Activity continued to decline through 15 September, with only three ash/steam emissions since about 25 August. Heavy monsoon rains triggered numerous mudflows and secondary explosions from the 15-16 June pyroclastic-flow deposits. Two large secondary pyroclastic flows occurred, producing associated ash clouds to 15 km height. The press reported continued fatalities from debris/mudflows and disease in evacuation camps, bringing the number of casualties attributed to the eruption to at least 740 by 20 September. Study of the June deposits has resulted in preliminary estimates of 7-11 km³ of material erupted.

5-11 August. Radar at Clark Air Base detected 13 ash/steam emissions rising to 4.5-13.5 km height; plumes were carried NE by the wind. Most RSAM peaks coincided with these emissions. The majority of seismicity was shallow (<=1 km depth), with magnitudes <1. Seven high-frequency earthquakes were felt at Clark Air Base.

12-18 August. Thirteen ash/steam emissions were detected, three with columns >15 km high (maximum 17.5 km). Wind carried the plumes ENE and NE, and ashfall was reported at Clark Air Base on 13 and 16 August. Ejection velocities ranged from about 300-900 m/min, similar to the ejection velocity on 25 June (estimated at about 450 m/min). A large secondary pyroclastic flow occurred sometime on 12-13 August, in the Marunot drainage on the NW flank. The flow was not observed, but satellite imagery was used to identify the deposits and estimate a deposit volume of 31 x 10⁶ m³ (1.25 km² areal coverage). The flow, ~10 km long, created a headwall scarp about 20 m high along a 240° arc in the primary pyroclastic-flow deposit source region. During aerial observations, the still-steaming secondary deposits could be differentiated from those of earlier pyroclastic flows by the absence of rills and dissected morphology.

Seismic energy release decreased notably from the previous week (figure 19), although the number of earthquakes remained about the same (102 recorded events/day compared to 95/day the week before). Several shocks were felt at Clark Air Base. RSAM peaks reflected high-frequency earthquakes generated by mudflows, and occasional long-period signals associated with ash/steam emissions from the caldera. Geologists suggested that small long-period events may also be related to secondary explosions from pyroclastic-flow deposits.

Figure 19. Accumulated RSAM energy at Pinatubo, 28 July-18 August 1991. Courtesy of PHIVOLCS.

19-25 August. Ash/steam emissions averaging ~9.8 km high (maximum 15 km) were detected eight times during the week. Ash was carried E. Some may have originated from secondary explosions at the E flank (Sacobia valley) pyroclastic-flow deposits. Seismicity consisted mostly of high-frequency earthquakes (M < 1.0) centered below the caldera or ~3 km NW, at 0-18 km depths (figure 20). Four events (M 2-4) were felt at Clark Air Base, with intensities to IV (adapted Rossi-Forel scale). RSAM peaks coincided with the larger high-frequency earthquakes, and long-period events were associated with ash/steam emissions.

Figure 20. Epicenters of 648 earthquakes recorded near Pinatubo, 19-25 August 1991. Courtesy of PHIVOLCS.

26 August-1 September. Only two ash/steam emissions were detected; plume heights ranged from about 10 to 16 km. Light ashfall occurred to 40 km SE (San Fernando) during secondary explosions that produced columns to 16 km. Ash related to these events caused poor visibility (300 m) on the highway between San Fernando and Angeles (25 km E of the volcano). The number of felt shocks (M < 4.2) increased to 17, with intensities to V (adapted Rossi-Forel scale). Multiple peaks in RSAM plots were due to mudflows, while single peaks were caused by long-period events associated with the two ash/steam emissions.

2-8 September. One ash/steam emission was detected (2 September), producing a 9-km plume that was carried W (highest portion) and NE (lower portion). Secondary explosions, three of which were recorded as low-amplitude, low-frequency earthquakes, generated ash clouds 2-4.5 km high. Geologists proposed that the heavy ashfall and 15-km-high ash column observed at 1400 on 4 September (figure 21) were from a secondary pyroclastic flow, whose fresh deposits were discovered two days later. The absence of a long-period earthquake coincident with the ash cloud suggested that it had not been generated by caldera explosions. The secondary pyroclastic-flow deposits about 3 km SSW of the caldera (in the upper Marella drainage) were estimated to be 1-2 km wide, and 4-6 km long, with a headwall scarp 15-25 m high. The deposit appeared very recent and seemed water-saturated. It was not known whether the ash cloud was generated purely by convection, or by phreatic explosions resulting from an encounter with water on the river bed. A helicopter overflight of the caldera on 6 September revealed no evidence of activity during the prior several days. Steaming was observed along the margins of the caldera and a bluish lake was present. No evidence of a lava dome was found.

Figure 21. Visible and infrared image from the NOAA 11 polar-orbiting weather satellite on 4 September at about 1445, showing a large, 15-km-high ash cloud above Pinatubo believed to have been generated by a secondary pyroclastic flow. Courtesy of G. Stephens.

Recorded earthquakes averaged 88/day, similar to 89/day the previous week. The majority were of high-frequency, and geologists believed that they were caused by tectonic readjustments. Most of the few low-frequency signals coincided with observed secondary explosions. Seismicity remained shallow (about 38% at less than 2 km depth), centered beneath or NW of the caldera. Long-duration, high-frequency earthquakes corresponding to mudflows created peaks in RSAM plots. A magnitude 5.1 earthquake at 0627 on 5 September, centered ~17 km NNW (15.53°N, 120.31°E) at 10 km depth, was felt at Clark Air Base (intensity RF V).

On 4 September, due to the continued decrease in caldera activity, the volcanic alert was reduced from Level 5 (eruption in progress) to Level 3 (numerous magma-related earthquakes, fumaroles, and gas emission), and the danger zone radius was reduced from 20 to 10 km. The principal remaining hazards and their probable durations were identified (table 4).

Table 4. Principal hazards associated with Pinatubo following the 15-16 June 1991 eruption (as of 4 September 1991). Courtesy of PHIVOLCS.

9-15 September. Although no ash/steam emissions were detected, ash clouds 2-10 km high were produced by secondary explosions. Vigorous steam emission was noted from the S side of the caldera, and the blue crater lake was still present during observations on 10 September. The average number of earthquakes decreased to 54 recorded daily, most centered ~2 km NW or 2 km S of the caldera, at <2 km and 5-10 km depths. The majority of events were M <2. RSAM and accumulated energy both showed decreases corresponding to the drop in seismicity. Multiple RSAM peaks coincided with mudflows, while single peaks were caused by moderate-sized earthquakes.

Debris flows. All of Pinatubo's major drainage systems experienced debris flows, ranging from mudflows to hyperconcentrated flows and floods. Numerous flows also occurred in more distant drainages in which significant quantities of tephra were deposited. To help alleviate hazards and to aid in studying debris-flow processes, rain gauges were installed, observation posts were set up at strategic locations along rivers, and cross sections were monitored at bridges. Timely warnings and evacuations considerably reduced the number of injuries and casualties. High rainfall (to > 30 cm/day) and still-hot pyroclastic-flow deposits generated numerous hot mudflows that deposited as much as several meters of material.

On the SE flank's Pasig-Potrero River, pyroclastic-flow deposits had formed a dam behind which a 1,000 x 600 m lake had formed. The lake drained on 7 September, causing muddy flash floods that reached 1.2 m high in about 5-10 minutes at Bacolar (35 km SE of the volcano). Press reports indicated that 800 homes were destroyed and seven people were confirmed dead. By 15 September, continued flooding and mudflows resulted in the deaths of 12 more people at Bacolar, where 45,000 of the 68,000 residents had fled.

News reports placed the death toll from the eruption, mud flows, and disease at more than 740 by 20 September [see also 16:9]. Of the fatalities in evacuation camps, an estimated 95% were Aeta tribesmen and 75% were children. The Aeta reportedly refused most medical assistance such as vaccinations.

Fieldwork on June eruptive products. Preliminary estimates have been calculated for pyroclastic-flow deposits and airfall tephra from the paroxysmal June eruptive activity. The bulk of the material erupted was found in pyroclastic flow deposits (5-7 km³); several drainage systems included more than 1 km³. An estimated 0.48 km³ of airfall tephra was deposited within the 15-cm isopach (table 5); the total volume of tephra-fall material erupted, including that deposited in the South China Sea or lost to the atmosphere, was believed to be between 2 and 4 km³. The total volume, therefore, is estimated as 7-11 km³ (roughly 3-5 km³ dense rock equivalent).

Table 5. Preliminary volume calculations (±10% error) of June 1991 eruptive products from Pinatubo. Total tephra deposit volume: 0.48 km³. Total pyroclastic-flow volume: 7.0 km³. Courtesy of PHIVOLCS.

Geological Summary. Prior to 1991 Pinatubo volcano was a relatively unknown, heavily forested lava dome complex located 100 km NW of Manila with no records of historical eruptions. The 1991 eruption, one of the world's largest of the 20th century, ejected massive amounts of tephra and produced voluminous pyroclastic flows, forming a small, 2.5-km-wide summit caldera whose floor is now covered by a lake. Caldera formation lowered the height of the summit by more than 300 m. Although the eruption caused hundreds of fatalities and major damage with severe social and economic impact, successful monitoring efforts greatly reduced the number of fatalities. Widespread lahars that redistributed products of the 1991 eruption have continued to cause severe disruption. Previous major eruptive periods, interrupted by lengthy quiescent periods, have produced pyroclastic flows and lahars that were even more extensive than in 1991.

Information Contacts: R. Punongbayan, PHIVOLCS; K. Rodolfo, Pinatubo Lahar Hazards Taskforce, Univ of Illinois; W. Scott, USGS CVO; G. Stephens, NOAA/NESDIS; NEIC; AP; Reuters; UPI.