FIRE and MUD Contents

Secondary Pyroclastic Flows from the June 15, 1991, Ignimbrite of Mount Pinatubo

By Ronnie C. Torres,1 2 Stephen Self,2 and Ma. Mylene L. Martinez1

1 Philippine Institute of Volcanology and Seismology.

2 Hawaii Center for Volcanology, Dept. of Geology and Geophysics, School of Ocean and Earth Science and Technology (SOEST), University of Hawaii, 2525 Correa Rd., Honolulu, HI 96822.


Secondary pyroclastic flows, some leaving deposits up to 10 kilometers long, 1 kilometer wide, and 10 meters thick, occurred at Pinatubo for more than 2 years after the June 15, 1991, eruption. Avalanching of the valley-ponded facies of the 1991 ignimbrite generates these secondary pyroclastic flows and leaves an avalanche scarp at their origin. Material is transported in hot, high-concentration laminar flow over a gentle slope, driven by gravity and gas fluidization, and is deposited as massive, valley-filling ignimbrite. Secondary pyroclastic flows are very similar to vent-derived pyroclastic flows and their deposits are virtually indistinguishable from primary ignimbrite in the field. However, in deposits so far studied, secondary ignimbrite tends to be fines depleted and its coarse clasts have non-uniform thermoremanent magnetic polarity. Secondary pyroclastic flows differ from other secondary movements in ignimbrite noted by previous workers, principally in that they occur long after the deposition of primary ignimbrite. Such flows, apparently a common mechanism for redistributing ignimbrite after initial deposition, present posteruption hazards that were previously unrecognized.

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.


Explosive activity of Mount Pinatubo climaxed on June 15, 1991, when thick, nonwelded ignimbrite was emplaced in all major river valleys and on low-lying pyroclastic fans around the volcano. The eruption coincided with the onset of the rainy season, and new drainage systems were rapidly reestablished within ignimbrite fans, preferentially over former river axes. Intense and prolonged rainfall after the climactic eruption generated hot lahars, and, in some instances, avalanches of hot primary ignimbrite. Several of the latter events involved collapse of large masses of primary ignimbrite that, after transport as secondary pyroclastic flow, were deposited as secondary ignimbrites. We use the term "secondary pyroclastic flow" for the moving system and "secondary ignimbrite" for the deposit.

Secondary pyroclastic flows were first observed on August 12-13, 1991, and more have been documented over the succeeding 2 years, up to the time of this writing (August 1993).

This paper describes some of the largest secondary pyroclastic flows, remobilized from thick, nonwelded, vent-derived ignimbrite and transported as dry, gas-fluidized flows. We also describe their deposits, contrasting them with the primary ignimbrite at Mount Pinatubo. Deposits described herein are part of a dynamic system involving large ignimbrite sheets and lahar conveyance channels. The exact units and the outcrops we describe will most likely not remain exposed for long and will be eroded or buried by new events.

In effect, ignimbrite emplacement is a continuum that begins with the vent-derived pyroclastic flow and continues during remobilization as secondary and tertiary pyroclastic flows long after the initial deposition. The second and third orders in this ignimbrite continuum have been overlooked in previous studies of ignimbrite sheets because the deposits are so alike. However, secondary and even tertiary pyroclastic flows are hazardous, and we hope that this paper will alert future workers to this hazard.


Secondary mass flow (Chapin and Lowell, 1979; Ellwood, 1982) and secondary flow structures (Wolff and Wright, 1981) have been recognized in high-temperature rheomorphic ignimbrites and other welded tuffs. Such flow is thought to occur during or shortly after emplacement, while the material is still a coherent viscous fluid (Schmincke and Swanson, 1967; Ellwood, 1982; Branney and Kokelaar, 1992), and permit, for instance, slumping down the oversteepened sides of paleovalleys toward the valley axis (Chapin and Lowell, 1979). Fierstein and Hildreth (1992) recognized segregation (pseudoflow) units at the Valley of Ten Thousand Smokes that they attribute to internal shear at the margin and termini of the main ignimbrite during the last stages of deposition.

A different kind of secondary flow was inferred at Mount St. Helens, where parts of the main blast flow condensed into secondary pyroclastic flows down South Coldwater Creek, and were emplaced contemporaneously with blast deposits (Hoblitt and others, 1981; Fisher and others, 1987). However, because blast material might not have come to rest before evolving into pyroclastic flows, Walker and others (in press) propose that flows like those in the South Coldwater valley be called drain-down deposits and that the term secondary pyroclastic flows be reserved for phenomena such as those observed at Pinatubo, which occur well after deposition of the primary ignimbrite.

Still another kind of secondary, pumice-rich flow deposit occurs in the Valley of Ten Thousand Smokes, Alaska. It strongly resembles the primary ignimbrite except for a lack of physical evidence for hot emplacement and a ratio of dacite to rhyolite that suggests remobilization of tephra rather than of primary ignimbrite (Hildreth, 1983).


By far the most voluminous pyroclastic-flow deposits of the 1991 eruptions are pumice-rich ignimbrites emplaced during the climactic phase on June 15. These are mostly massive, valley-confined, and reach 200 m in thickness in places; 40 to 80 m in thickness is more common. The bulk of the volume was deposited within 3 to 9 km from the vent, but some deposits extend 16 km from the vent.

The 1991 primary ignimbrite at Mount Pinatubo infilled and buried the deeply incised slopes and all preexisting major river valleys: Sacobia, Pasig-Potrero, Marella, Balin Baquero, Maraunot, Bucao, and O'Donnell (fig. 1; see also W.E. Scott and others, this volume). Peaks of former highlands were left protruding out of the ignimbrite fans like isolated islands. Small deposits of ignimbrite were also emplaced within the Bangat and Gumain River valleys.

Figure 1. Distribution of primary and secondary ignimbrites after the 1991 eruption of Mount Pinatubo. River valleys affected by major secondary pyroclastic flows are (1) north Balin-Baquero, (2) Maraunot, (3) Marella, (4) Sacobia, (5) Pasig-Potrero, (6) south Balin-Baquero, and (7) upper Bucao. Secondary ignimbrites are shown as densely dotted areas. (Distribution of primary ignimbrite is by R.S. Punongbayan, in PHIVOLCS-NEDA (1992)).

The fresh ignimbrite is typically hot, pumiceous, fines rich, unconsolidated, and extremely unstable along oversteepened sides, where small avalanches commonly occur. Deposits contain >60% pumice clasts in the -2 phi to -4 phi size range, and the remainder are dense juvenile and accidental lithic clasts. The pumice component includes coarsely vesiculated, phenocryst-rich and finely vesiculated, phenocryst-poor types. Pumice-rich ignimbrites are moderately sorted to poorly sorted (sigmaphi = 1.5 to 3.5) and have a grain-size distribution dominated by the finer fraction (fig. 2). Median grain sizes range from 0 to 2.5 phi (fig. 3).

Figure 2. Grain-size distribution of the primary ignimbrite. Samples consist of pumice-rich ignimbrite at (A) Sacobia, (B) Pasig-Potrero, (C) Marella, and (D) Maraunot.

Figure 3. Median grain diameter (Mdphi) and sorting (sigmaphi) of the primary ignimbrite (squares with dots) and secondary ignimbrites (solid triangles). Undotted squares are primary ignimbrite of W.E. Scott and others (this volume). Ignimbrite field (broken lines) after Sparks (1976) is delineated for comparison.

Within the pumice-rich ignimbrite, a lithic-rich facies occurs as pods, flow-segregated layers, and distinct flow units (W.E. Scott and others, this volume). The lithic clasts, many of which bear hydrothermal alteration, are dominantly dacitic, derived from pre-1991 lava domes and lava flows of Mount Pinatubo. Although some dense clasts occur in secondary ignimbrites, pumice-rich ignimbrites are the source of virtually all of the remobilized material discussed here.

Numerous fumaroles and phreatic explosion craters formed on the surface of the 1991 ignimbrite. Secondary explosions and hot avalanches persist to this time of writing (August 1993), although they significantly decreased within a year after the eruption. Fumarolic activity and secondary explosions were most intense along axes of former river valleys, where the ignimbrite is also thickest. A year and a half after the eruption, temperatures of fumaroles in primary ignimbrite of the Marella River valley were as high as 390°C at 1 m in depth. Actual emplacement temperature must have been higher. Juvenile pumice clasts exhibit uniform both normal and reversed thermoremanent magnetic (TRM) polarity (fig. 4 and C.G. Newhall, written commun., 1991).

Figure 4. Thermoremanent magnetic (TRM) polarity distribution of clasts of primary and secondary ignimbrites and lahars at Mount Pinatubo. Most clasts used in TRM polarity determination consist of pumice, except a set of samples from the southern Balin-Baquero watershed, which are dense juvenile fragments. Primary ignimbrite exhibits both normal and reversed TRM polarity. Secondary ignimbrite and lahars generally display a non-uniform distribution. A secondary ignimbrite in the middle Sacobia valley shows a distinct reversed mode. Circle radius represents five samples.


Collapses of primary ignimbrite, generating secondary pyroclastic flows, are contemporaneous with lahar-forming events. Avalanches originate in the medial and distal parts of valley-filling primary ignimbrite and travel along active lahar channels. All large secondary pyroclastic flows have an associated headscarp region in which interconnected or en-echelon scarps suggest multiple avalanche failures of the primary ignimbrite. Details of some individual events are given below.


The first known occurrence of a secondary pyroclastic flow was August 12-13, 1991, following days of heavy rains and lahars events. A 20-m-thick mass, occupying 1.25 to 2.0 km2 of the medial part of the ignimbrite fan at the former Maraunot River valley, was remobilized and transported about 10 km downslope (Pinatubo Volcano Observatory Team, 1991; Smithsonian Institution, 1991). The redeposited material could be distinguished from the surrounding primary ignimbrite by its relatively undissected surface and lighter tone, which indicated new deposition of pyroclastic materials. This deposit could be traced upslope until it terminated in a headscarp. The development of a deep headscarp and long scar paved the way for the establishment of a new river course that now feeds the Bucao drainage system. The volume of this secondary ignimbrite is estimated to be 0.04 to 0.05 km3 on the basis of the calculated volume loss at the headscarp. However, some of the volume lost at the headscarp may have been transported earlier as lahars.

During August 12-13, a ground-based weather radar at Cubi Point detected five ash clouds that reached 10 to 15 km in altitude (Smithsonian Institution, 1991). Correlative peaks appear in Realtime Seismic Amplitude Measurements (RSAM) at 0106, 0338, 0537, 1410, and 2250 on August 13, 1991 (Philippine Institute of Volcanology and Seismology (PHIVOLCS) Pinatubo Volcano Daily Update, August 14, 1991). These high ash clouds were assumed to have come from the crater, but, in retrospect, it is quite possible that they were generated by secondary explosions and (or) progressive avalanche failures in ignimbrite of the Maraunot River valley. Three felt earthquakes were recorded on this day and one of these, felt at Poonbato (on the Bucao River) with intensity I on the Rossi-Forel scale, coincided with the ash cloud at 0537.


Another secondary pyroclastic flow was generated on September 4, 1991, on the Marella River ignimbrite fan (fig. 5). The upper 20 to 25 meters of 100- to 200-m thick ignimbrite broke loose and flowed about 5 km (Pinatubo Volcano Observatory Team, 1991). The flow event was not actually observed, but the timing was reckoned from the conspicuous buildup of a 15-km-high ash plume at about 1400 (PHIVOLCS Pinatubo Volcano Daily Update, September 5, 1991), which was also documented by the NOAA-11 polar-orbiting weather satellite (Smithsonian Institution, 1991). The headscarp and a fresh deposit of pyroclastic materials were found 2 days after the event. Neither the seismicity nor aerial investigation of the crater gave any indication of a renewed magmatic eruption, so the ash plume was not vent-derived. In fact, the eruption alert level was lowered by PHIVOLCS on September 4 from Level 5 (eruption in progress) to Level 3 with a corresponding reduction of the boundary of the danger zone from 20 to 10 km in radius from the vent (PHIVOLCS Pinatubo Volcano Daily Update, September 5, 1991). However, the secondary pyroclastic flows traveled beyond the prescribed 10-km danger zone. The ash cloud and already overcast weather in the Pinatubo area caused darkness over Clark Air Base and vicinity for about 3 h. Light to moderate ash fall affected the towns to the east and northeast of the Pinatubo caldera.

Figure 5. Oblique aerial photograph of the headscarp area of the Marella secondary pyroclastic flow. Sketch made from the above photograph shows the configuration of the headscarp and the distribution of the pre-1991 topography (old topography, OT), 1991 primary ignimbrite (pyroclastic flow, Pf), and lahar (Lh). Arrow indicates downstream direction. Small patches of proximal facies or secodary ignimbrite are not mappable at this scale. Medial and distal ignimbrite were eroded or buried by subsequent lahar events.


A secondary pyroclastic flow was directly observed on April 4, 1992, in the Sacobia River valley. The event was probably initiated at 1514, during which time a secondary explosion and (or) avalanche triggered an ash column at least 1 to 2 km high (PHIVOLCS Pinatubo Volcano Daily Update, April 5, 1992). Eyewitnesses in the Sacobia River valley, upstream from Clark Air Base, described the secondary pyroclastic flow as a valley-confined "boiling ash flow."

This secondary pyroclastic flow originated just above the divergence of the Sacobia and Abacan valleys. Flows traveled down both valleys, 4 km down the Sacobia and about 3 km down the Abacan, and caused about 5 m of aggradation along both rivers (figs. 6 and 7). Several 5- to 7-m high dams that had been built to trap sediment and slow lahars along the Abacan and Sacobia Rivers were completely buried by the secondary ignimbrite. The vegetation on both sides of the channel was singed or smothered by wet fallout. A temperature of 240°C was measured in an exposed section 3 months after emplacement. Fumaroles on the secondary ignimbrite surface steamed for several months.

Figure 6. Photographs taken before (above) and a week after (below) the April 4, 1992, secondary pyroclastic-flow event in the Sacobia river valley, which caused channel aggradation of 5 to 7 m. The secondary pyroclastic flow buried the earlier lahar deposits. This is just upstream from the terminus of 1991 primary pyroclastic flows in the Sacobia valley. The channel walls consist of deposits of lahars and pyroclastic flows from several prehistoric eruptive episodes.

Figure 7. A, Location of the Sacobia, Abacan, and Pasig-Potrero River valleys. B, Primary and secondary ignimbrites of the Sacobia and Abacan Rivers. A 10- to 15-m-thick block of primary ignimbrite was remobilized as secondary pyroclastic flow into both the Sacobia and Abacan Rivers. The degradation at the headscarp area caused the Abacan River to be cut off from the upper drainage system it shared with the Sacobia River. C, Secondary pyroclastic flow that travelled 5 to 6 km from the source headscarp along the main channel of the Pasig-Potrero River. A fines-rich facies spilled from a notch in the escarpment. This secondary pyroclastic flow also dammed a tributary to the main channel, which broke out 6 weeks later and caused devastating lahars in the Mancatian area.

Associated ash and muddy rain fell heavily at 1545 and reduced the visibility over Clark Air Base from about 1600 to 2030. Wetted ash fall caked on the valley walls along paths of the secondary pyroclastic flow. A 2- to 3-cm-thick laminated and accretionary lapilli-rich fallout deposit draped the topography surrounding the headscarp area at Sacobia and Abacan Rivers; correlative fallout thins to <1 cm near the terminus of the secondary ignimbrite along the Sacobia channel at 3 km from the source.


A secondary ignimbrite was emplaced in the Pasig-Potrero River valley on the late afternoon and evening of July 13, 1992 (fig. 8). An ash column at least 4 km high was observed from Clark Air Base and from Barangay Mancatian (fig. 7B), along the Pasig-Potrero River, and light ash fell over Clark Air Base and adjacent areas from 1830 to 1930 (PHIVOLCS Pinatubo Volcano Daily Update, July 14, 1992).

Figure 8. Photographs of the main channel of the Pasig-Potrero River taken a month before (above) and 5 days after (below) the July 14, 1992, secondary pyroclastic-flow event, near the confluence of the main channel and the fines-rich arm of that flow (fig. 7C). The secondary pyroclastic flow caused about 10 to 12 m of aggradation in the main channel. Channel walls consist of old columnarly jointed ignimbrites. Recent lahars form the terraced deposits in the channel.

On July 15, we observed a new secondary ignimbrite that was traceable back to a 0.5-km-wide and 1.0-km-long escarpment in primary ignimbrite of the Pasig-Potrero watershed (fig. 7C). About 0.010 to 0.015 km3 of primary ignimbrite was remobilized and transported 5 to 6 km downriver. The bulk of the flow, rafting abundant pumice, went down the main valley of the Pasig-Potrero River. A fines-rich facies spilled from the rim of the headscarp, into a northern tributary of the Pasig Potrero River, and joined the main flow at a confluence 4 km from the source. The fines-rich facies covered the pumice-rich main-channel facies from their confluence point to their full runout distance. Near the foot of Mount Cutuno (fig. 7A), the main mass of secondary ignimbrite blocked a tributary to the main Pasig-Potrero River, and created a small impoundment of water.

The secondary ignimbrite contains partly to completely charred wood chunks. Temperature measurements of the secondary ignimbrite taken 5 days after emplacement ranged from 240 to 260°C at 1 m in depth.


Secondary pyroclastic flows were unknown hazards at Mount Pinatubo before the eruption. Because there was no specific effort to monitor their occurrence, it is very likely that some events, particularly those of late June and July 1991, which occurred at night or during inclement weather, escaped observation. In many cases, secondary pyroclastic flows were inferred after a new avalanche scar, an undissected pumice plain, or a fresh fallout deposit was identified by aerial or ground surveys. However, field inspection of ignimbrite fans was not possible on a regular basis, especially during the first 2 months after emplacement of primary ignimbrite. Sporadic ash venting, overcast or stormy weather, frequent phreatic explosions, and widespread fumarolic activity made the ignimbrite fans largely inaccessible even by helicopter, as most pilots refused to operate under these conditions. Consequently, the timing of some secondary pyroclastic-flow events cannot be established.

One large inferred event left an escarpment in the northern portion of the ignimbrite fan of the Balin Baquero watershed, about 1 to 2 km southwest of the Maraunot headscarp, that was noted from posteruption radar imagery flown in November 1991 (see fig. 4B of Newhall and others, this volume). Because the headscarp and valley morphology are very similar to those created by secondary pyroclastic flows in the Maraunot and Marella River watersheds, we believe that the same process operated in the northern Balin Baquero watershed. Although it is difficult to fix a specific date for the secondary pyroclastic flow in the northern Balin Baquero, greater incision of its headscarp and valley floor hints that it occurred earlier than the nearby, August 1991 Maraunot secondary pyroclastic flow.

Two additional secondary ignimbrites--one in the southern part of the Balin Baquero watershed and another in the upper Bucao watershed--were identified during an aerial survey a few days after heavy rainfall on September 20-21, 1992. At 0957 on September 21, an approximately 18-km-high ash plume was observed by the weather radar station at Cubi Point Naval Air Station, south of Pinatubo. Fallout from this ash cloud affected wide areas, including Manila, 90 km southeast of Pinatubo. The southern Balin Baquero and upper Bucao secondary pyroclastic flows each remobilized 0.035 to 0.045 km3 of primary ignimbrite. It is not certain whether these two secondary pyroclastic flows are coeval or not. No other distinct buildup of ash clouds was noted between July 13, 1992, and the time these deposits were found in late September 1992.

Smaller secondary pyroclastic flows with runout distances of about 1 km were also noted during fieldwork in the middle Sacobia valley in November 1992 and in the Balin Baquero and Marella valleys during fieldwork in July 1993. These smaller secondary pyroclastic flows are surely common, but their deposits are quite ephemeral, as lahars rework or bury them within a short time.


Secondary ignimbrites of Pinatubo are distributed along major river valleys that drain from the 1991 pyroclastic fans, namely the Sacobia, Pasig-Potrero, Marella, Balin Baquero, Maraunot and Bucao Rivers. The largest secondary pyroclastic flows occurred along the western flanks of Mount Pinatubo in the Marella, Balin Baquero, and Maraunot River valleys, which also received the most extensive and thickest primary ignimbrite. These large secondary ignimbrites have volumes in the range 0.01 to 0.05 km3; smaller secondary ignimbrites also occur, but, as noted above, are easily lost from the geologic record. Deposit characteristics are similar regardless of size.

Because secondary pyroclastic flows at Pinatubo followed valleys, it is suggested that they were dense flows. In at least one instance, along the Pasig-Potrero River, a dilute, fines-rich facies overflowed banks near the source. The flow did not erode substrate over which it flowed. Compared to other pyroclastic flows of comparable volume, Pinatubo secondary pyroclastic flows had a low ratio of vertical drop (H) to runout distance (L) (fig. 9, after Hayashi and Self, 1992), which indicates unusual mobility.

Figure 9. Log of H/L versus log of volume (in cubic kilometers) of secondary pyroclastic flows. H/L is the ratio of the vertical drop (H) to runout distance (L) of the secondary pyroclastic flows. Fields of pyroclastic flows, volcanic debris avalanches, and nonvolcanic debris avalanches after Hayashi and Self (1992).

The main body of the secondary ignimbrite is dry, massive, loosely consolidated, and poorly sorted (sigmaphi = 2.0-3.0). Grain-size distribution of the secondary ignimbrite (fig. 10) is very similar to that of the primary ignimbrite (fig. 2; comparison in fig. 12). Calculated grain-size parameters, such as median grain diameter and sorting, plot within the field of the pumice-rich primary ignimbrite at Mount Pinatubo. A major mode appears in the 1- to 2-phi (crystal-dominated) size range. Internally, the secondary pyroclastic-flow deposits strongly resemble a massive primary ignimbrite. A section of a proximal facies of secondary ignimbrite (fig. 11) exhibits inverse grading, gas escape pipes, surface fumaroles, and a dense lithic-rich basal layer. Medial and distal facies are generally massive and have a poorly defined basal layer (layer 2a of Sparks, 1976) and dense lithic segregations. Incorporated accretionary lapilli remained as coherent particles near the base. Secondary deposits lack the characteristic cohesiveness of laharic debris-flow deposits found in the same area.

Figure 10. Grain-size distribution of secondary ignimbrites in the Sacobia River (A and B), Pasig-Potrero River (C and D), and Marella River (E and F).

Figure 11. Vertical section of secondary ignimbrite and underlying lithic-rich primary ignimbrite in the Marella River valley. Sketch is made from the above photograph. The section is close to the valley wall and the source escarpment. The secondary ignimbrite is deposited on an erosional surface cut into ash fall and reworked deposits overlying a fines-depleted facies of the primary ignimbrite. The massive section of secondary ignimbrite also shows a basal layer and fumarolic pipes. When the photograph was taken (November 1992), the underlying primary ignimbrite displayed active fumaroles that deposited yellow sublimates on the wall of the section. About 10 to 20 cm of ash fall and reworked deposits cover the section and drape the surrounding topography. Upright scale is 1 m long.

Clasts in the secondary ignimbrite consist dominantly of pumice, which is strongly concentrated in the upper part of the flow. Charred wood debris on the surface was oriented perpendicular to the flow direction along the flow axis but subparallel along the margin. The "frozen" fabric of the rafted pumice and floating wood debris indicates that the maximum flow velocity was along the axis of flow and decreased toward the margin. Insignificant overbanking around bends of the channel (as deduced from absence of swash marks and nearly horizontal cross-sectional profiles) suggests a very low flow velocity, approximately a few meters per second.

TRM polarity distribution of the juvenile pumice clasts revealed no consistent orientation (fig. 4). Temperatures of 240-260°C were measured at points of steam emission on the surface of secondary ignimbrite at Pasig-Potrero valley five days after its emplacement. A temperature of 240°C was determined from an exposed 5- to 7-m-thick section of secondary ignimbrite 3 months after it was deposited at Sacobia valley. Together, these observations suggest typical emplacement temperatures of between 250°C and 300°C.

Secondary pyroclastic flows were commonly accompanied by secondary explosions and ash columns. Several of these ash columns were detected by weather radar and ground-based observation at heights comparable to violent vent ejections. Some events were not observed at all because they occurred during the night or cloudy weather. However, ash plumes were inferred from new ash layers on the surface of the ignimbrite fan and traces of ash on downwind vegetated areas. Though widely dispersed, the fallout deposits are preserved only near their source escarpments and are ephemeral elsewhere.



The generation of secondary pyroclastic flows involves a massive remobilization of hot, uncompacted primary ignimbrite. We envisage that two processes govern secondary pyroclastic-flow generation: overall reduction of friction within the primary ignimbrite and subsequent gravitational collapse. A decrease in the yield strength may result from an increase in pore pressure; gravitational collapse may be facilitated by the decrease in yield strength and by reduction of lateral support during channel erosion. The various thermodynamic and hydrologic conditions needed to cause a significant pore pressure increase and slope instability are yet to be resolved. Nevertheless, general aspects of the generation, triggering, and transport of secondary pyroclastic flows are outlined below.

Much of the pore fluid in the ignimbrite consists of water and steam in the originally dry interior of the primary ignimbrite. The manner by which water is introduced is unclear, but possibilities include upflow of ground water at the base of the ignimbrite induced by static loading on a porous and saturated river bed, percolation of rain water through cracks and open fumarolic pipes, and seepage from lahars, streamflows, and springs. As water recharges the hot primary ignimbrite, it is heated and a significant proportion is vaporized. As a result, the ignimbrite body becomes a pressurized system and remains as such prior to collapse. Poor internal permeability (due to a high proportion of fines), high lithostatic gradient, and a confining layer of fine ashfall prevent a significant release of pore pressure. Pore water vaporization decreases the stability of primary ignimbrite, priming it for collapse.

Heavy rainfall, channel erosion, secondary explosions, and earthquakes are possible triggers for collapse of the primary ignimbrite and, thus, for the generation of secondary pyroclastic flows. The close association of heavy rain and active lahars with secondary pyroclastic flows suggests a dominant role for rain and lahars in triggering secondary pyroclastic flows. Rain and lahars on the surface can load and perhaps destabilize an unstable block of primary ignimbrite. Lahars also cause deep incision on ignimbrite fans, erode channel walls into oversteepened and overhanging sections, and induce localized slope failures along the channel. The failing mass disintegrates into a small-scale cohesionless turbulent flow, accompanied by a convection of fine elutriates. Scarp collapse exposes fresh sections of primary ignimbrite, which sometimes react explosively in contact with lahars. Secondary explosions promote additional slope instability by forming steep escarpments and phreatic explosion craters. The combined effects of channel erosion and secondary explosions critically reduce the lateral support of potential failure surfaces. We also envisage that some avalanching involves rapid decompression of pore pressure at the collapsing front and triggers a runaway upslope propagation of the failure surface.

Felt earthquakes occurred on the same day as the Maraunot event, but many other earthquakes of similar magnitude were not followed by secondary pyroclastic flows. We do not claim an association, but we can envision, where unstable blocks are primed for failure, that ground shaking could induce compaction of the primary ignimbrite and, therefore, increase the pore pressure. If the pore pressure approximated the lithostatic load, the ignimbrite would lose its yield strength, undergo liquefaction with gaseous fluids as fluidizing agents, and flow downvalley over a very gentle gradient.

Transport of remobilized ignimbrite on a very gentle slope for several kilometers requires a mechanism that sustains fluidity. Characteristically low H/L ratios, the abundance of fumaroles on secondary ignimbrite surfaces, and the common association of high ash columns suggest that high gas pressure fluidizes the secondary pyroclastic flows until they are degassed to the point that they begin to deposit, from the bottom up. The absence of basal erosion, the massive character of deposits, channel-confined flow, and apparent slow flow velocities all suggest a high-concentration laminar flow in the depositional regime. Density and size segregation occur during transport of secondary pyroclastic flows, where dense block-size lithics settle near the headscarp area while large pumices are carried downstream and progressively segregated to the top of the flow.


The recognition of secondary ignimbrite at Pinatubo makes detailed interpretation of older ignimbrites more difficult. Having very similar sedimentary characteristics, primary and secondary ignimbrites are hardly distinguishable in the field unless one has knowledge of the depositional site before and after the events. However, some points of distinction can be made, including slight fines depletion, random TRM polarity of clasts in secondary ignimbrites, and, in well-incised sections, the nature of the underlying unit.

Secondary ignimbrite tends to be fines depleted relative to the primary ignimbrite from which it was derived (fig. 12). Some of the primary fine fraction is lost as elutriate that forms associated ash columns and secondary co-ignimbrite ash. Fines depletion also suggests that the secondary flowage at Pinatubo is gentle and does not grind up a new generation of fines to replace that being lost.

Figure 12. Size frequency curves of the (A) primary and (B) secondary ignimbrites plotted as cumulative weight percent on normal probability scale versus the particle size in f units and millimeters. A primary and secondary ignimbrite from the Sacobia valley, shown as SC*, were found to be in depositional contact and are shown individually, for comparison.

TRM polarity distribution of the clasts provides a more promising, yet not infallible, way to distinguish primary and secondary ignimbrite deposits. The uniform or bimodal TRM polarity of pumice clasts in primary ignimbrite at Pinatubo is in contrast to the generally random distribution in the secondary ignimbrite. Elsewhere, in primary ignimbrite that is dominated by magnetite and lacks self-reversing ferrian ilmenite (Nord and Lawson, 1989), the magnetic distinction between primary and secondary ignimbrite would be even more pronounced.

Because the chronology of events is known at Mount Pinatubo, identification of the underlying unit provides an added basis for recognition of secondary flow deposits. Primary ignimbrite usually overlies either the preeruption surface, the June 12 pyroclastic flow, the blast deposits, or the plinian fallout. By contrast, most secondary ignimbrites were emplaced on lahar or stream deposits that postdate the June 15 eruption.


There were no known fatalities during major secondary flow events at the Maraunot, Marella, and Balin Baquero events. These events occurred in very remote locations and at a time when people were still reluctant to venture near designated danger zones. However, the April 4, 1992, event, which was smaller than earlier flows, nearly caught people who were working in the Sacobia and Abacan River channels. The more serious impacts of secondary pyroclastic flows at Mount Pinatubo, such as burial and burning, are limited to the channel and areas proximal to the source. Light fallout of very fine ash from secondary co-ignimbrite and phreatic explosion columns affects distal areas. Ash plumes formed during secondary pyroclastic flows can also pose serious hazards to air traffic above the vicinity of Mount Pinatubo, as they can rise to the cruising altitudes of commercial airplanes (see description of August 1993 encounter, Casadevall and others, this volume).

Changes of morphology in the headscarp region and the emplacement of thick secondary ignimbrites alter the channel conditions for lahars and indirectly trigger other hazards. The April 4, 1992, event allowed the Sacobia to capture all flow from the Abacan River. Thus, lahar hazard was sharply reduced along the Abacan and raised along the Sacobia. No true debris flow has occurred in the Abacan channel since the April 4 secondary pyroclastic flow. Rapid aggradation of the channel can dam tributary streams and create a less efficient lahar-conveyance system. After the July 13, 1992, secondary pyroclastic flow in the Pasig-Potrero watershed, lahars stopped for more than a month, even during periods when the nearby Sacobia River had large lahars. On August 29, 1992, a large-volume, sustained lahar came down the Pasig-Potrero River, due to newly reintegrated drainage and breakout of water that had ponded behind a dammed tributary at the foot of Mount Cutuno (Arboleda and Martinez, this volume). Residents of the Mancatian area were surprised, and much damage occurred.


Secondary pyroclastic flows at Pinatubo are a documented example of a little known process of ignimbrite remobilization in a relatively dry state. Secondary movements in this sense had been suspected at the Valley of Ten Thousand Smokes, Alaska, and may be common after large ignimbrite-producing eruptions. However, distinguishing between primary, vent-derived ignimbrite and secondary ignimbrite is difficult. So far, at Pinatubo, the distinction lies in the recognition of fines depletion and non-uniform TRM polarity of clasts in secondary ignimbrite. Problems may persist when the loss of fines during transport is negligible or when the primary ignimbrite is emplaced at relatively low temperature and (or) contains abundant self-reversing ferrian ilmenite.

Secondary pyroclastic flows at Pinatubo were generated by massive avalanching of the primary ignimbrite. Pore fluid vaporization provides the mechanism for a positive pore pressure within the ignimbrite, while erosion at the toe of the slope of ignimbrite further decreases the stability of the mass. Failure is triggered by heavy rain, lahars, erosion of critical footslopes, secondary explosions, and perhaps also by earthquakes.

Secondary pyroclastic flows differ from secondary movements in ignimbrites noted by previous workers in that they (1) occur sporadically for months or years after deposition of the primary ignimbrite, (2) are generated in all parts of ignimbrite fans, (3) involve both gravity and gas fluidization, and (4) exhibit internal sedimentary characteristics very similar to valley-ponded primary ignimbrite. Unlike rheomorphic ignimbrite that requires a steep slope or temperatures sufficiently high to permit movement of the agglutinating pumice and glass shards, secondary pyroclastic flows may be generated on gentle slopes and at much lower temperatures.

Primary and secondary ignimbrite deposition represents a continuum of flow events beginning from the initial emplacement and continuing long after deposition. Final distribution of ignimbrite will be the result of a series of events of reworking and remobilization, which persist until the thermal energy of the system, the internal fluid pressure, the available water, or the channel gradient is no longer capable of mobilizing the pyroclastic materials. Many secondary ignimbrites may well have been mapped in the field as primary ignimbrite, and, for most purposes, they are the same. However, at active volcanoes, secondary pyroclastic flows need to be recognized as an important redepositional process and a potentially serious, lingering hazard of explosive volcanism.


We thank Wes Hildreth, Jim Riehle, Ray Punongbayan, George Walker, Tess Regalado, and Chris Newhall for valuable comments; the helicopter pilots, officers, and staff of the Philippine Air Force at Clark Air Base for providing aerial views and access to remote sections; Chris Newhall and Jim Anderson for lending us fluxgate magnetometers; our Aeta field guides, especially Boy Tanglaw; and our colleagues at the Philippine Institute of Volcanology and Seismology (PHIVOLCS), especially the staff of the Pinatubo Volcano Observatory and the Lahar Monitoring Group, for generously sharing information. The study was supported by the U.S. National Science Foundation (NSF) and PHIVOLCS. This is SOEST contribution no. 3556.

Note: An ash column of the type observed during secondary pyroclastic flow generation occurred on 11 July 1995, reaching 9 to 10 km in altitude. This event demonstrates that this phenomenon has persisted and still poses a considerable hazard on the ground and in the air 4 years after the main eruption.


Branney, M.J., and Kokelaar, P., 1992, A reappraisal of ignimbrite emplacement: Progressive aggradation and changes from particulate to non-particulate flow during emplacement of high-grade ignimbrite, Bulletin of Volcanology, v. 54, p. 504-520.

Casadevall, T.J., Delos Reyes, P.J., and Schneider, D.J., this volume, The 1991 Pinatubo eruptions and their effects on aircraft operations.

Chapin, C.E., and Lowell, G.R., 1979, Primary and secondary flow structures in ash-flow tuffs of the Gribbles Run paleovalley, central Colorado, in Chapin, C.E., Elston, W.E., eds., Ash flow tuffs: Geological Society of America Special Paper 180, p. 137-154.

Ellwood, B.B., 1982, Estimates of flow directions for calc-alkaline welded tuffs and paleomagnetic data reliability from anisotropy of magnetic susceptibility measurements: Central San Juan Mountains, southwest Colorado: Earth and Planetary Science Letters, v. 59, p. 303-314.

Fierstein, J., and Hildreth, W., 1992, The plinian eruptions of the 1912 at Novarupta, Katmai National Park, Alaska: Bulletin of Volcanology, v. 54, p. 646-684.

Fisher, R.V., Glicken, H.X., and Hoblitt, R.P., 1987, May 18, 1980, Mount St. Helens deposits in South Coldwater creek, Washington: Journal of Geophysical Research, v. 92, no. B10, p. 10267-10283.

Hayashi, J.N., and Self, S., 1992, A comparison of pyroclastic flow and debris avalanche mobility: Journal of Geophysical Research, v. 97, p. 9063-9071.

Hildreth, W., 1983, The compositionally zoned eruption of the 1912 in the Valley of Ten Thousand Smokes, Katmai National Park, Alaska: Journal of Volcanology and Geothermal Research, v. 18, p. 1-56.

Hoblitt, R.P., Miller, C.D., and Vallance, J.W., 1981, Origin and stratigraphy of the deposit produced by the May 18 directed blast, 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. 401-419.

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.

Nord, G.L., Jr., and Lawson, C.A., 1989, Order-disorder transition-induced twin domains and magnetic properties in ilmenite-hematite: American Mineralogist, v. 74, p. 160-176.

PHIVOLCS-NEDA, 1992, Pinatubo Lahar Hazards Map: Manila, Philippine Institute of Volcanology and Seismology and the National Economic Development Authority, 1 sheet at scale 1:200,000; 4 sheets, Quadrants 1-4, at scale 1:100,000.

Pinatubo Volcano Observatory Team, 1991, Lessons from a major eruption: Mt. Pinatubo, Philippines: Eos, Transactions, American Geophysical Union, v. 72, no. 49, p. 545, 552-553, 555.

Schmincke, H-U., and Swanson, D.A., 1967, Laminar viscous flowage structures in ash-flow tuffs from Gran Canaria, Canary Islands: Journal of Geology, v. 75, p. 641-664.

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.

Smithsonian Institution, 1991, Pinatubo: Bulletin of the Global Volcanism Network, v. 16, no. 8, p. 4-6.

Sparks, R.S.J., 1976, Grain size variations in ignimbrites and implications for the transport of pyroclastic flows: Sedimentology, v. 23, p. 147-188.

Walker, G.P.L., Hayashi, J.N., and Self, S., in press, Travel of pyroclastic flows as transient waves: implications for the energy line concept and particle-concentration assessment: Journal of Volcanology and Geothermal Research.

Wolff, J.A., and Wright, J.V., 1981, Rheomorphism of welded tuffs: Journal of Volcanology and Geothermal Research, v. 10, p. 13-34.

FIRE and MUD Contents

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

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