Data Limitations and Discussion

The observation that the mean duration of swarms that precede eruptions is about twice as long as swarms that are not associated with eruptions may be due to: 1) reporting bias, 2) mis-indentification of small tectonic mainshock-aftershock sequences as swarms, or 3) different time scales for the processes involved with transport of magma to the surface, when compared with intrusion or other transient forcing phenomena (tidal, barometric pressure fluctuations, seasonal ocean-loading, etc.).

Reporting bias is a possible source of error which must be considered when interpreting these results. The database was compiled primarily from the BVE where the emphasis is the reporting of eruptions. Therefore, there may be a tendency more frequently to report seismic activity associated closely with eruptions as opposed to swarms that occur at volcanoes with little or no historic activity. Furthermore, if an eruption occurs, the reporter may examine the preceding seismicity more rigorously, and perhaps include a longer period of time as the precursory seismicity.

The mean magnitude for the largest shocks in volcanic areas is about M3, based on data from 113 swarms at 61 volcanoes in the GVESD. Using aftershock decay parameters given by Reasenberg and Jones (1989) for tectonic earthquakes in southern California, the duration of an aftershock sequence following a M3 is about half a day and for a M4, 3.5 days. Thus, the durations of a small mainshock-aftershock sequences are similar to the mean duration of Type III swarms. In areas where the magnitudes are not available or not reliable, small tectonic earthquakes and their aftershocks may be reported as swarms.

The above problems of reporting bias and mis-identification are certainly factors in some of the reports. However, we believe that given a large sample size, these effects will not unduly bias the general result. With these limitations in mind, we can speculate that differing earthquake swarm durations are due to several suites of physical processes operating at different time scales. For example, the ascent of magma to the surface may express itself in longer lasting swarms, while intrusions or failed eruptions are manifested by shorter swarms. Other factors not directly associated with the movement of magma may also lead to shorter duration sequences of earthquakes. Volcanic and geothermal areas have been shown to be sensitive to small strains. Such strains can be generated by earth and ocean tidal stresses (Rydelek and others, 1988; McNutt and Beavan, 1981; Klein, 1976), body and surface waves from regional or teleseimic earthquakes (Hill et al., 1993), seasonal ocean-loading (McNutt and Beavan, 1987), or changes in barometric pressure (Rinehart, 1980).

Volcanic earthquake swarms, unlike tectonic mainshock-aftershock sequences, do not release the majority of seismic energy in the largest earthquake of sequence. Swarms, by definition, have one or more shocks of similar magnitude. Therefore, the seismic energy released during a swarm is spread over a longer period of time than tectonic mainshock-aftershock sequences. The difference in swarm duration distributions suggests that duration is more likely to reflect future eruptive activity than the magnitude of the largest event within a given swarm.

Move on to the [NEXT] page.
Return to the GVESD or to the AVO home page.


This page is designed and maintained by John Benoit. Your comments and suggestion are welcome. benoit@giseis.alaska.edu