FIRE and MUD Contents

Volcanic Earthquakes following the 1991 Climactic Eruption of Mount Pinatubo: Strong Seismicity during a Waning Eruption

By Jim Mori,1 Randall A. White,1 David H. Harlow,1 P. Okubo,1 John A. Power,1 Richard P. Hoblitt,1 Eduardo P. Laguerta,2 Angelito Lanuza,2 and Bartolome C. Bautista2

1U.S. Geological Survey.

2 Philippine Institute of Volcanology and Seismology.


One of the stronger sequences of volcanic earthquakes this century (cumulative seismic energy of 6.3x1013 joules) started during the June 15 eruption of Mount Pinatubo. The locations of these events were spread over a larger area (out to 20 kilometers from the summit) and greater depth extent (25 kilometers) than the preeruption seismicity. The intense rate of high-frequency volcano-tectonic earthquakes decreased rapidly over the first 2 weeks, following a smooth exponential decay. The locations of many of the events fall on outward-dipping trends that surround a region with relatively few earthquakes, which may be the magma reservoir that provided much of the material for the eruption. As the volcano-tectonic events diminished in late June, the low-frequency seismicity increased to levels similar to those during the preeruption period. The increase of low-frequency seismicity accompanied the transition in eruptive style from a steady outpouring of ash to more intermittent explosive activity. Episodic seismic activity developed into regular 7- to 10-hour intervals of increased low-frequency events often accompanied by large ash columns. Both the smooth decay of the high-frequency events and the periodicity of the low-frequency seismicity were interpreted as signs that the volcano was not building toward another large eruption.

Note to readers: Figures open in separate windows. To return to the text, close the figure's window or bring the text window to the front.


During the 1991 paroxsymal eruption of Mount Pinatubo on June 15, the seismic activity significantly increased and changed in character compared to the preeruption period. For the next several weeks the seismicity was dominated by high-frequency earthquakes that were distributed over a large volume extending out to 20 km from the volcano and to depths of 25 km. This was a marked change from the seismicity that was localized closer to the volcano summit at shallower depths during the preeruption period (Harlow and others, this volume).

The seismicity was the main source of information used to evaluate the state of the volcano, as was the case during the preeruption period (Harlow and others, this volume). Following the destruction of most of the seismic network during the eruption, the station on Clark Air Base (CAB) was the only local source of seismic data for the first 2 weeks following the climactic eruption. Although the seismic monitoring was severely limited, this single station located 17 km northeast of Mount Pinatubo (fig. 1) still proved useful in assessments of the possibility for future strong eruptions following the climactic eruption. The data from this instrument were particularly important during the first week, when very few visual observations could be made of the volcano because of the dense ash cloud that was constantly present and precluded visual observations. By the end of June, a seismic network of seven stations had been reinstalled, and the capability for recording and locating earthquakes was comparable to the preeruption period.

This paper summarizes the seismicity for the 2 months following the eruption of June 15. We note observations made during this period of declining seismicity and eruptive activity that we interpret as a waning eruption. During this time, the eruptive activity was small compared to June 15; however, the volcano still produced substantial ash columns to heights of over 18,000 m and small pyroclastic flows.

Figure 1. Locations of over 3,500 events recorded from June 29 through August 16, 1991. Hypocenters were calculated using a three-dimensional velocity model (Mori, Eberhart-Phillips, and Harlow, this volume). The west-east cross section shows earthquakes that were located within 5-km-wide slices centered on the summit. The south-north cross section shows earthquakes in a 3-km slice centered on the summit. Triangles mark seismic stations.


We present a summary of the seismic and eruptive activity from the time the strong sequence of earthquakes began on June 15 through the end of the eruptive activity in September 1991. We use terminology similar to Harlow and others (this volume) in describing the high-frequency volcano-tectonic earthquakes and the low-frequency activity, which includes long-period and the hybrid events as well as continuous volcanic tremor. Figure 2 shows a consistent measurement of seismic amplitudes at station CAB and summarizes the seismic activity for June and July. Also included is information about the eruptive activity, although the record of large explosions is largely incomplete for the posteruption period.

Figure 2. Hourly averages of seismic amplitudes at station CAB. Top trace is the total seismic amplitude recorded on the Real-time Seismic Amplitude Measurement (RSAM) system. Middle trace is the amplitudes of volcano-tectonic earthquakes inferred from the hourly event counts. Bottom trace is the amplitudes of the low-frequency activity obtained by subtracting the high-frequency amplitudes from the total amplitudes. The record of large explosions is incomplete for the posteruption period.

The strong sequence of earthquakes started with a magnitude 5 event at 1539 (all times are Philippine local time), several hours into the paroxysmal eruption that began at 1342 on June 15 (Hoblitt, Wolfe, and others, this volume). There are no local seismograms of this activity because all seismic instruments, except CAB, had been destroyed by the eruption, and the recording of this last station at Clark Air Base stopped when the scientific team abandoned the Pinatubo Volcano Observatory (PVO) at around 1500. Some of the automated seismic amplitude measurements continued to operate after the team left; these measurements provided some information during the time PVO was unmanned (Power and others, this volume). U.S. Geological Survey (USGS) National Earthquake Information Center (NEIC) reported 29 events with body-wave magnitudes (mb) >=4.5 and six events with mb >=5.0 within the next 6 hours. Earthquakes were being felt at the rate of nearly one per minute at the Arayat evacuation site, 35 km east of Mount Pinatubo. Probably sometime during this intense earthquake activity, the summit area collapsed to form the caldera, which is 2.5 km in diameter (W.E. Scott and others, this volume). The teleseismic locations of the large events have horizontal uncertainties of 10-15 km (Waverly Person, USGS, oral commun., 1994) and are not accurate enough to resolve if all of the earthquakes occurred in the immediate vicinity of the volcano or if some were associated with nearby regional faults (B. Bautista and other, this volume). The largest event reported by NEIC was an mb 5.7 earthquake at 1911 on June 15 that was located close to the volcano summit.

After we returned to PVO on June 16, continuous recording on a heliocorder resumed on the remaining seismic station (CAB) starting from 1100. Earthquakes of magnitudes 2.0 or greater were occurring at the rate of more than 150 per hour, but the rate was declining rapidly following a smooth decay curve (fig. 2). During this first week following the climactic eruption, there were only a few observable low-frequency events among the thousands of high-frequency volcano-tectonic events. At this time the volcano was continuously venting large amounts of ash without producing large explosions. On June 21 from 1500 to 2300, the low-frequency events significantly increased for the first time since June 15 (fig. 2). The seismogram recorded at station CAB during this period shows approximately equal numbers of high-frequency and low-frequency events, 20 to 50 events per hour over about magnitude 2.0. Associated with the low-frequency seismic activity were large explosions with ash columns up to 12,000 m.

Following the June 21 episode, volcanic tremor began to appear more frequently on the CAB record. Smaller amplitude tremor probably occurred continuously from June 15, but it could not be distinguished with the one station at CAB during the intense sequence of high-frequency events. Over the next 10 days the tremor increased steadily in duration and amplitude. By the end of June, the tremor was a continuous feature on the seismic record. During this time, individual low-frequency events also became more numerous. As the tremor and the low frequency events increased, the rate of volcano-tectonic earthquakes continued to decline smoothly (fig. 2). This change in the character of the seismicity correlated with a change in the character of the eruption. For the first week after June 15, the volcano was in a continuous mode of spewing out ash. There were only a few large explosions observed, and the ash column was fairly constant. In contrast, toward the end of June the continuous ash emissions became more intermittent, and more discrete, stronger explosions were observed (fig. 2).

Seismic stations were redeployed around Mount Pinatubo in late June (fig. 1), and by June 29 there were 6 operational stations to locate earthquakes. The hypocenters showed that the earthquakes were occurring throughout a much larger volume than before June 15. During the preeruption period, the earthquakes were restricted to two clusters of events within 7 km of the summit at depths from the surface to 8 km (figs. 4, 7, 9 in Harlow and others, this volume). In contrast, the posteruption volcano-tectonic earthquakes were distributed out to distances of more than 20 km from the summit and to depths greater than 25 km (fig. 1).

On June 30, low-frequency seismicity again increased to levels similar to those of June 21 and started a sequence of episodic activity. From July 7, the occurrence of the low-frequency activity became strikingly regular at 7- to 10-hour intervals. Low-frequency events that could be located all occurred at shallow depth near the summit area. During times that the volcano was visible, the peaks in the seismic activity correlated with increased eruptive activity. The near-predictability of these episodes enabled observers to photograph some of the larger ash columns that rose to over 20,000 m (fig. 3). The regular occurrences of the seismic and eruptive activity continued through July 11, although at increasing time intervals.

Figure 3. Ash column of July 7, 1991, at 1700, during one of the periodic peaks in low-frequency seismicity.

From the second week of July, the continuous tremor and episodes of low-frequency events began to occur at more irregular intervals. There were still episodes of large-amplitude tremor and low-frequency events, and throughout July the overall level of amplitudes for the tremor and low-frequency events remained comparable to the most intense preeruption activity on June 14 and 15 (fig. 2). The eruptive activity during this time was declining as explosions from the crater became fewer. As the rains of the monsoon season increased, the emphasis of the hazard evaluations concentrated less on the actual eruption and more on the numerous lahars (Pierson and others, this volume). From late July the seismicity began a slow decline that lasted through the next several months, reflecting the waning of the 1991 eruption. The eruptive activity had essentially stopped by early August, and the last reported explosion was on September 4, although that event could have been a secondary explosion. Occasional periods of continuous tremor and low-frequency activity were observed throughout 1991 and 1992 (Ramos and others, this volume).


In order to estimate the rate of seismic activity, we tabulated events from the paper recordings of station CAB. Events with maximum trace amplitudes of greater than 2 cm were counted (fig. 4). This would correspond to magnitudes of 2.0 to 2.5 near the summit area and smaller magnitude for events closer to Clark Air Base. The high rate of earthquakes during June 16 to 19 made counting the events difficult, so the numbers in figure 4 are probably underestimates during this period. The smooth decline of the volcano-tectonic earthquakes qualitatively resembles an aftershock sequence, but the shape of the decay is different. Aftershock rates following large earthquakes follow a power law that decays proportional to time (t-p), where p is close to 1.0 (Utsu, 1961). The rate at which the earthquakes diminished at Pinatubo follows an exponential decay, not a power-law decay. This is illustrated by the linear trend in the bottom panel of figure 4 which shows the logarithm of the event counts versus time. A power-law decay would produce a concave upward trend to the data in this plot. We can fit the data to an exponential curve,

N(t) = Ke-qt

where N(t) is the event count, K is a constant, t is time in days. The decay constant (q) was calculated to be 0.0095. If we force the data of the first 9 days to fit the power-law decay of aftershocks, using the method of Ogata (1983),

N(t) = K(t+C)-p

we obtain a p value of 2.42+-0.79 and a c value of 7.79+-3.76. Both of these of values are much higher than observed for earthquake aftershocks. Using longer time windows will yield even higher p values. An exponential decay or a power-law decay with large p value, points out that the rates of the volcano-tectonic earthquakes at Pinatubo diminished much faster than rates for aftershock sequences. This suggests differences in the geometry or mechanism for earthquakes following large volcanic events compared to aftershocks following large tectonic earthquakes.

Figure 4. Hourly event counts of high-frequency volcano-tectonic earthquakes tabulated from CAB. Top panel plots the data on a linear scale. Bottom panel plots the logarithm of the event counts.

Magnitudes for volcano-tectonic events in the range of 3.2 to 5.0 were estimated by using coda durations, and over 2,500 events were recorded in this range from June 16 to August 12 (table 1). We calculated a b-value for the sequence using these data along with magnitudes for the larger events (M> 5.0) which were taken from body-wave magnitudes determined by NEIC. The b-value calculated by use of the maximum likelihood method was 1.1+-0.3, which is typical of crustal tectonic earthquake sequences (Utsu, 1961). There were no significant differences in the rate of occurrence of larger events compared to smaller events; that is, no large changes in b-value were observed during the intense activity of high-frequency events in June. We also used the magnitude data (ML, local-wave magnitude) to estimate the radiated energy (E) by using a revised energy-magnitude relation from Kanamori and others (1993),

log E = 1.96 ML + 9.05

The cumulative radiated energy for the sequence from June 16 to August 12 was 6.3 x 1013 J. For comparison, this is larger than the radiated seismic energy for eruptions at Mount St. Helens, Washington but smaller than for Mount Katmai, Alaska (table 2).

Table 1. Magnitude (M) distribution of volcano-tectonic earthquakes.






June 16-20





June 21-30





July 1-31





August 1-12





Earthquake hypocenters were determined using a three-dimensional P-wave velocity structure determined from an inversion of arrival times from 298 selected events. Details about the velocity inversion and earthquake relocations are given in Mori, Eberhart-Phillips, and Harlow (this volume). The locations using this more complicated velocity structure, rather than a simple one-dimensional model, should be more accurate, considering the complex pattern of velocities often associated with volcanoes. The strengths of this method are especially useful around regions where there may be large low-velocity (magma) bodies. Mori, Eberhart-Phillips, and Harlow (this volume) show comparisons of the locations using the three-dimensional structure and a best-fitting one-dimensional structure. Locations plotted in figure 1 are for over 3,500 events recorded at more than five stations from June 29 through August 16. The horizontal uncertainties of the locations are less than 1 km, with vertical uncertainties of 1 to 3 km. The resultant earthquake locations are spread diffusely over a large area around the volcano with concentrations in the summit area (fig. 1). From the resumption of recording on June 16, the high-frequency events showed substantial variations in the relative arrival time between the S- and P-waves, indicating that the earthquakes were located at varying distances from the station and scattered throughout the volume. In cross section, these concentrations form trends that dip steeply away from Mount Pinatubo. The strongest clusters of earthquakes to the east and west of the summit clearly show trends that dip outward away from the volcano (east-west cross section in fig. 1). A group of earthquakes to the north of the summit also appears to dip steeply to the north, although the pattern is not as clearly defined for these events (north-south cross section of fig. 1). There are many earthquakes in the region directly below the summit from the surface to depths of about 4 km and then a striking lack of the earthquakes in the deeper central area.

The spatial distribution of the volcano-tectonic earthquakes can be described as three legs of a tripod around a volume relatively free of earthquakes. The tripod of earthquake locations may be partially outlining a magma reservoir that supplied much of the material for the eruption. This interpretation is supported by the observation of a very large low-velocity zone that corresponds to the central region between the earthquakes (Mori, Eberhart-Phillips, and Harlow, this volume).


Following the decline of the volcano-tectonic events in late June, episodes of low-frequency seismic activity dominated the seismic records from about July 1 (fig. 2). The low-frequency activity consisted of discrete events that resembled those recorded during the preeruption period and often grade into continuous tremor, which remained at a high level during episodes of increased activity. There was always some level of continuous tremor observable on the CAB (17 km from the summit) record from June 20 through about July 7. A few low-frequency events that had discernible P-waves were located at shallow (less than 3 km) depths near the summit. The shallow depth is also consistent with the relative amplitudes recorded on the network, which decrease rapidly with distance from the summit.

When the summit area was visible, the increases in the low-frequency seismic activity were correlated with periods of large ash explosions from the volcano. Some of ash columns produced during the periods of stronger seismicity were quite large, reaching heights of 18,000 m (fig. 3). There is a similarity in the amplitudes (fig. 2) and character (fig. 5) of the low-frequency waveforms recorded at CAB during the posteruption with those recorded during the large ash explosions on June 14. This is further evidence that much of the low-frequency activity is associated with the large ash eruptions.

Figure 5. Examples of low-frequency events taken from a section of a heliocorder record from seismic station CAB during one of the periodic peaks in activity on July 1, 1991, at 0900. Each line shows about a 3-minute portion of the 15-minute revolution of the recording drum.

The episodes of increased activity can be seen most easily on the Real-time Seismic Amplitude Measurement (RSAM) data shown in the top trace of figure 2. The RSAM data is a cumulative measure of the seismic amplitude for a specified time window (Murray and Endo, 1989). The data in figure 2 are from station CAB averaged over 1-hour time windows. From June 16 through late June the amplitude is dominated by the high-frequency earthquakes. With the gradual increase of the continuous tremor starting around June 24 and the decline of the high-frequency earthquakes, the RSAM is measuring predominantly the low-frequency seismic activity by July 1. Large-amplitude peaks on June 21, June 30, and the many subsequent ones are caused by the episodes of low-frequency earthquakes that were associated with increased ash emissions when the volcano was visible. Since some of the largest amplitudes are offscale, the heights of the peaks during the strongest activity are underestimated. However, the duration of the clipped signals is relatively short compared to the hour or more duration of the low-frequency activity, so the underestimate may not be significant.

We tried to separate the amplitudes of the volcano-tectonic earthquakes from the low-frequency earthquakes in the RSAM data by using the event counts shown in figure 4. We assumed that during the first 10 days following the June 15 eruption, the RSAM record is dominated by the volcano-tectonic events and that the shape of the time series matches the shape of the event-count curve. Theoretically, this should be the case for a b-value of 1.0 and amplitudes that add constructively. The event-count curve was fit to the RSAM curve to find an amplitude constant between the two curves for that time period. The amplitude-corrected event-count curve was subtracted from the RSAM curve to give a resultant time series that represents amplitudes of the low-frequency events and continuous tremor (bottom trace of fig. 2). The inferred low-frequency amplitudes (fig. 2) illustrate the progression of low-frequency activity described above. There were single peaks of activity on June 21 and June 30 and an increase in activity to the sequence of regularly spaced episodes starting on July 1.

The amplitudes of the low-frequency events and continuous tremor throughout most of July were relatively large and were comparable to the amplitudes measured during the intense low-frequency activity that occurred on June 14, 1 day before the climactic eruption (fig. 2). The large amplitudes for the low-frequency activity during the first 2 weeks of July (fig. 2) correspond to reduced displacements (Fehler, 1983) on the order of 102 cm, assuming the predominant frequencies are 2 Hz with wavelengths of 1 km. This estimate of reduced displacement and the corresponding explosive activity (fig. 3) are consistent with results of McNutt (1994) which show that the amplitude of tremor is indicative of the size of the corresponding eruptions.

This low-frequency activity during the posteruption period is similar in amplitude and character to the preeruption period after 1309 on June 14 (compare fig. 5 to fig. 13 of Harlow and others, this volume). The seismic activity on June 14 is associated with a time when the eruptive activity became more continuous, with pyroclastic flows emanating from multiple vents (Hoblitt, Wolfe, and others, this volume). This type of activity may be characteristic of an open system in which much of the low-frequency seismicity is directly associated with eruptive activity. This is in contrast to the sequence of large, discrete, long-period events observed earlier on June 14 (see fig. 12B in Harlow and others, this volume), which may be more indicative of pressurization inside the volcano (Chouet and other, 1994).


One striking feature of the low-frequency events and continuous tremor during the posteruption period was the regular peaking of activity observed in early July. During this period, the seismic activity increased to relatively high levels for about 2 to 3 hours and then declined to background levels. From July 7 through 11, the pattern became nearly periodic with increased activity occurring at 7- to 10-hour intervals. These regular intervals can be seen clearly in the RSAM data recorded for 10-min time windows from five stations (fig. 6). The time interval between the peaks was not strictly periodic but increased with time. On July 7, the time interval was about 7 hours and by July 11 had increased to about 10 hours. All the peaks in figure 6 can be correlated from station to station, except for the small peak of July 9 seen on BUG, which was produced by a burst of high-frequency amplitudes. This peak followed a period of heavy rain and was inferred to be a lahar signal.

Figure 6. Real-time Seismic Amplitude Measurement (RSAM) data from five stations showing the regularity of the low-frequency seismicity during the first week in July. Amplitudes are in counts which are proportional to the cumulative amplitudes recorded on the velocity sensors. Distances to the summit are shown after the station name.

Since the regular increases in seismicity were also associated with increases in ash production, the recognition of the periodic pattern made it possible to anticipate several of the larger ash columns. Weather conditions allowed good visibility from PVO for three of the eruptive periods (July 6 at 0600 and July 7 at 0500 and 1700). Concurrent visual and seismic observation showed that the ash column could attain a significant height (9,000-12,000 m) before the increases in seismic activity were observed. On August 1, there was a similar observation that the appearance of an ash column preceded the increases of low-frequency seismic activity by a few minutes.

Since the occurrence of low-frequency seismic events sometimes follows the beginning of ash emissions, we suggest that they may be triggered by the release of the overburden pressure. This is consistent with interpretations that the low-frequency events are associated with the process of gas vesiculation and may provide an explanation for the low level of low-frequency activity during the first week following the climactic eruption. During that time, the newly formed conduit system was wide open, as a constant stream of outpouring ash that did not allow any significant buildup of gas pressures. As the vent area cooled, small volumes within the volcano became sealed and provided regions where the pressure could increase to sufficient levels to produce small explosions and associated low-frequency events. In this model, the appearance of low-frequency activity is an indication of the transition from continuous ash emissions to periods of discrete eruptive episodes.



The occurrence of many felt earthquakes during the few days after the June 15 eruption raised concern of further strong eruptions; however, the smooth decay in the number of earthquakes (fig. 4) suggested the opposite. A systematic decay in the rate of seismicity that follows a smooth decay indicates that the earthquakes were a time-dependent response to a large perturbation, analogous to the occurrence of many aftershocks following a large mainshock. The removal of several cubic kilometers of material from under the volcano drastically changes the local stress conditions, and it is not surprising that a high level of seismicity was induced around that volume of material. Even though people near Mount Pinatubo were being shaken by many more earthquakes following the eruption compared to before, this seismicity was interpreted to be a response to the previous eruption and not precursory to further activity.

The decreasing rate of earthquakes following the June 15 eruption is superficially similar to an aftershock sequence; however, the decay law is different. Typical aftershock sequences decay proportional to t-1 (Utsu, 1961; Kisslinger and Jones, 1991). In contrast, the Pinatubo data show a more rapid exponential fall-off. This disagreement might be attributed to a difference in the source region or physical process. Hirata (1987) suggested that aftershocks associated with a fractured surface follow a power-law decay, while intact volumes of material subjected to stress produce seismic activity with exponential decays. The seismicity that occurred in the large volume of material around Pinatubo is more similar to the latter case, rather than aftershocks associated with a fault surface. An exponential decay of seismic activity was also observed following eruptions at Usu Volcano, Japan (Yokoyama and others, 1981) and may be a general characteristic of seismicity following volcanic activity.

The periodicity of the low-frequency seismicity and accompanying ash eruptions was interpreted as another sign that the volcano was not building toward more strong eruptions. Periodic activity usually implies a system that is near equilibrium with recurrent events that are controlled by repeating processes. Examples of these systems would be eruption cycles of geysers (Kieffer, 1984) or the banded tremor observed at Karkar Volcano, Papua New Guinea (McKee and others, 1981). The banded tremor at Karkar was eventually followed by a moderate eruption, but the periodicity was probably due to a local hydrothermal system. At Pinatubo, the regularity in eruptions was interpreted to mean that the pressures were at a generally stable level and were not building toward another large eruption. Furthermore, the time interval between eruptive episodes, which was lengthening slightly, suggested that the rates at which the pressures accumulated to produce the eruptions were decreasing.

One observation that raised concern was that the levels of low-frequency activity during the beginning of July were as large or larger than during much of the preeruption period. The RSAM levels (fig. 2) show that the amplitudes were comparable to the intense low-frequency activity on June 14 preceding the climactic eruption. Also, the character of the low-frequency activity in the posteruption period is similar to the seismicity observed after 1309 on June 14. However, we suggested above that this type of low-frequency seismicity be may characteristic of an open system and associated with ongoing eruptive activity, as distinguished from discrete long-period events, which may be more indicative of pressurization inside the volcano. Therefore, despite the large amplitude of the low-frequency seismicity, this activity was not interpreted to be an imminent sign of a large eruption.


The spatial distribution of the volcano-tectonic earthquakes was described as a tripod shape around a volume relatively free of earthquakes. The interpretations from the three-dimensional velocity inversion (Mori, Eberhart-Phillips, and Harlow, this volume) indicate that there may be a large magma body located within this boundary of earthquakes. The observed pattern of seismicity that partly surrounds a large magma body suggests the beginning of a ring-fault structure. The clusters of earthquakes seen in figure 1 would bound a block that has east-west dimensions of 5 to 8 km and north-south dimensions of 10 to 12 km. This raises the possibility that subsequent eruptions could continue to contribute to the formation of a ring structure, and eventually the pieces would coalesce into a complete ring fault and lead to formation of a large caldera during a future eruption of Pinatubo.

Outward-dipping trends of earthquakes have been observed at several other volcanoes, such as Rabaul caldera, Papua New Guinea (Mori and McKee, 1987), Mount St. Helens, Washington State (Scandone and Malone, 1985), and Mount Spurr, Alaska (Jolly and others, 1994), although only Rabaul caldera is associated with a well-identified ring fault. If these trends reflect the geometry of faults, they would be consistent with orientations that would form in response to a dilatational source (Anderson, 1936). Following an eruption, the source region of the magma could be viewed as a dilatational source because of the volume change from the loss of the erupted magma. The depths of the earthquakes at Mount Pinatubo, Mount St. Helens, and Mount Spurr, which extend to depths greater than 20 km, may indicate the extent of the dilatational source, suggesting relatively deep source regions for the erupted magma. For the smaller eruptions with smaller volume changes, such as at Mount St. Helens and Mount Spurr, the regional stress field can also be an important factor (Barker and Malone, 1991).


The occurrence of large earthquakes (M>5) during volcanic eruptions is often associated with large-scale deformation such as caldera formation or sector collapse (Okada, 1983). The activity at Pinatubo was one of the stronger seismic sequences related to a volcanic eruption this century and is probably directly or indirectly associated with the collapse of the summit area to form the small caldera (W.E. Scott and others, this volume), although it is unknown if the earthquakes were triggers or results of the collapse. The caldera with a diameter of 2.5 km has a collapse volume of about 2.5 km3.

Table 2 shows comparisons of the maximum size earthquake, total seismic energy, collapse volume, and other estimates of source parameters for several other large eruptions plus the smaller 1980 Mount St. Helens eruption. The cumulative energies in table 2 were recalculated by using the magnitude-energy relationship from Kanamori and others (1993) to make them consistent with the energy estimate for Pinatubo. The seismic energy releases at Katmai (Alaska) Pinatubo, and Fernandina (Galapagos Islands) occurred primarily during the few days of or following the climactic eruptions. The Pinatubo activity was stronger than the seismic sequence associated with a 2-km3 caldera collapse at Fernandina in 1968 (Filson and others, 1973) but much smaller than the series of large earthquakes that occurred at the 5-km3 collapse of Mount Katmai in 1912 (Abe, 1992). Abe (1992) suggests that there is a correspondence between the seismic energy and the collapse volume, which appears to be roughly supported by table 2; however, there have been some notable exceptions, such as the 1914 eruption of Sakurajima, Japan, which had an associated earthquake with a surface wave magnitude of 7.0 but no significant collapse (Abe, 1979). Furthermore, it seems unlikely that there can be a direct relation between surficial collapse volume and seismic energy of earthquakes at depth under the volcano. Large earthquakes following volcanic eruptions may be controlled more by local stress changes induced by volume changes of the magma chamber rather than by collapse volumes of the volcano. The seismic energy associated with the Mount St. Helens eruption is much smaller and may have a different mechanisms, since almost all of the energy release occurred during the 2 months prior to the cataclysmic eruption.

Table 2. Comparison of seismic activity for several significant eruptions of the century.

[Values are from the following references: Mount Katmai--Abe, 1992, Hildreth, 1983; Mount Pinatubo--W.E. Scott and others, this volume; Fernandina--Filson and others, 1973, Simkin and Howard, 1970; Mount St. Helens--Weaver and others, 1981, Moore and Albee, 1981. Seismic energies were recalculated by using the magnitude-energy relation from Kanamori and others (1993)]


Mount Katmai, Alaska,

Mount Pinatubo,

Fernandina, Galapagos Islands,

Mount St. Helens, Washington State,

Largest earthquake (Ms)





Cumulative energy (x1013 J)





Cumulative moment (x1017 Nm)










Erupted volume Magma eq. (km3)



< 0.2







Collapse volume (km3)





* includes both collapse and landslide volume at Mount St. Helens.

The eruptive activity at Pinatubo declined relatively quickly after June 15. The duration of strong activity following climactic eruptions at other large explosive volcanoes has also generally been short (Simkin and others, 1981), Krakatau (6 months), Santa María, Guatemala (1 month), Katmai (2 months). In many instances, the relatively short length of these larger eruptions makes them easier to deal with in terms of issuing hazard evaluations, compared to smaller eruptions that can continue at a sustained or irregular level of activity for a year or more, such as recent activity at Unzen, Japan, or Galeras, Colombia.


The strong sequence of volcano-tectonic earthquakes that began during the climactic eruption of June 15 is thought to reflect the local response to a large volume change in the magma chamber underlying the volcano. The large areal and extended depth distribution of the earthquakes suggests that there were significant volume changes down to depths of 20 km. The smooth exponential decay in the rate of these events supports the idea that the high-frequency earthquakes were a response to the eruption and not indications of increasing volcanic activity. The smooth decay of the seismicity was similar to the decreasing rates of aftershocks following a large tectonic earthquake, except that the decay curve at Pinatubo was exponential, in contrast to power-law decays observed in aftershocks. Much of the seismicity is located on outward-dipping trends that surround a volume directly under the volcano that is relatively free of earthquakes. This region at 7 to 20 km in depth may be the magma reservoir that supplied much of the material for the 1991 eruption (Mori, Eberhart-Phillips, and Harlow, this volume).

Strong increases in the low-frequency activity observed in late June were comparable to levels observed during the preeruption period. These episodes of low-frequency events were often accompanied by large ash columns. The low-frequency seismicity may reflect the transition from continuous ash emissions during the first week following June 15 to periods of discrete eruptive episodes. During the first week of July, the episodic behavior became very regular, having increases in seismicity and eruptive activity at 7- to 10-hour intervals. The near-periodicity of the activity and the gradual lengthening of the time interval were also thought to be indications that the volcano was relatively stable and not building toward another large eruption.

The decrease of high- and low-frequency seismicity along with the decline of the eruptive activity throughout July indicates an end to the 1991 eruption in early August, excepting one possible small explosion in September.


The data and interpretations presented in this paper were made possible because of the considerable efforts of T.L. Murray, F. Fischer, L. Bautista, E.T. Endo, J.W. Ewert, A.B. Lockhart, J. Lockwood, J.N. Marso, A. Miklius, C.G. Newhall, E. Ramos, E.W. Wolfe, and others. We thank U.S. Air Force personnel at Clark Air Base for their support. Helpful comments on the manuscript were provided by S. McNutt, C.G. Newhall, and 2 anonymous reviewers.


Abe, K., 1979, Magnitudes of major volcanic earthquakes of Japan 1901 to 1925: Journal of the Faculty of Science, Hokkaido University, Series VII (Geophysics) v. 6, p. 201-212.

------1992, Seismicity of the caldera-making eruption of Mount Katmai, Alaska 1912: Bulletin of the Seismological Society of America, v. 82, p. 175-191.

Anderson, E.M., 1936, The dynamics of the formation of cone-sheets, ring-dykes, and caldron-subsidences: Proceedings of the Royal Society of Edinburgh, v. 56, p. 128-156.

Barker, S.E., and Malone, S.D., 1991, Magmatic system geometry at Mount St. Helens modeled from the stress field associated with posteruptive earthquakes: Journal of Geophysical Research, v. 96, p. 11883-11894.

Bautista, B.C., Bautista, M.L.P., Stein, R.S., Barcelona, E.S., Punongbayan, R.S., Laguerta, E.P., Rasdas, A.R., Ambubuyog, G., and Amin, E.Q., this volume, Relation of regional and local structures to Mount Pinatubo activity.

Chouet, B.A., Page, R.A., Stephens, C.D. Lahr, J. C., and Power, J.A., 1994, Precursory swarms of long-period events at Redoubt Volcano (1989-1990), Alaska: Their origin and use as a forecasting tool: Journal of Volcanology and Geothermal Research, v. 62, p. 95-135.

Fehler, M, 1983, Observations of volcanic tremor at Mount St. Helens Volcano: Journal of Geophysical Research, v. 88, p. 3476-3484.

Filson, J., Simkin, T., and Leu, L.-K., 1973, Seismicity of a caldera collapse: Galapagos Islands: Journal of Geophysical Research, v. 79, p. 8591-8622.

Harlow, D.H., Power, J.A., Laguerta, E.P., Ambubuyog, G., White, R.A., and Hoblitt, R.P., this volume, Precursory seismicity and forecasting of the June 15, 1991, eruption of Mount Pinatubo.

Hildreth, W., 1983, The compositionally zoned eruption of 1912 in the Valley of Ten Thousand Smokes, Katmai National Park, Alaska: Journal of Volcanology and Geothermal Research, v. 18, p. 1487-1494.

Hirata, T., 1987, Omori's power law aftershock sequences of microfracturing in rock fracture experiments: Journal of Geophysical Research, v. 92, p. 6215-6221.

Hoblitt, R.P., Wolfe, E.W., Scott, W.E., Couchman, M.R., Pallister, J.S., and Javier, D., this volume, The preclimactic eruptions of Mount Pinatubo, June 1991.

Jolly, A.D., Page, R.A., and Power, J.A., 1994, Seismicity and stress in the vicinity of Mt. Spurr volcano, south central Alaska, Journal of Geophysical Research, v. 99, p. 15305-15318.

Kanamori, H., Mori, J., Hauksson, E., Heaton, T.H., Hutton, L.K., and Jones, L.M., 1993, Determination of earthquake energy release and ML using TERRAscope: Bulletin of the Seismological Society of America v. 83, p. 330-346.

Kieffer, S.W., 1984, Seismicity at Old Faithful Geyser: An isolated source of geothermal noise and possible analogue of volcanic seismicity: Journal of Volcanology and Geothermal Research, v. 22, p. 59-95.

Kisslinger, C., and Jones, L.M., 1991, Properties of aftershock sequences in southern California: Journal of Geophysical Research, v. 96, p. 11947-11958.

McKee, C.O., Wallace, D.A, Almond, R.A., and Talai, B., 1981, Fatal hydro-eruption of Karkar volcano in 1979: Development of a maar-like crater, in Johnson, R.W., ed., Cooke-Ravian volume of volcanological papers: Geological Survey of Papua New Guinea Memoir 10, p. 63-84.

McNutt, S. R., 1994, Volcanic tremor amplitude correlated with eruption explosivity and its potential use in determining ash hazard to aviation, in Casadevall, T.J., ed., Proceedings at the First International Symposium on Volcanic Ash and Aviation Safety: U.S. Geological Survey Bulletin 2047, p. 377-385.

Moore, J.G., and Albee, W.C., 1981, Topographic and structural changes at Mount St. Helens: Large-scale photogrammetric data, in Lipman, P.W., and Mullineaux, D.R., eds., The 1980 eruptions of Mount St. Helens, Washington: U.S. Geological Survey Professional Paper 1250, p. 123-134.

Mori, J., and McKee, C.O., 1987, Outward-dipping ring-fault structure at Rabaul Caldera as shown by earthquake locations: Science, v. 235, p. 193-195.

Mori, J., Eberhart-Phillips, D., and Harlow, D.H., this volume, Three-dimensional velocity structure at Mount Pinatubo, Philippines: Resolving magma bodies and earthquake hypocenters.

Murray, T.L., and Endo, E.T., 1989, A real-time seismic amplitude measurement system (RSAM): U.S. Geological Survey Open-File Report 89-684, 21 p.

Ogata, Y., 1983, Estimation of the parameters in the modified Omori formula for aftershock sequences by the maximum likelihood procedure: Journal of Physics of the Earth, v. 31, p. 115-124.

Okada, H., 1983, Comparative study of earthquake swarms associated with major volcanic activities, in Shimozuru, D., and Yokoyama, I., eds., Arc volcanism: Physics and tectonics: Tokyo, Terra Scientific Publishing Company, p. 43-61.

Pierson, T.C., Daag, A.S., Delos Reyes, P.J., Regalado, M.T.M., Solidum, R.U., and Tubianosa, B.S., this volume, Flow and deposition of posteruption hot lahars on the east side of Mount Pinatubo, July-October 1991.

Power, J.A., Murray, T.L., Marso, J.N., and Laguerta, E.P., this volume, Preliminary observations of seismicity at Mount Pinatubo by use of the Seismic Spectral Amplitude Measurement (SSAM) system, May 13-June 18, 1991.

Ramos, E.G., Laguerta, E.P., and Hamburger, M.W., this volume, Seismicity and magmatic resurgence at Mount Pinatubo in 1992.

Scandone, R. and Malone, S.D., 1985, Magma supply, magma discharge, and readjustment of the feeding system of Mount St. Helens during 1980: Journal of Volcanology and Geothermal Research, v. 23, p. 239-262.

Scott, W.E., Hoblitt, R.P., Torres, R.C., Self, S, Martinez, M.L., and Nillos, T., Jr., this volume, Pyroclastic flows of the June 15, 1991, climactic eruption of Mount Pinatubo.

Simkin, T. and Howard, K.A., 1970, Caldera collapse in the Galapagos Islands, 1968: Science, v. 169, p. 429-437.

Simkin, T., Siebert, L., McClelland, L., Bridge, D., Newhall, C., and Latter, J.H., 1981, Volcanoes of the World, a regional directory, gazetteer, and chronology of volcanism during the last 10,000 years: Stroudsburg, Pa., Hutchinson Ross Publishing Company, 232 p.

Utsu, T., 1961, A statistical study on the occurrence of aftershocks: Geophysical Magazine, v. 30, p. 521-605.

Weaver, C.S., Grant, W.C., Malone, S.D., and Endo, E.T., 1981, Post-May 18 seismicity: Volcanic and tectonic implications, in Lipman, P.W., and Mullineaux, D.R., eds., The 1980 eruptions of Mount St. Helens, Washington: U.S. Geological Survey Professional Paper 1250, p. 102-122.

Yokoyama, I., Yamashita, H., Watanabe, H., and Okada, Hm., 1981, Geophysical characteristics of dacite volcanism--the 1977-1978 eruption of Usu volcano: Journal of Volcanology and Geothermal Research, v. 9, p. 335-358.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey

This page is <>
Contact: Chris Newhall
Last updated 06.11.99