Some Seismic Precursors to Eruptions at Pavlof Volcano, Alaska, October 1973–April 1986

This chapter examines 13 years of seismicity data (1973–1986) at Pavlof Volcano for eruption precursors. Changes in the rate of occurrence of earthquakes give the most consistent information. Tests for changes in rate of occurrence were performed using different values for short-term (STA) and long-term (LTA) average seismicity, and STA/LTA thresholds. The best parameters of LTA = 3 days, STA = 2 days, threshold = 3x, and absolute number of events \u3e150 per day yielded 46% precursors, 33% false alarms, and the remainder missed eruptions. Seismic activity at Pavlof is monitored by an array of five seismometers which are located at distances of 4–11 km from the summit vent. The numbers of earthquakes increase from a background level averaging 33 low-frequency or B-type events per day to an average of 453 events per day during phreatomagmatic eruptions, and to 1919 events per day during magmatic eruptions (volcanic tremor, which occurs during magmatic eruptions, is converted to an equivalent number of B-type earthquakes, and included in this total). Data were also systematically searched for eight other general types of seismic precursors: (1) changes in event locations; (2) changes in event sizes; (3) changes in frequency content; (4) changes in b-value; (5) seismic quiescence; (6) changes in focal mechanisms; (7) changes of event types; and (8) onset of tidal (or other) periodicities. None of these gave results as good as changes in rate of occurrence. Seismic monitoring alone has provided some useful data on eruption precursors, but, based on comparisons with other volcanoes, geodetic monitoring of the volcano is considered likely to improve the short-term eruption forecasting capability