FIRE and MUD Contents

Precursory Seismicity and Forecasting of the June 15, 1991, Eruption of Mount Pinatubo

By David H. Harlow,1 John A. Power,1 Eduardo P. Laguerta,2 Gemme Ambubuyog,2 Randall A. White,1 and Richard P. Hoblitt1

1U.S. Geological Survey.

2Philippine Institute of Volcanology and Seismology.


ABSTRACT

Seismic monitoring was the primary tool used to assess the evolving eruptive potential of Mount Pinatubo prior to a climactic eruption on June 15, 1991. We used a seven-station seismic network and a portable, PC-based data acquisition and analysis system to track seismic activity in real time and near-real time. We divide seismic activity prior to the eruption into five distinct phases: (1) Metastable (through May 31)--characterized by 50 to 150 volcano-tectonic events per day; (2) Predome (June 1 to 7)--escalating seismic activity below the summit of Mount Pinatubo in early June which evolved into a strong swarm of volcano-tectonic earthquakes that culminated in the extrusion of a lava dome at the surface; (3) Preexplosive Buildup (June 8 to 12)--characterized by a variety of seismic activity leading up to the first explosive eruption on June 12, including volcano-tectonic earthquakes accompanying the extrusion of a dome, tremor episodes, and the occurrence of hybrid and long-period events; (4) Long-Period Buildup (June 12 to 14)--characterized by explosive eruptions with small pyroclastic flows and a remarkable progression of increasingly large long-period events; and (5) Preclimactic (June 14 to 15)--a 24 hour period of high-amplitude, widely diverse seismic activity that coincided with a series of increasingly frequent explosive eruptions; these produced large pyroclastic flows that evolved into the continuous climactic eruption that began about 1342 on June 15.

We were able to forecast the escalating eruptive potential on the basis of shifts in the locus of earthquake hypocenters, increases in total seismic energy release, changes in the character of earthquake waveforms, escalating nonseismic precursory activity, and an interpretation of magmatic processes during the precursory phenomena. The key changes in seismic activity on which we based our forecasting evaluations included (1) a shift in the locus of the dominant seismic source region during late May and early June from a cluster 5 kilometers northwest of the summit to a cluster just beneath the summit, (2) an intense swarm of volcano-tectonic earthquakes culminating in the extrusion of a dome, (3) the increase in amplitude of continuous background tremor prior to the first explosive eruption on June 12, and (4) the dramatic increase in seismic energy release and in the number and magnitude of long-period events prior to the climactic eruption on June 15.

The successful forecast of the Mount Pinatubo eruption is a confirmation of our current capability to quickly install a network of monitoring instruments at remote volcanoes, analyze the data, and then provide an adequate interpretation of ongoing magmatic processes for eruption forecasting. The evolution of seismic activity leading up to the Mount Pinatubo eruption has critical implications for eruption forecasting at large volcanic systems where crisis-level seismicity persists for as long as years but has not led to eruptions. At Mount Pinatubo, only 1 week elapsed from the time we recognized that seismic activity differed significantly from crisis-level seismicity at other large volcanic systems to the time of the first explosive eruption on June 12.

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.

INTRODUCTION

The value of seismic data for forecasting volcanic eruptions was recognized as early as 1910 for eruptions of Usu Volcano, Japan (Omori, 1911) and Mauna Loa Volcano, Hawaii (Wood, 1915). Long-term seismic monitoring of volcanoes led to the discovery of a diverse range of seismic signatures and patterns associated with eruptive activity (Sassa, 1935, 1936). A widely used classification scheme developed by Minakami (1960) was based on the characteristics of seismic event signatures and their hypocentral locations. This classification system was used for eruption forecasting by correlating the progression of various parameters of each event type and then matching emerging data patterns with seismicity patterns that preceded eruptions at the same and other volcanoes. The drawback to this technique, however, is that the wide variety of seismic activity and different eruption styles can make the application of pattern recognition for eruption forecasting difficult and unreliable.

Over the last decade, significant improvements have been developed in using the different types and characteristics of seismic activity to identify the causative processes. This approach has evolved from theoretical modeling of source mechanism of different event types (Chouet, 1992) and the study of data from eruptions at "instrumented" volcanoes including Kilauea (Klein and others, 1987), Mount St. Helens (Endo and others 1981; Malone and others, 1981), Augustine (Power, 1988), Redoubt (Power and others, 1994), and Spurr (Power and others, in press) in the United States and Nevado del Ruiz (Nieto and others, 1990) in Colombia. While precursory seismicity at those volcanoes varied in duration, intensity, and character, a synthesis of those data sets provides us with a set of practical guidelines for identifying seismic activity associated with critical preeruptive processes.

At Mount Pinatubo (fig. 1), we were able to forecast eruptive activity by closely observing earthquake hypocenters, the waveform character of individual events, seismic energy release, and seismicity rate. The overall lessons learned from our Pinatubo crisis response have been summarized by the Pinatubo Volcano Observatory Team (1991). We focus here on the interpretation of seismic data from Mount Pinatubo and discuss how this experience can contribute to the overall forecasting effort. Specifically, we describe the temporal and spatial evolution of seismic activity leading to the climactic eruption of Mount Pinatubo on June 15, 1991, and discuss how the seismic data were used for eruption forecasting and volcano hazard management. We begin with a description of the seismic network and the data acquisition and analysis system used to monitor the volcano and then discuss the various types of seismic events observed at Mount Pinatubo. We follow with a chronology of seismic events and related phenomena observed before the climactic eruption and conclude with a discussion of the implications of preeruptive seismicity for the evolution of the Mount Pinatubo eruption. In this narrative we include insight into the forecasting deliberations as they occurred, without the comforting clarity of hindsight.

Figure 1. Location map of the vicinity of Mount Pinatubo.

INSTRUMENTATION, DATA ACQUISITION, AND ANALYSIS

Shortly after the phreatic explosions on April 2, several portable seismographs were deployed on the west side of Mount Pinatubo (Sabit and others, this volume). In late April and early May, a network of seven radio-telemetered seismic stations was installed as part of a joint Philippine Institute of Volcanology and Seismology (PHIVOLCS)/U.S. Geological Survey Volcano Crisis Assistance Team effort. The stations were located at distances of 1 to 19 km from the volcano's summit (fig. 2). Single-component vertical seismometers with a natural frequency of 1-Hz were used at all stations except PPO, which used a 2-Hz three-component geophone. The first station was installed on April 28, followed by three more during the next 9 days, allowing the reliable calculation of earthquake hypocenters by May 7. Installation of the network was completed on May 13. The network operated satisfactorily until shortly after 2200 (all dates and times are Philippine local time; to convert to G.m.t. subtract 8 h) on June 11, when station UBO failed. Vandals subsequently disabled the radio-relay site for stations PPO and BUR on June 12. Stations PIE, GRN, and BUG were lost during the climactic eruption on June 15. Thus, CAB was the only station to operate continuously throughout the eruptions of Mount Pinatubo, although its recording was interrupted for several hours on June 15 when power to the acquisition system was lost. The installation, operation, and technical specifications of this network are described in Lockhart and others (this volume).

Figure 2. Map of radio-telemetered seismic stations installed around Mount Pinatubo, April-June 1991.

The seismic signals from these stations were transmitted to Clark Air Base and recorded on a PC-based data acquisition system as well as a number of drum recorders. The networked system of PC's represents a significant improvement in our ability to monitor remote volcanoes effectively, analyze the data quickly, and provide accurate eruption forecasts. The system provides for the digital recording of seismic waveforms, hypocenter and magnitude calculation, Real-time Seismic Amplitude Measurements (RSAM) (Endo and Murray, 1991), and Seismic Spectral Amplitude Measurements (SSAM) (Stephens and others, 1994). The acquisition system is described in detail by Murray and others (this volume).

Seismic data from the Pinatubo network were processed by a variety of computer programs. For each seismic event detected and saved by the acquisition system, phase arrival and signal duration data from individual stations were determined by use of the program PCEQ (Valdez, 1989). These data were then used to calculate earthquake hypocenters and coda magnitudes through the use of the program HYPO71 (Lee and Valdez, 1989) with a horizontally layered velocity model. The large volume of data and severe time constraints allowed only a rapid first pass through the data analysis scheme during the crisis response in 1991. Low-quality hypocenter solutions, due to poor or incorrect phase data, were rejected and not reanalyzed. The earthquake hypocenters and magnitudes presented in this paper were recalculated by use of more sophisticated analysis techniques than were available at Clark Air Base at the time of the eruption. Phase-arrival times were repicked by use of the program XPICK (Robinson, 1992), and hypocenters were again calculated using a layered structure. This posteruption processing reduced the dispersion of earthquake hypocenters and thereby make the earthquake clusters appear more compact than they appeared during our crisis response. Hypocenters were then recalculated by use of a three dimensional-velocity model derived by Mori, Eberhart-Phillips, and Harlow (this volume).

The computer program ACROSPIN (Parker, 1990) permitted a two-axis rotation of plots on a computer screen either in a continuous or stepwise mode. During the crisis response, we gained an important impression of the three-dimensional development of seismic activity by using ACROSPIN to rotate earthquake hypocenter plots. Following the eruption Hoblitt, Mori, and Power (this volume) developed a computer program called VOLQUAKE, which displays the evolution of earthquake hypocenters throughout the Pinatubo eruption on a computer screen. The program uses colored symbols to represent earthquakes at various depths and allows the user to view the hypocenters in three dimensions from any perspective. This program offers what is perhaps the best representation of the development of earthquake activity at Pinatubo through the 1991 eruption sequence.

The RSAM system provides 10-min averages of the absolute seismic amplitude for each seismic station. RSAM data are a particularly useful tool for monitoring the overall level of seismic activity during periods of high and rapidly changing activity. RSAM data have been successfully used as an evaluation and forecasting tool at Mount St. Helens (Endo and Murray, 1991) and during the 1989-90 eruptions of Redoubt Volcano, Alaska (Power and others, 1994). RSAM data proved to be particularly valuable during the Pinatubo crisis, especially during the final days of the precursory seismic sequence, when activity escalated and a number of stations were destroyed or disabled. RSAM data were analyzed and displayed in realtime by use of the program BOB (Murray, 1990).

A Seismic Spectral Amplitude Measurement (SSAM) system was also used at Mount Pinatubo. This relatively new tool (Stephens and others, 1994) was developed during the 1989-90 eruptions of Redoubt Volcano, Alaska, and was successfully used to forecast some eruptions of Redoubt Volcano in the spring of 1990. SSAM is a refinement of the RSAM system and provides 1-min averages of the spectral amplitudes within several narrow frequency bands for each seismic station. SSAM results were not incorporated in formulating eruption forecasts at Mount Pinatubo because only cumbersome software was available to display the data in 1991. The SSAM data, however, have proved valuable in reconstructing the spectral character of seismicity preceding the Pinatubo eruption (Power and others, this volume) and promise to be a valuable tool for future eruption forecasting.

EVENT CLASSIFICATION

Chouet and others (1994) argue convincingly for classifying the diverse seismic signatures produced by volcanic activity on the basis of the physics of their source processes. The advantage of a process-oriented scheme is that the observed seismic activity is directly linked to ongoing volcanic processes. Chouet and others (1994) separate volcanic seismicity into two basic families of processes. The first family includes events caused by the brittle failure of rock as a result of stresses induced by magmatic processes. The second family consists of sources in which fluid plays an active role in the generation of seismic waves. Included in this family are long-period events, tremor, and signatures produced by degassing activity. We follow a process-based classification system for the Pinatubo seismic data and describe below the various types of events observed.

Volcano Tectonic (VT) Earthquakes.--The signature of a typical VT earthquake is illustrated in figure 3A. Such earthquakes are characterized by sharp, mostly impulsive onsets, and their spectra are typically broad, between 1 and 15 Hz (Lahr and others, 1994). They are called VT earthquakes because their signatures are indistinguishable from those of tectonic earthquakes, although the stresses that trigger them are derived from magmatic processes rather than large-scale tectonic movements (Chouet and others, 1994). VT earthquakes occur as single events as well as sequences of rapidly occurring events with overlapping codas, called flurries by Hill and others (1985).

Long-period (LP) Events.--The signatures of LP events are characterized by a high-frequency onset followed by a lower frequency quasimonochromatic coda (fig. 3A). The spectra of LP events typically show a strongly peaked response in contrast to the broad spectra of VT earthquakes (Lahr and others, 1994). This spectral criterion is used to distinguish an LP event from the classic B-type earthquake (Minakami, 1960), which is a VT earthquake occurring at a depth of 1 km or less. Because path effects can significantly modify their signatures, such VT earthquakes are often confused with true LP events. However, the broad spectra of VT earthquakes serve to distinguish them from LP events. Theoretical modeling by Chouet (1985, 1988, 1992), Ferrazzini and others (1990), and Chouet and others (1994) suggests that the signal characteristics of LP events are comparable to the synthetic signatures generated by the excitation of a fluid-filled crack or conduit in response to pressure transients. Thus, LP events are considered to be indicators of pressurization in volcanic conduits and, as such, represent critical precursors to volcanic eruptions. The source of LP events and tremor are in many cases thought to be closely related (Chouet, 1992; McNutt, 1992), and for this reason we refer to periods when both individual and LP events are occurring as LP seismicity.

An example of an LP event is shown in figure 3A, and the distinct character of its signature is evident compared with that of a VT earthquake (also shown in fig. 3A). Another illustration of LP activity from Mount Pinatubo is the sequence of events shown in figure 3C. Posteruption analysis revealed that this sequence of events originated at depths of more than 30 km (White, this volume).

Hybrid Events.--Hybrid events exhibit signatures that combine characteristics of both LP and VT earthquakes (Chouet and others, 1994; Lahr and others, 1994, Power and others, 1994). These events generally have broad spectra with a distinct spectral peak similar to an LP event. Hybrid events are thought to be generated by brittle fracturing either through or near fluid-filled cracks, which thereby generates seismic waves from source processes associated with both end member types (Lahr and others, 1994).

Volcanic Tremor.--Tremor denotes a wide range of irregular amplitude seismic signals with durations that can range from minutes to days, weeks, and even longer. The source mechanisms for tremor vary widely (Chouet, 1988, 1992; McNutt, 1992), and several types of tremor tied to specific source processes are found in the literature (Aki and others, 1977; Aki and Koyanagi, 1981; Chouet, 1985, 1988, 1992; Chouet and Shaw, 1991; Chouet and others, 1994; Koyanagi and others, 1987; Ferrazzini and others, 1990; Shaw and Chouet, 1991). Harmonic tremor refers to continuous signals of quasi-monochromatic appearance with dominant frequencies of a few hertz (figs. 3B and 3C). Volcanic tremor and LP events are intimately linked (Koyanagi and others, 1987), and the latter have been interpreted as the impulse response of the tremor-generating source (Chouet, 1985). Figure 3C shows a tremor episode composed of a closely spaced series of LP events that occurred at an unusual depth of 35 km (White, this volume).

Eruption Signals.--The signatures of eruptions generally have emergent onsets and extended codas, although the onset can be quite sudden in some cases (Power and others, 1994). At Pinatubo, the durations of eruption signals varied greatly; minor hydrothermal or ash eruptions were barely perceptible on the seismic records of the nearest high-gain stations and lasted only a few seconds, whereas the climactic eruption on June 15 produced a continuous signal for more than 12 h at regional stations at distances of more than 300 km. The signature of an explosive eruption at Mount Pinatubo is shown in figure 3D.

Figure 3. Examples of seismic event types observed at Mount Pinatubo (station PIE). A, Volcano-tectonic earthquake (VT) and long-period event (LP). B, Tremorlike episode of closely-spaced long-period events. C, Harmonic tremor. D, Explosive eruption at station CAB. Time marks represent 1-min interval.

SEISMIC CHRONOLOGY

We divide the period April 5 to June 15, prior to the climactic eruption, into five phases based on the level, type, character, and intensity of the seismic activity. These seismic phases roughly coincide with the first five of eight eruptive phases outlined by Hoblitt, Wolfe, and others (this  volume). The precursory phases are (1) Metastable--through May 31, (2) Predome--June 1 to 7, (3) Pre-explosive Buildup--June 8 to 12, (4) Long-Period Buildup--June 12 to 14, and (5) Preclimactic--June 14 to 15. Thus, the seismic phases have respective durations of a minimum of 2 months (beginning with seismic monitoring on April 5), 7 days, 5 days, 3 days, and 1 day. In the following section we describe the seismicity that typified each of these phases.

Metastable Phase: To May 31

The Metastable Phase is characterized seismically by a high but roughly constant rate of VT earthquake activity. Epicenters of most located events cluster together roughly 5 km northwest of the volcano's summit (fig. 4). During this phase the fumarolic activity on the volcano's flanks, which began with the small explosions on April 2, remained roughly constant.

Figure 4. A, Epicenter map of events located during the interval from May 6 to May 31, 1991. B, Vertical cross section through A-A' shown in A.

We do not know when seismic activity first reached a minimum level indicative of volcanic unrest. When the first portable seismographs were installed on April 5, seismic activity was already at an elevated level that would persist through the Metastable Phase (40 to 200 VT earthquakes per day). Sabit and others (this volume) note reports of felt earthquakes in the vicinity of Mount Pinatubo beginning on March 15, 1991.

Almost all of the recorded seismic activity during the Metastable Phase consisted of VT earthquakes with magnitudes ranging from less than 1 to 3. The majority of VT earthquakes at Mount Pinatubo occurred as separate events or in small flurries of events with overlapping codas lasting from 1 to a few minutes. The most prolonged flurry occurred on May 29 and lasted for approximately 20 min. The most active seismic source region during the Metastable Phase was a cluster of earthquakes located approximately 5 km northwest of the summit (fig. 4). At least 75 percent of the earthquakes occurred in that northwest cluster, over a range in depth from 1 to 10 km, with the majority of events occurring at depths of 3 to 7 km. Only a few events were detected under the energetic fumaroles that persisted after the April 2 phreatic eruptions or beneath what would eventually become the crater.

Sporadic episodes of low-level tremor with frequencies between 1 and 8 Hz were also recorded during this phase. The amplitude of these tremor episodes decreased rapidly with distance from the summit of Mount Pinatubo, suggesting a shallow source. Such tremor appears to have been associated with shallow hydrothermal activity coincident with the vigorous fumaroles observed on the north flank of Pinatubo.

A more vigorous episode of shallow tremor was associated with a small explosion that occurred between approximately 1800 on May 26 and 0200 on May 27. Several shallow LP events also occurred during this episode. Post-eruption analysis showed that the shallow tremor and LP events roughly coincided with the onset of deep LP events and that tremor from a deep source also occurred during this time (see figs. 3A and 12 of White, this volume). Deep LP events continued until 1210 on May 28. A second occurrence of deep LP events has also been identified (White, this volume). We did not recognize the deep LP events as they occurred, due to limitations in our acquisition and analysis software and in the time available to analyze the data.

Although the number of earthquakes varied somewhat about a stable average during the Metastable Phase, SO2 emissions changed drastically (fig. 5A). The emission rate of SO2 determined from COSPEC measurements increased tenfold, from 500 t/d to 5,000 t/d, between May 13 and May 28. The next measurement on May 30 indicated that SO2 had dropped abruptly to 1,400 t/d. SO2 emissions continued to drop, reaching a low of only 260 t/d on June 5 (Daag and others, this volume). We note that the shallow LP event illustrated in figure 3A occurred on May 26 during the rapid buildup in SO2 emission. This event and several other small LP events occurred during an episode of phreatic activity that lasted several hours. This activity was significantly more energetic than had been observed previously. No other shallow LP events were recorded until June 9.

Figure 5. Plot of (A) the SO2 volumes from May 10 to June 12, 1991, estimated from COSPEC measurements and (B) 4-h average RSAM values during the same time interval showing the marked division between the Metastable and Preexplosive Buildup seismic phases.

Predome Phase: June 1-7

Distinct changes in seismic activity began in early June that included an increase in the number of locatable VT earthquakes beneath the active fumaroles, an increase in small explosions, and an increase in the intensity and durations of episodes of tremor. Seismic activity during this phase eventually evolved into an intense swarm of shallow VT earthquakes that heralded the beginning of dome growth on the northwest flank of the volcano (Hoblitt, Wolfe, and others, this volume). RSAM values varied slightly around a low constant level until June 1, when they began to increase steadily and manifest larger fluctuations (fig. 5B). We use this change in RSAM values from station PIE to identify the transition from the Metastable to the Predome Phase. The Predome Phase begins at the change in the trend of RSAM averages on June 1 and ends at about 2300 on June 7, with a rapid drop in the high level of seismicity that we assume coincided with the inception of dome growth at the surface.

Small explosions and tremor episodes began to increase in number during the first few days of June. The largest explosions occurred at 1939 and 2258 on June 3. The relative amplitudes of this tremor on various stations in the network suggested it was associated with surficial hydrothermal activity. On June 4, a 22-min-long episode of tremor (fig. 3C) occurred with a dominant frequency of about 2 Hz and roughly equal amplitudes at each seismic station. At the time, we speculated that this tremor episode occurred at between 2 and 5 km in depth. Posteruption analysis, however, reveals that this tremor consists of a series of overlapping LP events and occurs at a depth of about 35 km (White, this volume).

At about 0700 on June 6, VT activity beneath the active fumaroles began to increase rapidly (fig. 6). This earthquake swarm continued to intensify until 1630 on June 7, when it evolved into an hour-long episode of sustained activity. Earthquakes in that swarm ranged in depth from -1 to 3 km. This swarm continued until 2300 on June 7, when activity abruptly declined. Map and cross-sectional views of earthquake hypocenters from June 1 through June 7 are shown in figure 7 and illustrate the concentration of seismic activity associated with the swarm beneath the summit of Mount Pinatubo. Figure 8 shows the number of earthquakes located each day in the northwest cluster and beneath the summit dome. Visual observations made early on June 8 established that a small dome had been extruded near a weak fumarole on the north-northwest flank of the volcano (Hoblitt, Wolfe, and others, this volume).

Figure 6. Drum record from station PIE showing a sample of the swarm of volcano-tectonic earthquakes associated with the dome extrusion on June 7. Time marks represent 1-min interval.

Figure 7. A, Epicenter map of events located between June 1 and June 7, 1991. B, Vertical cross section through A-A' shown in A.

Preexplosive Phase: June 8 to 0851 June 12

The Preexplosive Phase is characterized by swarms of VT earthquakes, a gradual increase in the incidence and intensity of volcanic tremor, and the reappearance of a few shallow LP events. This phase begins with a 33-hour period of relative seismic quiescence and ends with the first major explosive eruption on June 12.

The number of earthquakes detected and located in the northwest cluster continued more or less unchanged from the predome phase, at 20 to 25 per day (fig. 8). In contrast, seismicity beneath the summit decreased dramatically on June 8 and consisted only of occasional VT earthquakes of less than magnitude 1.5, 1- to 4-min episodes of tremor, and a few small explosion events that were likely produced by activity on the new dome. Visual observations suggest the dome continued to grow throughout this phase, and small amounts of ash were continuously emitted from the dome's margins (Hoblitt, Wolfe, and others, this volume). The level of seismicity beneath the new dome began to escalate again at about 0800 on June 9. Earthquakes occurred in episodic swarms lasting from 0.5 to 2.0 h separated by 2- to 4-h intervals of quiescence. Most events during this period are concentrated at shallow depth beneath the dome (fig. 9). Preliminary spectral analysis of the earthquakes recorded on June 9 suggests that some events exhibit characteristics of hybrid earthquakes, while a few are LP events. This style of earthquake activity continued until the first explosive eruption on June 12. More detailed analysis will be required to determine the relative rate of occurrence of hybrid and LP events during this period.

Figure 8. Number of located seismic events per 4-h intervals between June 1 and June 12 from (A) the entire Pinatubo network, (B) the cluster of seismic activity 5 km northwest of the summit of Mount Pinatubo, and from (C) the region beneath the summit. Events in the northwest cluster were of M„1.2; events beneath the summit were of M„0.4.

Figure 9. A, Epicenter map of events located between June 8 and June 12, 1991. B, Vertical cross section through A-A' shown in A.

Tremor episodes also began to appear on June 9 with a narrower frequency content than previously observed. The most vigorous episode began at 1030 on June 9 and lasted about 4 h. This tremor exhibited predominant frequencies in the 2- to 3-Hz range, and recorded amplitudes appeared to correlate with the vigor of the continuous, low-volume ash and steam emissions from the dome. We interpret tremor during this period as reflecting low-level ash emission rather than hydrothermal boiling that preceded extrusion of the dome. On June 9, the seismic energy release began to show bigger fluctuations than had been observed previously and its average level began to increase steadily, as reflected in the RSAM values shown in figure 5. The steady increase in RSAM values was caused by the onset of sustained low-level tremor that began at about 1300 on June 10 and continuously increased in amplitude until early on June 12.

An episode of high-amplitude tremor with a broad spectrum began at about 0310 on 12 June and lasted for about 2 h (fig. 10). An eruption event was recorded at 0341 during that tremor episode, but because of darkness no visual observations of the eruption were made. An aerial observation at approximately 0700 on June 12 confirmed that a small eruption had occurred during the night (Hoblitt, Wolfe, and others, this volume). Low amplitude tremor began again at 0841 on June 12, followed 10 min later by a rapid increase in amplitude that signaled the onset of the first explosive eruption at 0851.

Figure 10. Station PIE drum record showing (A) the eruption signals for the small eruption at 0341 and (B) the first explosive eruption at 0851 on June 12, 1991. Time marks represent 1-min intervals; the pen was manually offset following each explosion.

Long-Period Buildup Phase: 0851 June 12 to 1309 June 14

The Long-Period Buildup Phase is characterized seismically by a strong increase in the number and size of LP events, episodes of strong tremor, swarms of VT earthquakes, as well as signals from explosive eruptions. The overall level of seismic activity increased dramatically following the first large explosive eruption at 0851 on June 12.

By this time, seismic stations UBO, BUR, and PPO had been lost. Of the four remaining stations, the closest three, BUG, GRN, and PIE, recorded continuous seismic activity, and their seismic signals were often electronically off scale. The high level of seismic activity saturated the detection algorithm in the computer system, and in the absence of continuous seismic recording, most events were not recorded digitally. For the few events that were recorded, the poor geometry of the remaining four stations relative to the earthquake source region resulted in severe degradation in location quality. Seismic signals from the most distant station, CAB, were still on scale after the 0851 eruption. Thus, after that eruption, the identification of seismic events for eruption forecasting depended almost exclusively on drum records of the CAB station.

Additional explosive eruptions occurred at 2252 on June 12, 0841 on June 13, and 1309 on June 14 (fig. 11). The durations of the signals for the four eruptions during this phase are 45, 20, 8 and 4 min, respectively, as measured on the CAB drum record. All of these eruptions produced vertical ash columns; at least two of them produced minor pyroclastic flows. From the thickness of their ash-fall deposits, the successive eruptions became progressively less vigorous (Hoblitt, Wolfe, and others, this volume). The decreasing pyroclast production is consistent with the decreasing durations of their eruption signals.

Figure 11. Estimated column heights (spikes) versus time for explosive eruptions that occurred between June 12 and the onset of the climactic eruption on June 15, 1991. The cumulative RSAM values (solid line) are shown for comparison.

Following the striking increase in seismicity after the 0851 eruption, and in contrast to the decreasing vigor of the first four explosive eruptions, the rate of seismic energy release continued to increase throughout the Long-Period Buildup Phase. This increase is most evident in the RSAM data shown in figure 11. A remarkable buildup of LP events, from which this phase derives its name, evolved from 1- and 2-h swarms, respectively, of LP events preceding the explosive eruptions at 2252 on June 12 and at 0841 on June 13, to a sustained series of increasingly larger LP events that dominated seismic activity. The rapid increase in the RSAM values early on June 14 (fig. 11) shows the intense buildup of LP seismicity. The increasing size of LP events between June 12 and 14 is illustrated in figure 12 for events at station CAB. Note that the signature of the explosive eruption at 1309 in figure 12 is not significantly different from the signatures of the preceding LP events. The largest of these events were recorded on regional seismic stations at distances of up to 350 km, and preliminary estimates suggest that the largest events were roughly equivalent to earthquakes of magnitude 3.5 to 4.

Although the Long-Period Buildup Phase was characterized by the LP seismicity, an uncountable number of VT earthquakes were also recorded. Few VT earthquakes were large enough to be recorded at station CAB, and most have magnitudes between 1 and 2.

Figure 12. Sections of station CAB drum records showing the increase in size of long-period events between (A) June 12 and (B) June 14, 1991. Time marks represent 1-min intervals; pen was manually offset in B.

Preclimactic Phase: 1309 June 14 to 1342 June 15

Following the eruption at 1309 on June 14, the character of seismicity changed to a diverse mix of large VT earthquakes, and LP and eruption events. This phase coincided with a significant change in eruptive style from convectively driven vertical eruption columns at intervals of 10 to 24 h to increasingly frequent eruptions that generated pyroclastic surges. The seismic and pressure record (Hoblitt, Wolfe, and others, this volume) indicates that explosive eruptions occurred at 1410, 1516, 1853, 2218, and 2320 on June 14 and 0115, 0257, 0555, 0810, 1027, 1117, 1158, 1222, 1252, and 1315 on June 15 before the climactic eruption began at 1342.

RSAM average amplitudes show the escalation in overall seismic activity throughout the phase and indicate that the rate of activity increased at about 0300 on June 15. A portion of the CAB drum record is shown in figure 13, which illustrates the diverse seismic character during the phase. SSAM measurements indicate that seismic activity with both VT and LP spectral characteristics increased during this phase (Power and others, this volume).

Figure 13. Station CAB drum record showing diverse character of seismic activity during the Preclimactic Phase on June 14, 1991. Time marks represent 1-min intervals.

DISCUSSION

In this section we review the role that seismological observations played in formulating eruption forecasts and evaluating volcanic hazards and review the factors that influenced our interpretations at the time. The overall forecasting strategy used at Pinatubo relied on synthesizing data from a number of monitoring techniques, which included a variety of seismic measurements, visual observations, gas emissions, and geologic observations. The successful forecasts and evacuations prior to the onset of explosive activity validate this approach. Our seismic interpretations were based largely on monitoring hypocentral locations, tracking seismicity rate, and energy release, and associating the various event types with possible source processes. Recent advances in computer technology allowed us to apply a number of relatively sophisticated analysis techniques, even at the relatively remote field locations.

During the Pinatubo crisis we used a numbered alert system to communicate our interpretations and forecasts of volcanic activity to civilian and military leaders, as well as the general public. A detailed description of this system and its application throughout the 1991 eruption crisis is given by Punongbayan and others (this volume) and Wolfe (1992).

In the following subsections, we discuss our interpretations of the seismic data and their implications for evaluating magmatic processes and anticipating future eruptions prior to the climactic eruption of Mount Pinatubo at 1342 on June 15. Our discussion follows the order of successive seismic phases presented earlier.

Metastable Phase: To May 31

Our interpretation of preeruption seismicity necessarily begins when seismic stations were installed on April 5, 3 days after the initial phreatic eruptions. By that time as many as 200 VT earthquakes a day were being recorded, and the Metastable Phase was already in progress. The significance of the elevated level and widespread distribution of VT earthquake activity was the focus of our initial evaluations of eruptive potential. Experiences at other volcanoes indicate that, although a high level of VT earthquake activity is a clear indication of volcanic unrest, VT earthquakes are not always a reliable indicator of an impending volcanic eruption (Newhall and Dzurisin, 1988). For example, a marked increase in the rate of VT earthquakes preceded the 1986 eruption sequence at Augustine Volcano, Alaska (Power, 1988) and the initial eruption of Mt. Spurr, Alaska (Power and others, in press), by 9 and 10 months, respectively. However, in large volcanic systems such as Rabaul, New Britain, Papua-New Guinea (McKee and others, 1984), Campi Flegrei, Italy (Aster and others, 1992), and Long Valley, Calif. (Hill and others, 1985; Rundle and Hill, 1988), VT earthquakes have been observed to continue at elevated levels for months to years without (or, in the case of Rabaul, before) leading to eruptive activity. At each of these large complexes, evidence suggests that magma intrusions have taken place, but no eruptions have occurred.

Additionally, we were concerned that most of the seismic energy released during the Metastable Phase was associated with the cluster of hypocenters located about 5 km northwest of the summit of Mount Pinatubo (fig. 4). We considered the possibility that the earthquake activity resulted from tectonic processes that had merely perturbed the Pinatubo hydrothermal system. A rapid examination of aerial photographs indicated that the northwest cluster lay beneath an area of geologically young fault traces, and this fact raised the possibility that these earthquakes were tectonic in origin. The rationale for such a conclusion comes from a study of 35 destructive near-surface earthquakes along the volcanic chain of Central America made by White and Harlow (1993). They found that eruptive activity was associated with only two of the earthquakes, whereas the rest occurred on active faults in the vicinity of the volcanic chain and were associated with tectonic activity rather than magmatic or volcanic processes.

The COSPEC measurements that indicated that a large volume of SO2 was being emitted from the volcano left little doubt that we were dealing with an active magmatic system. The rapid increase and later decrease in SO2 emission were interpreted at the time as evidence that magma was rising and that gas escape was temporarily choked off (Daag, Tubianosa, and others, this volume).

Through the end of May, therefore, we estimated that there was very roughly a 50 percent probability that this particular episode of unrest would end in eruptive activity. This assessment was conditioned by several points: (1) the previous observations of prolonged seismic unrest exhibited by other large volcanic systems that did not evolve into eruptions (Newhall and Dzurisin, 1988), (2) the confusing situation seen at Mount Pinatubo where most of the seismic energy was being released 5 km northwest of the April 2 phreatic vent area, and (3) the large SO2 flux observed during May.

Predome Phase: June 1 to 7

Seismic activity evolved beyond the relatively steady occurrence of VT earthquakes in the early days of this interval to episodes of tremor and an intense swarm of VT earthquakes associated with a dome extrusion on June 7. The activity leading to June 7 suggested to us the increased probability of an eruption. Both the pace and the gravity of our decisions increased dramatically.

During the first days of June, the changes in seismic activity were subtle and we were unsure of their significance. The RSAM level, which began to increase on June 1 (fig. 5), is not visually obvious on the analog drum recorders and affirms the value of this straightforward tool to detect small changes in seismicity level. During our necessarily rapid assessment of seismicity, we failed to identify a tremor episode (fig. 3B) on June 4 as a sequence of LP events at depths of more than 30 km (White, this volume). We did, however, realize from a visual inspection of the drum records that this tremor episode was deeper than previous tremor episodes. That conclusion was based on the roughly equal signal amplitudes at all stations of the sequence of deep LP events compared to the rapid decrease of signal amplitude with distance for shallow tremor. Therefore, we considered this a significant change in the seismic character.

The shift in the locus of VT earthquake activity from the northwest cluster to beneath the active fumaroles also began gradually in early June. By June 6, however, that shift had become obvious with the onset of the swarm of VT earthquakes that preceded the emplacement of the dome (fig. 7). In evaluating this activity on June 6 and 7, we were cognizant of the VT nature of the swarm. Strong swarms of VT earthquakes have frequently been observed in association with shallow intrusions of magma and dome formation (Fremont and Malone, 1987, Klein and others, 1987). Measured ground deformation from tiltmeters recently installed near the volcano (Ewert and others, this volume) strongly supported this interpretation.

The fivefold jump in SO2 emission between May 24 and May 28, followed by the dramatic decrease by May 30 (fig. 5) (Daag, Tubianosa, and others, this volume), was possibly the initial step in a process that led to the extrusion of magma at the surface. Following the decrease in SO2 emissions at the end of May, VT earthquake activity began to increase beneath the fumarole vents. Hypocenters from the VT earthquake swarm on June 6 and 7 suggested that the final seismic step in the dome formation process was an episode of intense brittle fracturing in a zone extending from a depth of about 3 km to the surface. The appearance of the dome on June 7 proved that activity at Mount Pinatubo was being driven by magmatic processes.

The decision to move to a higher alert level, which declared that the volcano could erupt at any time (Punongbayan, and others, this volume), was made during the most intense phase of the swarm associated with dome formation and before the dome had been visually observed. The decision was made in spite of the absence of LP seismicity. We speculated that magma had begun ascending a conduit system formed by the phreatic activity on April 2 and that this action was responsible for the pulse of SO2 observed at the end of May. We speculated further that, as the upper tip of the magma column began to degas and cool, the system progressively sealed itself and caused the drop in SO2 emissions observed after May 28.

Preexplosive Phase: June 8 To 0851 June 12

Significant seismic changes during the Preexplosive Phase included continued swarms of VT earthquakes at shallow depths beneath the still growing lava dome, episodes of strong tremor, a gradual increase in low-level tremor that occurred throughout the phase, and the reappearance of a few shallow LP events as well as hybrid events. Interestingly the phase began with roughly 30 hours of relative seismic quiescence.

In our reflections during this period, some of us felt we had overreacted in raising the alert level on June 7, while others were confident that seismic activity would reappear and increase to higher levels than had yet been observed. This interval of seismic quiescence strongly suggested to us, however, that prior to the dome extrusion, magma movement was inducing high strain rates over a broad area. Correspondingly, the extrusion of the dome had acted as a temporary strain relief valve.

When seismic activity resumed on June 9, we observed that more seismic activity was occurring beneath the dome rather than in the now seismically subdued northwest cluster (fig. 9). This shift in seismicity also suggested that a fundamental change from the predome pattern of strain release had taken place. Continuous background tremor began on June 10, and we became increasingly concerned that an explosive eruption might be imminent, particularly when tremor increased and when we received the results of the June 10 measurement of SO2 flux, which had risen sharply to 10,000 t/d (Daag, Tubianosa, and others, this volume).

We failed to recognize that many of the events recorded after June 9 were hybrid events, nor did we fully appreciate the appearance of a few LP events, because we were fatigued and increasing demands had been made on our time by civil defense officials and base commanders as evacuation plans were made and carried out.

Between June 9 and June 12 we grew more convinced that a large explosive eruption was imminent. This confidence was based on (1) the shift in the locus of seismic energy release from the northwest cluster to depths of 1 to 3 km beneath the dome (this suggested that the magma was inducing strain over a wide area, in which case we inferred that the volume of magma involved was likely to be large), (2) the steadily increasing level of background tremor after June 9, (3) the volume of SO2 emanating from the volcano, (4) the emergence of a dome, which proved conclusively that the system was magmatic, and (5) geologic information on recent eruptive style (Newhall and others, this volume). We communicated our observations and interpretations to responsible officials frequently throughout this period. Additional communities close to the volcano, as well as nonessential personnel from Clark Air Base were evacuated on June 10, 2 days before the first explosive eruption.

Long Period Buildup Phase: 0851 June 12 to 1309 June 14

Seismic activity during this phase is characterized by increasing LP seismicity, as well as continued swarms of VT earthquakes and eruption events. Following the first large explosive eruption at 0851 on June 12, our interpretations of the Mount Pinatubo seismicity relied on the RSAM system to indicate the relative level of seismic energy release and on the waveform character of the seismic events as identified on station CAB drum records. As mentioned earlier, the abrupt increase in seismic activity observed at this time all but overwhelmed the earthquake detection algorithm, and the severely degraded accuracy and precision of calculated earthquake hypocenters resulting from the loss of three seismic stations prevented us from following the spatial development of earthquake hypocenters.

The large number of VT earthquakes following the first explosive eruption indicated that intense brittle rock fracture was occurring. These earthquakes were interpreted to be the result of readjustment of the system following the sudden evacuation of magma. At Mount St. Helens (Weaver and others, 1981), Nevado del Ruiz (Munoz and others, 1990), and at Redoubt Volcano (Power and others, 1994), significant increases in the number and distribution of VT earthquakes followed rather than preceded eruptive activity. In those cases, VT earthquakes were attributed to the readjustment of brittle rock following magma withdrawal. Similar VT earthquakes followed the June 15 eruptions of Mount Pinatubo (Mori, White, and others, this volume). That the level of VT activity remained high during this phase, however, suggests that at some point rock fracturing also may have been affected by the influx of new magma.

The dramatic buildup of LP events that dominated this phase suggested to us that the system was moving toward more energetic eruptive activity. Swarms of LP events occurred before the explosive eruptions of 2252 on June 12 and 0841 on June 13. Those swarms were minor in terms of duration and average event amplitude in comparison to the large buildup of LP events that began on June 13 and that became increasingly energetic by early June 14.

As a result of this buildup in LP seismicity, we were expecting an eruption larger than had yet occurred. We were surprised, therefore, that the eruption of 1309 on June 14 was a relatively small event compared to the previous explosive eruptions. Although the 1309 eruption was not as large as we expected, it was the last preclimatic eruption characterized mainly by vertical ash emission.

Preclimactic Phase: 1309 June 14 to 1342 June 15

After the 1309 eruption on June 14, the character of seismic activity changed from one dominated by LP events to one characterized by a diverse mix of VT earthquakes, LP events, tremor, and eruption signals (fig. 13). RSAM values show that seismic energy release continued to increase during this phase (fig. 11).

With this change in seismicity came a change in eruption style from explosive eruptions with vertical ash columns to eruptions generating large pyroclastic surges (Hoblitt, Wolfe, and others, this volume). The first large pyroclastic surge was observed from the west of the volcano at about 1510 on June 14. An infrared video system recorded two night eruptions that generated pyroclastic density currents at 2320, on June 14 and 0114, on June 15. The first observation of a large pyroclastic density current from Clark Air Base occurred later that morning during the 0555 eruption. After the 0555 eruption, clouds from the leading edge of Typhoon Yunya obscured the volcano and precluded further visual observations from Clark Air Base.

During the period between 1500, on June 14 and the onset of the climatic eruption at 1342, June 15, our interpretations were based solely on visual inspections of station CAB drum records and on the level of seismic energy release as depicted by RSAM values. At this point, as civil defense decisions had become the most critical, RSAM again proved to be an invaluable tool. Exhausted from the intense efforts needed to react to the escalating volcanic activity of the previous few days, we relied upon RSAM as an indicator of overall seismic energy release. This was all the information we had time to use and, indeed, all we could absorb. RSAM was an easy concept to grasp, and the data were automatically updated and displayed on a computer screen in an easy-to-understand graph.

The early dawn view of large pyroclastic density currents moving in all directions from Mount Pinatubo at 0555 on June 15 made us apprehensive about our safety. Once the view of the volcano was concealed, with RSAM values increasing, we decided to move to a more distant site. At 0730, our monitoring team and the remaining military personnel abandoned Clark Air Base. We reevaluated this decision a few kilometers from the base and elected to return. The team and a small military detachment returned to Clark Air Base at about at 1000. Although we could not see the volcano, eruption signals observed on the station CAB drum record were becoming more frequent, and RSAM values increased further. We retreated again at about 1500 after Mount Pinatubo had been in continuous eruption for more than an hour.

SUMMARY AND CONCLUSIONS

Our success in identifying the escalating eruptive potential between late April and mid-June can be credited to four factors: (1) our ability to deploy a seismic network around the volcano rapidly, (2) the capability to analyze seismic data from the network in near-real-time and to combine these results into geologic observations and measurements of other volcanic phenomena (3) our experience in working with seismic data from other recently active volcanoes, and (4) real-time synthesis of seismic and a number of other data sets and observations. On the basis of these forecasts, over 60,000 people were evacuated from high-risk areas surrounding the volcano, and an enormous loss of life was averted. The forecasts also prevented the loss of hundreds of millions of dollars of equipment, such as the military aircraft that were evacuated.

The pivotal changes in seismic activity that led us to believe that a large eruption was increasingly imminent included (1) a shift in seismic energy release from the northwest cluster to beneath the summit during early June, (2) a swarm of VT earthquakes associated with the extrusion of a dome, (3) increasing seismic activity beneath the summit and the appearance of continuous background tremor between June 9 and the first explosive eruption on June 12, (4) the remarkable buildup of LP events between June 12 and 14, and (5) the dramatic increase in seismic energy release as indicated by RSAM values between June 14 and the onset of the climactic eruption on June 15.

A critical context for interpreting seismic data was formed by results from geologic reconnaissance work, and measurements of SO2 flux, as well as ground-deformation measurements. Geologic information indicated that volcanism at Mount Pinatubo was distinguished by infrequent, large eruptions that produced extensive pyroclastic-flow deposits. Thus, if the seismic data indicated that an eruption was imminent, then there was a high probability that the eruption would be large and generate widespread pyroclastic flows.

The evolution of seismic activity during the Predome and Preexplosive Buildup Phases has critical implications for the early preparation of hazards maps, risk assessments, and civil defense plans at other large volcanic systems. Until June 1, seismic activity at Mount Pinatubo had been indistinguishable from that recently observed at other large volcanic complexes such as Long Valley, Calif., Rabaul, Papua New Guinea, and Campi Flegri, Italy, only one of which (Rabaul) has erupted. Two to 3 days passed before the subtle changes in the character of seismic activity that began on June 1 caught our attention, and those changes became significant enough for us to declare on June 5 that an eruption was possible within weeks. On June 7, in the midst of the swarm of earthquakes associated with the emergence of a dome, we declared that an eruption was possible within days. Thus, there was only about 1 week between the time we could state with confidence that an eruption was imminent, a confidence that wavered somewhat during the seismic quiescence on June 8, and the first explosive eruption on June 12. This short time span forcefully demonstrates that volcanic hazard studies and civil defense plans need to be completed well in advance.

During the 24 h interval preceding the onset of the climatic eruption, the value of being able to immediately display the level of seismic activity as depicted by RSAM values cannot be overstated. At a certain point, responsible officials needed to make evacuation decisions quickly, and the easily understood RSAM format proved invaluable. We strongly recommend, therefore, that computer software be developed to display clearly all data being collected in an easily understood and comparative format.

ACKNOWLEDGMENTS

We thank the many individuals who helped to make the response to the 1991 eruptions at Pinatubo a success. In particular, Ray Punongbayan, Chris Newhall, Richard Janda, C. Dan Miller, and members of the 13th Air Force. This paper benefited greatly from discussions and interactions with A. Lockhart, T. Murray, J. Ewert, P.J. Delos Reyes, B. Tubianosa, Jim Mori, Bernard Chouet, W.H.K. Lee, J. Lahr, S. Marcial, and M.L.P. Bautista. Jim Mori and Donna Eberhart-Phillips provided the hypocenter data used in figures 4, 7, 8, and 9. Bernard Chouet, E.G. Ramos, and Barry Voight provided helpful reviews of the text and figures.

References Cited

Aki, K., Fehler, M., and Das, S., 1977, Source mechanism of volcanic tremor: Fluid driven crack models and their application to the 1963 Kilauea eruption: Journal of Volcanology and Geothermal Research, v. 2, p. 259-287.

Aki, K., and Koyanagi, R., 1981, Deep volcanic tremor and magma ascent mechanism under Kilauea, Hawaii: Journal of Geophysical Research, v. 86, p. 7095-7109.

Aster, R.C., Meyer, R.P., Denatale, G., Zollo, A., Martini, M., Del Pezzo, E., Scarpa, R., and Iannaccone, G., 1992, Seismic investigation of Campi Flegrei: A summary and synthesis of results, in Gasparini, P., Scarpa, R., and Aki, K., eds., Volcanic seismology: International Association of Volcanology and Chemistry of the Earth's Interior (IAVCEI) Proceedings in Volcanology, Berlin, Springer-Verlag, p. 462-483.

Chouet, B.A., 1985, Excitation of a buried magmatic pipe: A seismic source model for volcanic tremor: Journal of Geophysical Research, v. 90, p. 1881-1893.

------1988, A seismic source model for the source of long-period events and harmonic tremor: Journal of Geophysical Research, v. 93, p 4373-4400.

------1992, A seismic model for the source of long-period events and harmonic tremor, in Gasparini, P., Scarpa, R., and Aki, K., eds., Volcanic seismology: International Association of Volcanology and Chemistry of the Earth's Interior (IAVCEI) Proceedings in Volcanology, Berlin, Springer-Verlag, p. 133-156.

Chouet, B.A., and Shaw, H.R., 1991, Fractal properties of tremor and gas piston events observed at Kilauea Volcano, Hawaii: Journal of Geophysical Research, v. 96, p. 10177-10189.

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.

Daag, A.S., Tubianosa, B.S., Newhall, C.G., Tuñgol, N.M, Javier, D., Dolan, M.T., Delos Reyes, P.J., Arboleda, R.A., Martinez, M.L., and Regalado, M.T.M., this volume, Monitoring sulfur dioxide emission at Mount Pinatubo.

Endo, E.T., Malone, S.D., Nosen, S.D., and Weaver, C.S., 1981, Locations, magnitudes, and statistics of the March 20-May 18 earthquake sequence, 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. 93-107.

Endo, E.T., and Murray, T.L., 1991, Real-time seismic amplitude measurement (RSAM), a volcano monitoring and prediction tool: Bulletin of Volcanology, v. 53, p. 533-545.

Ewert, J.W., Lockhart, A.B., Marcial, S., and Ambubuyog, G., this volume, Ground deformation prior to the 1991 eruptions of Mount Pinatubo.

Ferrazzini, V., Chouet, B.A., Fehler, M., and Aki, K., 1990, Quantitative analysis of long-period events recorded during hydrofracture experiments at Fenton Hill, New Mexico: Journal of Geophysical Research, v. 95, p. 21871-21884.

Fremont, M., and Malone, S.D., 1987, High precision relative locations of earthquakes at Mount St. Helens, Washington: Journal of Geophysical Research, v. 92, p. 10233-10236.

Hill, D.P., Bailey, R.A., and Ryall, A.S., 1985, Active tectonic and magmatic processes beneath Long Valley Caldera, Eastern California: an overview: Journal of Geophysical Research, v. 20, p. 111-120.

Hoblitt, R.P., Mori, J., and Power, J.A., this volume, Computer visualization of earthquake hypocenters.

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.

Klein, F.W., Koyanagi, R.Y., Nakata, J.S., and Tanagawa, W.R., 1987, The seismicity of Kilauea's magma system, in Decker, R.W., Wright, T.L., and Stauffer, P.H., eds., Volcanism in Hawaii: U.S. Geological Survey Professional Paper 1350, p. 1019-1186.

Koyanagi, R.Y., Chouet, B.A., and Aki, K., 1987, Origin of volcanic tremor in Hawaii, Part 1, Data from Hawaiian Volcano Observatory 1969-1985, in Decker, R.W., Wright, T.L., and Stauffer, P.H., eds., Volcanism in Hawaii: U.S. Geological Survey Professional Paper 1350, p. 11121-1259.

Lee, W.H.K., and Valdes, C.M., 1989, User manual for HYPO71PC, in Lee, W.H.K., ed., Toolbox for seismic data acquisition, processing, and analysis, IASPEI software volume 1:. Seismological Society of America, International Association of Seismology and Physics of the Earth's Interior (IASPEI), El Cerrito, Calif., p. 203-236.

Lahr, J.C., Chouet, B.A., Stephens, C.D., Power, J.A., and Page, R.A., 1994, Earthquake location and error analysis procedures for a volcanic sequence: Application to 1989-1990 eruptions: Journal of Volcanology and Geothermal Research. v. 62, p. 137-152.

Lockhart, A.B., Marcial, S., Ambubuyog, G., Laguerta, E.P., and Power, J.A., this volume, Installation, operation, and technical specifications of the first Mount Pinatubo telemetered seismic network.

Malone, S.D., Endo, E.T., Weaver, C.S., and Ramey, J.W., 1981, Seismic monitoring for eruption prediction, 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. 803-814.

McKee, C.O., Lowenstein, P.L., de Saint Ours, P., and Talai, B., 1984, Seismic and deformation crisies at Rabaul Caldera--prelude to an eruption: Bulletin of Volcanology, v. 47, no. 2, p. 397-410.

McNutt, S.R., 1992, Volcanic tremor, Encyclopedia of earth science system, Volume 4: San Diego, Calif., Academic Press, p. 417-424.

Minakami, T., 1960, Fundamental research for predicting volcanic eruptions. I--Earthquakes and crustal deformation originating from volcanic activities: Bulletin of the Earthquake Research Institute, University of Tokyo, v. 38, p. 497-544.

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

Mori, J., White, R.A., Harlow, D.H., Okubo, P., Power, J.A., Hoblitt, R.P., Laguerta, E.P.,Lanuza, L., and Bautista, B.C., this volume, Volcanic earthquakes following the 1991 climactic eruption of Mount Pinatubo, Philippines: Strong seismicity during a waning eruption.

Munoz, F.C., Nieto, A.E., Hansjurgen, M., 1990, Analysis of swarms of high frequency seismic events at Nevado del Ruiz volcano, Colombia (January 1986-August 1987): Development of a procedure: Journal of Volcanology and Geothermal Research, v. 41, p. 327-354.

Murray, T.L., 1990, A user's guide to the PC-based time-series data management and plotting program BOB: U.S. Geological Survey Open-File Report 90-56, 53 p.

Murray, T.L., Power, J.A., March, G.D., and Marso, J.N., this volume, A PC-based realtime volcano-monitoring data-acquisition and analysis system.

Newhall, C.G., Daag, A.S., Delfin, F.G., Jr., Hoblitt, R.P., McGeehin, J., Pallister, J.S., Regalado, M.T.M., Rubin, M., Tamayo, R.A., Jr., Tubianosa, B., and Umbal, J.V., this volume, Eruptive history of Mount Pinatubo.

Nieto, A.H., Brandsdottir, B., and Munoz, F.C., 1990, Seismicity associated with the reactivation of Nevado del Ruiz, Colombia, July 1985-December 1986: Journal of Volcanology and Geothermal Research, v. 41, p. 315-326.

Newhall, C.G., and Dzurisin, D., 1988, Historical unrest at large calderas of the world: U.S. Geological Survey Bulletin 1855, 1108 p.

Omori, F., 1911, The Usu-san eruption and earthquake and elevation phenomena: Bulletin of the Imperial Earthquake Investigation Committee, v. 5, no. 1, p. 1-38.

Parker, D.B., 1990, User manual for Acrospin, in Lee, W.H.K., ed., Toolbox for seismic data acquisition, processing, and analysis, IASPEI software volume 1: Seismological Society of America, International Association of Seismology and Physics of the Earths Interior (IASPEI), El Cerrito, Calif., p. 119-164.

Pinatubo Volcano Observatory Team, 1991, Lessons from a major eruption: Mount Pinatubo, Philippines: Eos, Transactions, American Geophysical Union, v. 72, p. 545-555.

Power, J.A., 1988, Seismicity associated with the 1986 eruption of Augustine Volcano, Alaska: M.Sc. thesis, University of Alaska, Fairbanks, 142 p.

Power, J.A., Jolly, A.D., Page, R.A., McNutt, S.R., in press, Seismicity and forecasting of the 1992 eruptions of Crater Peak vent, Mount Spurr volcano, Alaska: An overview, in Keith, T.E.C., ed., U.S. Geological Survey Bulletin.

Power, J.A., Lahr, J.C., Page, R.A., Chouet, B.A., Stephens, C.D., Harlow, D.H., Murray, T.L., and Davies J.N., 1994, Seismic evolution of the 1989-90 eruption sequence of Redoubt Volcano, Alaska: Journal of Volcanology and Geothermal Research. v. 62, p. 69-94.

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.

Punongbayan, R.S., Newhall, C.G., Bautista, M.L.P., Garcia, D., Harlow, D.H., Hoblitt, R.P., Sabit, J.P., and Solidum, R.U., this volume, Eruption hazard assessments and warnings.

Robinson, M.R., 1992, XPICK user's manual V4.2: Seismology lab of the Geophysical Institute, University of Alaska, Fairbanks, 119 p.

Rundle, J.B., and Hill, D.P., 1988, The geophysics of a restless caldera; Long Valley, California: Annual Review of Earth and Planetary Sciences, v. 16, p. 251-271.

Sabit, J.P., Pigtain, R.C., and de la Cruz, E.G., this volume, The west-side story: Observations of the 1991 Mount Pinatubo eruptions from the west.

Sassa, K., 1935, Volcanic micro-tremors and eruption earthquakes, (Part 1 of the Geophysical Studies on the volcano Aso): Memoirs of the College of Science, Kyoto Imperial University, Kyoto, p. 255-293.

------1936, Micro-seismic study on eruptions of the volcano Aso (Part 2 of the Geophysical studies on the volcano Aso): Memoirs of the College of Science, Kyoto Imperial University, Kyoto, p. 11-54.

Shaw, H.R., and Chouet, B.A., 1991, Fractal hierarchies of magma transport in Hawaii and critical self organization of tremor: Journal of Geophysical Research, v. 96, p. 10191-10207.

Stephens, C.D., Chouet, B.A., Page, R.A., Lahr, J.C., and Power, J.A., 1994, Seismological aspects of the 1989-1990 eruptions at Redoubt Volcano, Alaska: The SSAM perspective: Journal Volcanology and Geothermal Research. v. 62, p. 153-182.

Valdez, C.M., 1989, User manual for PCEQ, in Lee, W.H.K., ed., Toolbox for seismic data acquisition, processing, and analysis, IASPEI software volume 1: Seismological Society of America, International Association of Seismology and Physics of the Earths Interior (IASPEI), El Cerrito, Calif., p. 175-202.

Weaver, C.S., Grant, W. C., Malone, S.D., and Endo, E.T., 1981, Locations, magnitudes, and statistics of the March 20-May 18 earthquake sequence, 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. 109-121.

White, R.A., this volume, Precursory deep long-period earthquakes at Mount Pinatubo, Philippines: Spatio-temporal link to a basalt trigger.

White, R.A., and Harlow, D.H., 1993, Destructive upper-crustal earthquakes of Central America since 1900: Bulletin of the Seismological Society of America, v. 83, p. 1115-1142.

Wolfe, E.W., 1992, The 1991 eruption of Mount Pinatubo, Philippines: Earthquakes and Volcanoes, v. 23, no. 1, p. 5-38.

Wood, H.O., 1915, The seismic prelude to the 1914 eruption of Mauna Loa: Bulletin of the Seismological Society of America, v. 5, p. 39-50.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey
This page is <https://pubs.usgs.gov/pinatubo/harlow/>
Contact: Chris Newhall
Last updated 06.10.99