sequences and suggests diffusion is a controlling process. The portions of the swarms prior to the peaks behave differently, however. These may represent invasion of hot fluids and opening or reopening of cracks prior to magma intrusion. We infer that the growth in event size reflects activation of a preferred magma pathway. Recognition of such patterns, linked to processes, may help to improve monitoring and reduce risks from eruptions. Comparison is recommended between patterns observed here and those associated with induced seismicity from hydraulic fracturing and deep well injection.

Was this content written or created while at USF?


Citation / Publisher Attribution

Presented at the 9th Biennial Workshop on Japan-Kamchatka-Alaska Subduction Processes in May 2016 in Fairbanks, AK