FIRE and MUD Contents
1U.S. Geological Survey.
2Philippine Institute of Volcanology and Seismology.
3Department of Geology and Geophysics, Hawaii Center for Volcanology, University of Hawaii at Manoa, Honolulu, HI 96822.
About 5.5 cubic kilometers of pyroclastic-flow deposits were emplaced during the climactic eruption of Mount Pinatubo volcano on June 15, 1991, which, combined with plinian pumice-fall deposits, distinguishes the event as one of the five greatest eruptions of the 20th century. Pyroclastic flows traveled as much as 12 to 16 kilometers from the vent in all sectors, impacted directly an area of almost 400 square kilometers, and profoundly altered the landscape. In proximal areas, flows were highly erosive and left little deposit, but, in medial and distal areas, they created broad, thick valley fills and fans of ponded pyroclastic-flow deposits as well as veneers on ridges and uplands.
Pyroclastic-flow deposits comprise three facies. First, by far the most voluminous are massive pumiceous deposits that form the valley fills and fans. Little evidence exists for individual flow or emplacement units except in some upland areas and near distal limits. Second, stratified pumiceous pyroclastic-flow deposits veneer uplands in medial areas and consist of numerous beds several centimeters thick. Stratified deposits grade laterally into, and are therefore cogenetic with, massive pyroclastic-flow deposits. Third, a prominent lithic-rich facies formed of clasts of Pinatubo's former summit dome overlies, or is interbedded in the upper parts of, pumiceous pyroclastic-flow deposits in medial areas of all major drainages.
Stratigraphic evidence indicates that pyroclastic-flow deposits were emplaced during most, and perhaps all, of the plinian pumice fall. We infer that pyroclastic flows were generated by either quasicontinuous, low-level collapse of portions of a sustained plinian column or by repeated brief collapse of the entire column. Either origin is different from classic models of an eruption featuring a period of high plinian column followed by column subsidence into low pyroclastic fountains that fed the bulk of the pyroclastic-flow deposits. The climactic eruption culminated with formation of a caldera 2.5 kilometers in diameter, the onset of which is marked stratigraphically by the abrupt appearance of the lithic-rich facies. We hypothesize that foundering of the caldera (1) produced abundant lithic clasts that were incorporated into the final pyroclastic flows, (2) choked the conduit, and (3) shut down the eruption.
Note to readers: Figures open in separate windows. To return to the text, close the figure's window or bring the text window to the front.
The climactic eruption of Mount Pinatubo on June 15, 1991, is one of the largest volcanic eruptions of the 20th century, as measured by the volume of its products. Relative to past large eruptions, events at Pinatubo were well monitored, and many aspects of eruption dynamics are known or can be inferred with some confidence, even though the climax occurred under poor viewing conditions, owing to the arrival of Typhoon Yunya. This report considers the role of pyroclastic flows in the climactic eruption only; small-volume pyroclastic flows associated with eruptions of June 12-13 (Hoblitt, Wolfe, and others, this volume; Torres and others, this volume), pyroclastic-surge deposits of June 14-15 (Hoblitt, Wolfe, and others, this volume), and secondary pyroclastic flows of 1991 and 1992 (Torres and others, this volume) are discussed elsewhere in this volume. This report includes discussion of (1) origin and behavior of pyroclastic flows, (2) distribution and volume of resulting pyroclastic-flow deposits, (3) types and facies of pyroclastic-flow deposits, (4) timing of pyroclastic flows relative to the major pumice fall of June 15 (Paladio-Melosantos and others, this volume) and other events, and (5) a model for formation of the 2.5-km-diameter caldera.
Owing to logistical and safety considerations, field investigations for this and a companion study (Hoblitt, Wolfe, and others, this volume) were limited, and we consider both to be reconnaissance studies. Fieldwork was restricted to sites around the margins of the area swept by pyroclastic flows that were accessible on foot and by vehicle and to scattered sites closer to the volcano reached by helicopter (fig. 1). Aerial photographs and views were useful in mapping the approximate extent of various units, but readers should keep in mind the lack of ground-based observations over much of the area.
Figure 1. Distribution of pyroclastic-flow and related deposits from the climactic June 15, 1991, eruption of Mount Pinatubo, geographic features referred to in text, and study sites. Triangle marks the former summit; heavy line outlines caldera; star marks approximate location of lava dome of June 7 and presumed location of initial vent. Locations of contacts are based on limited field study only and are inferred largely from aerial observations and photographic evidence. Distal margins of stratified pyroclastic-flow deposits and lithic-rich facies are poorly constrained; inferred presence of lithic-rich facies in eastern part of Marella valley is based on comparisons with other major valleys. Extent of ash-cloud deposits is largely inferred from outer limit of singed or killed vegetation. Patchy distribution of pyroclastic-flow deposits in northeast and southeast sectors (as seen in late June and early July 1991) reflects significant erosion of deposits from narrow canyon floors during heavy typhoon rains that accompanied eruption. Scarp symbol between sites 5 and 6 is drawn off massive pyroclastic-flow deposit for clarity; actual scarp lay in fill of Sacobia valley adjacent to symbol.
Pyroclastic flow, as used in this report, is a general term for hot, gravity-driven density currents of gas and particles that range from dense flows, whose flow regime is thought to be dominantly laminar, to turbulent flows having a lower particle concentration. Specific depositional facies are identified by modifiers that describe dominant clast lithology or sedimentary structures (for example, massive and stratified pumiceous pyroclastic-flow deposits; lithic-rich facies). We avoid the term ignimbrite (see Cas and Wright, 1987), even though appropriate, because the terms pyroclastic flow and pyroclastic-flow deposit were used widely in the Philippines before and after the eruption. Ash clouds, as used here, are composed dominantly of fine-grained particles (fine sand and silt) that are elutriated from and rise above moving pyroclastic flows and can affect areas beyond the margins of flows (see Cas and Wright, 1987). Ash-cloud deposits may originate by both surge (laterally moving, low-particle-concentration, turbulent currents) and fall (vertical fallout) processes from ash clouds.
Areal distribution, thickness, and volume of pyroclastic-flow deposits emplaced on June 15 provide perspectives on flow behavior and the major landscape changes they created, as well as key information for various measures of the size of the eruption. We refer to proximal, medial, and distal areas that lie at increasing distance from the current caldera rim. With variations of about 1 km, depending on azimuth, proximal areas lie within about 2 km, medial areas about 2 to 5 km, and distal areas about 5 to 15 km from the caldera rim. These modifiers were selected solely to fit the needs of this discussion; Hoblitt, Wolfe, and others (this volume) selected a twofold distance classification in their discussion of preclimactic events. The vent that produced the pyroclastic flows is thought to have been at or near the site of the June lava dome (Hoblitt, Wolfe, and others, this volume), which was located near the head of the Maraunot valley about 1.3 km northwest of the old summit (fig. 1).
Pyroclastic flows of June 15 and their associated ash clouds traveled as far as 12 to 16 km from the vent in all sectors and impacted directly an area of almost 400 km2 (fig. 1). The great bulk of their deposits is sandy and pumiceous; a minor and areally restricted component forms a lithic-rich facies, which is composed dominantly of gravel-size, dense lithic clasts. The remainder of this discussion of thickness and volume pertains to pumiceous pyroclastic flows and their deposits; lithic-rich facies will be discussed in a different section.
Distribution of pumiceous pyroclastic flows and their deposits was controlled chiefly by preexisting topography. The asymmetric distribution of deposits shown in figure 1 results from several conditions. Terrain in the west has relatively low relief, and terrain in the east is mountainous. High terrain south and southeast of the volcano may have deflected south- and some southeast-directed flows southwest into the Marella valley. Likewise, north- and northeast-directed flows may have been deflected by high terrain northeast of the volcano northwest into the O'Donnell and Bucao drainages. Furthermore, the vent was probably located in the deeply incised valley of the northwest-flowing Maraunot River. The possibility that strong typhoon winds enhanced pyroclastic-flow deposition on the west side can be rejected, as forecast wind data for June 15 show that winds below 7 to 9 km in altitude were blowing eastward, and only winds at higher altitudes were blowing westward (Paladio-Melosantos and others, this volume).
The preeruption landscape in the west sector consisted of variably dissected fans and terraces of older pyroclastic and epiclastic deposits (fig. 2A). Terrain varied from little-dissected plains to highly dissected badlands. Relief generally decreased outward from several hundred meters in proximal areas to several tens of meters in distal areas. A few prominent isolated hills rose hundreds of meters above this dissected surface. The new June pyroclastic flows were able to spread out across this broad terrain and bury large parts of the preexisting landscape (figs. 1, 2B). Medial areas contain typically thin veneers (less than several meters) of pyroclastic-flow deposits on undulating upland surfaces with thicker deposits (tens of meters) in some valley bottoms; mean surface gradients of valley fills over areas of several square kilometers are typically 0.08-0.12 (4-7°). Thin, discontinuous deposits occur on steep slopes of canyon walls, ridges, and isolated hills. Distal areas contain broad, thick (up to 200 m) fills of ponded pyroclastic-flow deposits in the southwest sector (Marella River valley), west to northwest sector (unnamed Balin Baquero tributaries and Maraunot valleys), and north sector (Bucao and O'Donnell valleys). Narrow canyon-filling deposits tens of meters thick extend outward from the thick ponds (figs. 1, 3); ash-cloud deposits mantle much of the intervening broad divides and tributary canyons.
Figure 2. Aerial views of Maloma and Marella River valleys comparing preeruption terrain (A) and posteruption terrain with fill of pyroclastic-flow deposits (B). Orientation of both views is northeast toward Mount Pinatubo, but B was taken from a position slightly east of and at higher altitude than that of A. Note hill surrounded by pyroclastic-flow deposits near center of B and compare with A. Main (east fork) channel of Marella River to east (right) of knob has been filled to a depth of about 200 m. (Both photographs by R.P. Hoblitt; A, June 3, 1991; B, June 23, 1991.)
Figure 3. Generalized profiles of former river channels and approximate depth of pyroclastic-flow deposits; approximate azimuth from vent labeled. Thin deposits not shown. Varied profiles near vent reflect rugged topography of preeruption Mount Pinatubo.
In contrast, terrain in much of the east sector is mountainous, with local relief typically 300 to 500 m, steep, rugged ridges, and deep, narrow canyons. The mountainous terrain is bisected by the broad Sacobia and Pasig River valleys, which contained a dissected fill of older pyroclastic and epiclastic deposits. Pyroclastic flows in this area overwhelmed much of the mountainous terrain in medial areas as well as large parts of distal areas, with deposition occurring chiefly in the Sacobia and Pasig valleys but also in several narrow canyons such as the Bangat, Malago, Porac, and Gumain (fig. 1).
The thickest pumiceous pyroclastic-flow deposits occur in broad ponds that cover areas of 10 to 30 km2 and bury former valley floors to depths of 100 to 200 m. Terraces and interfluves between the former valley floors have thinner, discontinuous accumulations. Figure 3 shows generalized profiles of former river channels and estimated surface altitude of pyroclastic-flow deposits. Gradients of most major channels and pond surfaces are about 0.05 (3°); those in the Marella valley are steeper, about 0.07 (4°).
Thinner tongues of pyroclastic-flow deposits lead downstream from the ponds and terminate where valley gradients are typically about 0.01 to 0.02 (0.5-1.0°; figs. 3, 4). In some narrow tributaries of the Balin Baquero and Bucao Rivers, pyroclastic-flow deposits terminate on steeper gradients of 0.02 to 0.04 (1-2°). Pyroclastic-flow deposits in narrow canyons in northeast and southeast sectors also terminate on relatively steep slopes of 0.03 to 0.4 (1.5-2°); their termination probably reflects increased resistance to flow in these narrow, sinuous canyons.
Figure 4. Aerial view to northwest showing ponded pyroclastic-flow deposits in upper Bucao River valley leading into narrow outlet channels. Note ridges of preeruption badlands in lower left. (Photograph by W.E. Scott, no. WES-91-12-6, July 5, 1991.)
Thick valley fills of pyroclastic-flow deposits blocked numerous tributaries to major drainages and created undrained or poorly drained areas, some of which became temporary lakes (fig. 5). Most lakes were shallow and short lived; many drained early in the rainy season of 1991 as they overflowed and outlet channels were rapidly incised through poorly consolidated pyroclastic-flow deposits. Such events produced lahars (Pierson and others, this volume; Umbal and others, this volume). Some lakes filled and drained repeatedly as secondary pyroclastic flows, generated by avalanching of thick pyroclastic-flow deposits (Torres and others, this volume), or lahars rebuilt blockages. Lakes and poorly drained areas in tributaries were also sites of secondary steam explosions driven by heated surface and shallow ground water; such areas were characterized by numerous craters with aprons of surge and fall deposits (fig. 5). Explosions began shortly after the June 15 eruption and continued through at least the next three rainy seasons.
Figure 5. Aerial view to west of blocked tributary along west (left) wall of lower Sacobia River valley (between sites 5 sites 6) showing a new lake, a scarp at head of failure that produced secondary pyroclastic flow shortly after original emplacement, and secondary phreatic explosion craters in right center. Height of failure scarp estimated to be 12 to 20 m. The lake drained during a rainstorm on July 25, 1991, and generated a lahar. (Photograph by W.E. Scott, no. WES-91-10-29, July 2, 1991.)
Pumiceous pyroclastic flows were highly mobile in proximal areas and much less so in distal areas. High mobility of pumiceous pyroclastic flows in proximal areas is illustrated by (1) the paucity of deposits in proximal areas, even on relatively flat terrain, and (2) the observation that flows crossed several drainage divides (fig. 1) and descended valleys that do not head directly on the volcano. Pumiceous pyroclastic flows swept over the steep slopes of proximal areas, removed virtually all vegetation and much soil and unconsolidated colluvium, and left only minor deposits in an area of almost 30 km2 (fig. 1). Most of the 1991 deposit in this area consists of fine-grained fall deposits from ash plumes that issued from the caldera for several weeks after June 15. The passage of energetic flows etched and fluted canyon walls--some as high as 200 m--composed of pre-1991 volcanic diamicts (fig. 6). Clearly, at least parts of some proximal flows were expanded in order to affect areas high above canyon floors, and all were sufficiently mobile to pass through the area and leave only small patches of pyroclastic-flow deposits in protected areas. We infer that much of the proximal area was in the deflation zone, the zone in which density currents of pyroclasts and gas were rapidly descending from the eruption column and segregating into density-stratified flows (Walker, 1985). If this inference is correct, the deflation zone extended 3 to 4 km out from the vent.
Figure 6. Aerial view of Holocene(?) diamicts exposed along valley wall on proximal west-southwest flank of Pinatubo. Note etching and fluting of surface by passage of pyroclastic flows, which moved from left to right. Height of slope in foreground is about 50 m. (Photograph by W.E. Scott, no. WES-92-22-20, March 17, 1992.)
Pyroclastic flows crossed the 200- to 400-m-high ridge 3 km northeast of the former summit of Pinatubo and flowed down the east tributaries of the O'Donnell River (Bangat and several unnamed streams between the Bangat and main O'Donnell) and the Malago River (western tributary of the Marimla River). The ridge altitude lay close to that of the vent. Flows also crossed lower divides (about 100 m) to enter the Gumain and Porac River canyons in the southeast sector. As previously discussed, the entire volume of flow material directed toward these high ridges did not pass over the ridge because much was transported into west-flowing drainages. Density stratification of flows would favor deflection of relatively dense, lower portions of flows. Blocking of flows (Valentine, 1987) by ridges would accentuate this process by increasing flow density. Although we have no field observations from the obstructing ridge tops, we infer that less dense upper portions of flows rode up and over ridges and then deflated and became concentrated in the narrow valley bottoms.
Yet as reflected by the following two proxies of mobility, pumiceous pyroclastic flows at Pinatubo were not as spectacularly mobile as some other examples. Ratios of vertical fall (measured from the vent altitude) to distance traveled for Pinatubo flows are about 1:11 to 1:14, whereas those of many other ancient and modern flows are as small as 1:30 to 1:75 (Sparks, 1976). Aspect ratios, which are represented by the fraction of average deposit thickness divided by a measure of areal extent (for instance, the diameter of a circle with an area equal to the that of the pyroclastic-flow deposits (Walker, 1983)), further suggest that Pinatubo pyroclastic flows were of comparatively moderate mobility. Aspect ratios for the entire 1991 sequence of Pinatubo flows range from about 1:480 to 1:1,200 (table 1), which fall in the upper range of values obtained for other pyroclastic-flow deposits (Walker, 1983). The larger fraction, or higher aspect ratio, was calculated by using the area of thick valley-fill deposits and thick upland veneer; the lower aspect ratio includes the area of thin veneer. The higher ratio is similar to the aspect ratio calculated by Walker (1983) for the Valley of Ten Thousand Smokes (Alaska) ignimbrite, a so-called high-aspect-ratio ignimbrite.
Table 1. Area, thickness, and volume estimates of pyroclastic-flow deposits derived by using valley-cross-section method (see text).
But both measures used here, vertical-fall-to-runout ratio and aspect ratio, employ assumptions that create uncertainties for comparisons. Actual fall-to-runout ratios for flows probably are larger than those measured, for many flows probably originated at varying heights above the vent from fountains or collapsing portions of eruption columns (Sparks and Wilson, 1976; Hoblitt, 1986). Aspect ratios suffer from comparing sequences containing variable numbers of flows, which obviously influences thickness comparisons. In addition, thin distal or upland veneers of older deposits may have been eroded prior to study, and this erosion would lead to an underestimated deposit area and higher ratios. Redistribution of primary deposits by secondary pyroclastic flows (Torres and others, this volume) beyond the border of primary flows would increase an area estimate and lead to lower ratios. Allowing for these uncertainties, the pyroclastic-flow deposits at Pinatubo probably represent flows of low to intermediate mobility.
Because pyroclastic-flow deposits were emplaced on a highly irregular landscape, measuring their volume is subject to substantial uncertainties. No detailed posteruption topographic maps are yet available that would allow volume determination by differencing preeruption and posteruption topography. We obtained two estimates of volume in the months after the eruption, but recent detailed photogrammetric measurements made by the U.S. Army Corps of Engineers as part of their study of sediment problems are probably the most accurate.
Our estimates used two different maps of valley-filling pyroclastic-flow deposits, one made by Scott and the other made by geologists at the Philippine Institute of Volcanology and Seismology (PHIVOLCS). Both maps used posteruption aerial photographs to map the extent of pyroclastic-flow deposits on the preeruption topographic base. All photographs were uncontrolled, so contacts had to be transferred by visual inspection. Locating contacts between pyroclastic-flow deposits and the preeruption surface entailed identifying features that had been buried versus those that had not. In areas of well-defined valleys with terraces, spurs, and hills at various heights, the altitude of the contact could locally be determined to within one contour interval (20 m). Contacts were then extended by interpolation between these relatively well-defined points. Uncertainties are greatest in western sectors, where pyroclastic-flow deposits bury broad areas of tens of square kilometers. Altitudes were estimated across these surfaces by constructing broadly convex contours from areas along margins where altitudes could be determined with more certainty. The differences in the two estimates stem largely from different judgments made in placing contacts and in the different methods used to calculate the volume.
Scott used a cross-section method. Figure 7 shows estimated 100-m contours on the surface of 1991 deposits in the Sacobia and Pasig valleys. Mean thickness of 1991 pyroclastic-flow deposits was calculated for cross sections at each 100-m contour interval and multiplied by the area of 1991 deposit between that and adjacent contours to obtain a volume for that reach of the valley. Reach volumes were summed to provide an estimate of the total volume of pyroclastic-flow deposits in the valley. Table 1 lists the estimated volume of pyroclastic-flow deposits for the Sacobia, Pasig, and other drainages.
Figure 7. Map and cross sections of pyroclastic-flow deposits in Sacobia and Pasig River valleys used in valley-cross-section method of estimating volume of deposits. Cross sections at 300- and 400-m contours not shown. Heavy line outlines caldera.
In addition, veneers of stratified and massive pyroclastic-flow deposits drape broad upland regions between major valleys in medial areas but contribute little to the total volume. If we assume a reasonable mean thickness (5 m) for the 30-km2 area patterned in figure 1, the volume is appreciably less than 10 percent of the total estimate of about 5 km3. A much thinner and discontinuous veneer drapes the rugged uplands in east sectors and elsewhere within the margin of ash-cloud deposits shown on figure 1. Mean deposit thickness is probably much less than 1 m in this large area, which accounts for only minor additional volume.
The PHIVOLCS group digitized (at 100-m contours) the preeruption topography and their estimates of posteruption topography, aided by numerous field measurements and estimates of thickness of pyroclastic-flow deposits exposed in gullies and canyons after the substantial erosion that occurred during the rainy season of 1991. From these data, a PC-based contouring program calculated a volume of 4.2 to 4.5 km3.
The Portland District of the U.S. Army Corps of Engineers (K. Eriksen and M. Pearson, oral commun., 1993) used controlled aerial photographs to map photogrammetrically the boundaries of valley-filling pyroclastic-flow deposits on the preeruption base. They also accounted for much of the veneer of deposits between major valleys. They digitized preeruption topography (20-m contours) and followed a cross-section method similar to, but more detailed than, that described above to measure volumes. Their total volume measurement is about 5.5 km3 (table 1).
An early volume estimate of 7.1 km3 made by geologists at PHIVOLCS (J. Daligdig and G. Besana, written commun., in Pierson and others, 1992) is considered to be too large. Overestimates of volume in the O'Donnell, Sacobia, Abacan, and Pasig valleys account for most of the discrepancy with the other data sets.
The wide range of volume estimates, about 4 to 7 km3, is thought to reflect chiefly uncertainties in locating contacts. Significantly, the Corps of Engineers' method, which most reduced this uncertainty through use of controlled photographs and analytical plotters, gave a figure in the middle of the range, 5.5 km3. But uncertainties in thickness of upland veneer deposits and ash-cloud deposits that cover broad areas still introduce an error of a few tenths of a cubic kilometer, as do uncertainties created by the 20-m contour interval on the preeruption base. Therefore we judge a realistic estimate of the total bulk volume of pyroclastic-flow deposits to be 5.5+-0.5 km3.
We have no measurements of in-situ bulk density from which we can calculate dense-rock or magma volume, because deposits were too friable to sample accurately. We poured representative samples of Pinatubo pyroclastic-flow deposits into 500-mL beakers in the laboratory and got initial bulk density values of 1.1 g/cm3; with shaking and tamping the density increased to 1.3 to 1.4 g/cm3. The maximum density we could attain was 1.5 g/cm3 by tamping each increment as it was added, but the resulting degree of compaction appeared greater than that observed in deposits in the field. Bulk-density values obtained from similar deposits at other volcanoes range from about 0.7 to 1.5 g/cm3 (Smith, 1960; Riehle, 1973; Riehle and others, 1992; Wilson and Head, 1981). Therefore, we consider that 1.0 to 1.3 g/cm3 appears a reasonable range of bulk density, which indicates a dense-rock volume of 2.1 to 3.3 km3. The maximum, probably unreasonably high, dense-rock volume obtainable by using a density of 1.5 g/cm3 and bulk volume of 6 km3 is 3.8 km3. All estimates use a density of dacitic magma of 2.4 g/cm3.
Pyroclastic-flow deposits of June 15 comprise three facies having contrasting clast composition and sedimentary structures--massive and stratified pumiceous facies and lithic-rich facies. The two pumiceous facies, consisting of pumiceous ash, glass shards, crystals, and subordinate pumice lapilli and blocks, constitute the bulk of the pyroclastic-flow deposits. They show no conspicuous compositional grading and are light gray to pinkish light gray when dry, becoming darker gray on wetting. Lithic clasts of older Pinatubo lavas are widely dispersed through the pumiceous units. Most of the pumiceous facies consists of massive deposits with little if any discernible bedding; thickness ranges from tens of centimeters to 200 m. Associated with the massive flow deposits in medial areas are stratified pyroclastic-flow deposits that have conspicuous beds a few to tens of centimeters thick of alternating lapilli and ash. The stratified pyroclastic-flow deposits veneer broad, undulating uplands and are interbedded with and transitional to massive pyroclastic-flow deposits that occupy low areas. Lithic-rich facies are confined to proximal, medial, and near-distal areas of major drainages. They include proximal to medial clast-supported breccias of dense lithic clasts from the former summit dome in a pumiceous matrix, and medial to near-distal pyroclastic-flow deposits that are rich in lithic clasts. In distal areas, ash-cloud deposits are interbedded with and veneer areas adjacent to massive pyroclastic-flow deposits. They were deposited by surge and fall from ash clouds elutriated from moving pumiceous pyroclastic flows.
Massive pumiceous pyroclastic-flow deposits constitute the majority of pyroclastic-flow deposits emplaced on June 15. They form large valley ponds in the western sector, broad, thick fills in the Sacobia and Pasig valleys, as well as thinner, narrower fills in canyons in the northeast and southeast sectors (fig. 1). Massive pyroclastic-flow deposits also lie in many lows in uplands of medial areas, where they are associated with stratified pyroclastic-flow deposits. The massive deposits are friable, except where they are moist; fresh exposures of still-hot material easily ravel and collapse. No evidence of welding has been observed. Basal layers (layer 2a of Sparks and others, 1973), from which coarse clasts have been excluded by high shear forces, are displayed locally at the bases of some massive pyroclastic-flow deposits. Ground layers (layer 1 of Sparks and others, 1973), fines-depleted sediments that may underlie the pyroclastic-flow deposit proper, have been observed widely in medial and distal areas. Most are thin (<20 cm) and relatively fine grained (coarse ash, locally containing fine lapilli); the scarcity of coarse-grained ground layers suggests that massive flows were emplaced at comparatively low velocity (Wilson and Walker, 1982).
As of November 1992, exposures revealed only sparse contacts between flow units, although fine ash that typically coats exposures makes observing or tracing flow contacts difficult. Crude bedding is defined by local concentrations of coarse pumice clasts (fig. 8). Sharper contacts, as illustrated in figure 9, are created by a slight enrichment of fines at the top of a flow, which contrasts with the better sorted, relatively fines-depleted base (layer 1) of an overlying flow (sample 23, fig. 10A). Such contact relations suggest sufficient time elapsed between flows for fine ash to settle out (layer 3 of Sparks and others, 1973) and for the surface to firm up enough that the subsequent flow did not disrupt it. Multiple flow units, each unit 1 to 5 m thick, are found mostly in narrow canyon fills downstream from major ponds and in medial-area uplands between ponds, where massive deposits lie in topographic lows. The lack of discernible contacts in deep exposures (up to 80 m) in major ponds (although few ponds have been scrutinized closely) suggests that either single flow units are at least tens of meters thick or the ponded deposits accumulated without sufficiently long time breaks to allow for development of marked contacts. Poorly defined or cryptic contacts in massive ponded deposits would result from the homogeneity of pyroclastic-flow material and the inflated, gas-charged character of recently deposited flow units, which would foster some mixing with successive flows.
Figure 8. Thick (about 20 m), massive pumiceous pyroclastic-flow deposits exposed in canyon cut by west fork of the Marella River (east of site 17, fig. 1); flow from right to left. On figure 2B this section is located between hill surrounded by pyroclastic-flow deposits and point formed by ridge of old terrain left of and slightly below hill. Note pumice-concentration zone (P); otherwise flow is massive without apparent flow-unit breaks. Lahar deposits and moist pyroclastic-flow deposits comprise darker colored material near top of exposure. (Photograph by W.E. Scott, no. WES-92-19-9, March 11, 1992.)
Figure 9. Contact between two pyroclastic-flow units in tributary of Balin Baquero River about 5 km west-southwest of caldera rim (site 23, fig. 1). Locality is in upland area where ridges are veneered with stratified pyroclastic-flow deposits and valley bottoms contain massive pyroclastic-flow deposits. A, Outcrop view of two flow units separated by sharp contact. Shovel head is at contact; handle is 43 cm long. Unit above shovel is overlain by darker colored lahar deposits. B, Closeup view of contact in A showing fine-ash-enriched top of lower pyroclastic-flow deposit (23-mm-diameter coin for scale) contrasting with relatively fine-ash-depleted base of overlying deposit (sample 23, fig. 10A; b in fig. 11). (Photograph by W.E. Scott, no. WES-92-22-13, March 17, 1992.)
Figure 10. Cumulative grain-size plots of pyroclastic-flow deposits. Numbers refer to sites shown on figure 1; letters indicate multiple samples from same site. A, Representative massive pyroclastic-flow deposits. B, Individual beds of stratified pyroclastic-flow deposits (11A-D) and matrix of lithic-rich facies (1B).
Massive pyroclastic-flow deposits are composed dominantly of sand-sized material, which accounts for 70 to 85 wt% of analyzed samples (fig. 10A). The samples are not large enough to represent a statistically valid sample of grain sizes coarser than 16 mm (-4 ), but such excluded coarse material probably constitutes no more than an additional 5 to 20 wt% of sampled volumes. Fine ash (finer than 0.0625 mm, or 4 ) accounts for only a few percent to a maximum of 18 wt%, which is lower than that of many other pumiceous pyroclastic-flow deposits (table 2), and may account, in part, for the comparatively low mobility of the Pinatubo flows. Possible reasons for this are discussed in another section. The median4 (Md) grain size of most massive pyroclastic-flow deposits analyzed is coarse sand (0 to 1 ); few are medium to fine sand (1 to 3 ; table 3, fig. 11). Sorting is typically poor (graphic standard deviation =1.71-3.07 ); almost all samples are near symmetrical to fine skewed. Coarse pumice lapilli (64-256 mm) and blocks (>256 mm) are scattered through pyroclastic-flow deposits, with local concentrations (fig. 8), especially near flow margins. Vertical fluid-escape pipes are ubiquitous and conspicuously fines depleted. Compared with certain other pyroclastic-flow deposits as compiled on sorting versus median grain-size plots (Sparks, 1976), massive deposits at Pinatubo are somewhat better sorted (fig. 11).
4Grain-size parameters used in this discussion are those of Inman (1952) and Folk (1980); modifiers indicating degree of sorting are those of Cas and Wright (1987).
Figure 11. Median grain size (Md) plotted against sorting () for various facies of pyroclastic-flow deposits defined in text. Stippled area encloses field of pyroclastic-flow deposits compiled by Sparks (1976).
Table 2. Percent fine ash (silt- plus clay-size fraction; finer than 4 , or 62.5 m) of pumiceous pyroclastic-flow deposits from historical and prehistorical eruptions.
Table 3. Grain-size parameters of pyroclastic-flow deposits plotted in figures 10 and 11.
Table 3 footnotes:
aNumerals and letters in parentheses correspnd with those in figures 1 and 10; * denotes samples plotted in figure 10.
bName of drainage basin: M, medial; and D, distal.
cType of pyroclastic-flow deposit: M, massive pumiceous; B, fines-depleted layer 1 of massive pumiceous; S, stratified pumiceous; L, lithic-rich facies.
Stratified pumiceous pyroclastic-flow deposits, typically less than 5 m thick, veneer much of the dissected uplands in medial areas between extensive ponds of massive pyroclastic-flow deposits (fig. 1). Within the uplands, stratified deposits are locally interbedded with thin, massive pyroclastic-flow deposits and grade laterally into massive pyroclastic-flow deposits as thick as tens of meters that occupy valley floors and other topographic lows. The resulting differential draping reduced and muted former topographic relief in the uplands (fig. 12), but posteruption erosion has largely followed preexisting drainage lines. Stratified deposits are being gullied and stripped from divides, and thick fills along valley bottoms are being intricately incised to near or locally below preeruption levels. The landscape in these areas is now a complex mosaic of 1991 and older deposits.
Figure 12. Stratified pyroclastic-flow deposits draping rolling preeruption upland surface between Gumain and Sacobia River basins about 2.5 km southeast of caldera rim (site 11, fig. 1); view to east. (Photograph by R.P. Hoblitt, March 18, 1992.)
Stratification is variable in both thickness and clarity (fig. 13A). Centimeter-scale stratification is defined by tens of alternating lenses and beds of (1) subround to round pumice lapilli, (2) ash, and (3) slightly cohesive, poorly sorted ash and scattered lapilli that are similar in texture to massive pyroclastic-flow deposits. Some beds of sand-size ash display lamination and cross-lamination on a millimeter scale (fig. 13B). Bedding attitudes in stratified pyroclastic-flow deposits range from planar beds that roughly parallel the underlying ground surface (figs. 12, 13A) to longitudinal-dune crossbeds with dips of about 5° to 20° that define bedforms tens of centimeters to 1 m or more high and meters to tens of meters long (fig. 13C). Contacts are chiefly gradational at scales of millimeters to centimeters; graded contacts at tops of lapilli beds may, in part, represent infilling by overlying ash of originally better sorted lapilli beds. Erosional discontinuities, in which a bed or bed set is locally scoured, occur in many sections. Grain size and sorting of individual beds vary greatly in stratified pyroclastic-flow deposits. Lapilli beds are well to poorly sorted (=1.68 and 2.55 ); sandy beds are well sorted (=1.31-1.63 ). Representative samples of stratified deposits plotted in figures 10B and 11 are typically better sorted than massive pyroclastic-flow deposits of similar mean grain size and fall near the better sorted edge of the field of pyroclastic-flow deposits of Sparks (1976) on a plot of median grain size versus sorting. Deposits having similar grain size, sorting, and stratification include ignimbrite-veneer deposits at Taupo, New Zealand (Walker and others, 1981; Wilson and Walker, 1982), proximal bedded pyroclastic-flow deposits at Mount St. Helens, Washington (Rowley and others, 1985), high-energy proximal ignimbrite at the Valley of Ten Thousand Smokes (Fierstein and Hildreth, 1992), and stratified lithofacies in the Neapolitan Yellow Tuff, Italy (Cole and Scarpati, 1993).
Figure 13. Stratified pyroclastic-flow deposits exposed in gullies near site 11 (A and C) and site 23 (B). A, Planar bedded to weakly crossbedded deposit showing centimeter-scale beds of alternating ash and fine lapilli; flow direction from left to right. Grain-size analyses of samples of some of these beds are plotted in figures 10 and 11. (Photograph by W.E. Scott, no. WES-92-24-30, March 18, 1992.) B, Crossbedded stratified pyroclastic-flow deposits showing small-scale dune forms (23-mm-diameter coin for scale); flow direction from left to right. (Photograph by W.E. Scott, no. WES-92-22-9, March 17, 1992.) C, Large-scale dune form in stratified pyroclastic-flow deposits showing foreset beds overlain by topset beds that become inclined foresets to right; flow direction from left to right; handle of shovel at base of exposure is 43 cm long. (Photograph by W.E. Scott, no. WES-92-24-2, March 18, 1992.)
Stratification changes laterally from sharply to vaguely defined as stratified pyroclastic-flow deposits grade into massive pyroclastic-flow deposits. Such stratigraphic relations, which we have observed most clearly within broad, rolling upland areas (fig. 14), especially on the west side of Pinatubo, demonstrate that both stratified and massive deposits were produced contemporaneously from the same flow; this point supports the inference of Cole and Scarpati (1993) that a pyroclastic flow can entail a continuum of flow processes, from dense, nonturbulent portions that form massive deposits to turbulent, low-particle- concentration portions that form stratified deposits. In addition, the relation of stratified and massive deposits suggests that the massive pyroclastic-flow deposits accumulated gradually, either from numerous discrete flows or from quasisteady flow, even though there is no evidence for fluctuating depositional conditions in the massive deposits themselves, except locally, as discussed previously. The textural and compositional homogeneity of the deposits would make recognition of contacts difficult. In addition, the inflated gas-charged character of recently deposited flow units would have fostered erosion and mixing by successive flows, further obscuring contacts. Our view of gradual accumulation of the massive fills fits the model of progressive deposition of ignimbrite proposed recently by Branney and Kokelaar (1992).
Figure 14. Simplified block diagram showing relationship of stratified and massive pyroclastic-flow deposits in dissected upland terrain of medial area. Veneer (thickness is exaggerated) of stratified pyroclastic-flow deposits (medium gray) on divides intertongues with thick, massive pyroclastic-flow deposits (dark gray) in valley bottom.
The widespread presence of stratified pyroclastic-flow deposits at the Valley of Ten Thousand Smokes and Pinatubo, both of which are high-aspect-ratio ignimbrites (see "Behavior of Pumiceous Pyroclastic Flows"), counters the inference of Walker and others (1981) that stratified deposits like their ignimbrite-veneer deposits are well developed only in low-aspect-ratio ignimbrites. Perhaps low-aspect-ratio ignimbrites have stratified deposits developed over a larger proportion of their area, but the presence of stratified deposits alone is not firm evidence of a low-aspect-ratio ignimbrite.
Lithic-rich facies of pyroclastic-flow deposits range between two endmembers that contrast in clast size and amount of matrix--coarse-grained, clast-supported lithic breccias and matrix-supported, though matrix-poor, lithic-rich pyroclastic-flow deposits. The lithic-rich facies is confined to major drainages that head on the volcano and extends up to 4 to 8 km from the caldera rim (fig. 1). Stratigraphically, lithic-rich facies occur at or near the top of the pumiceous pyroclastic-flow deposits; pumiceous pyroclastic-flow deposits that overlie lithic-rich facies rarely exceed a few meters.
Lithic breccia, typically less than 5 m thick, displays an interlocked framework of angular to subangular gray to red lithic clasts chiefly decimeters and rarely >=1 m in length (fig. 15A). The lithic clasts are fragments of Pinatubo's old summit lava dome; the matrix consists of pumiceous ash and lapilli with variable amounts of pulverized summit-dome lava. The two samples of matrix for which we have grain-size distributions are somewhat coarser and more poorly sorted than matrix of massive pumiceous pyroclastic-flow deposits (sample 1B, fig. 10B; fig. 11).
Some of the lithic breccias at Pinatubo resemble near-vent lithic breccias associated with other pumiceous pyroclastic-flow deposits as described by Druitt and Sparks (1982) at Santorini (Greece) and Druitt and Bacon (1986) at Mount Mazama (Oregon) but differ in that they occur stratigraphically very near or at the top of pumiceous pyroclastic-flow deposits rather than more widely interbedded through the proximal sequence of pyroclastic-flow deposits (fig. 15B). Also, lithic breccias at Pinatubo form a single bed rather than a stratified sequence of units as described elsewhere. Although lithic breccias lie directly on the eroded preJune-15 surface in many proximal areas, downstream tracing of these deposits demonstrates that they lie above all or nearly all of the pumiceous pyroclastic-flow deposits. Earlier pumiceous pyroclastic flows passing these proximal sites left virtually no deposits, whereas lithic breccias represent a lag from later flows; that is a size fraction that could not be transported farther.
The lithic-rich pyroclastic-flow deposits are intimately associated with lithic breccias but are distinguished from the breccias by having sufficient matrix to support coarse clasts. They appear to have moved as single flow units rather than to have accumulated as a lag. Lithic-rich pyroclastic-flow deposits typically overlie or grade upward from lithic breccia in medial areas and extend farther downstream into near-distal areas. Although we have no systematic measurements, clast size appears to decrease upward from lithic breccia to lithic-rich pyroclastic-flow deposits and to decrease downstream in lithic-rich pyroclastic-flow deposits.
Figure 15. Typical character and stratigraphic position of lithic-rich facies. A, Closeup view of lithic-rich facies exposed along O¹Donnell River about 5 km north-northeast of caldera rim (site 1, fig. 1) showing clast-supported fabric; much of light-colored area is surface mineral efflorescence. Shovel handle is 43 cm long. Photograph by W.E. Scott, no. WES-92-14-29, March 4, 1992.) B, View upstream of lithic-rich facies overlying June 15 massive pyroclastic-flow deposits in 10-m-deep tributary to west fork of Marella River; site is about 4.5 km southwest of caldera rim (site 21, fig. 1). Clasts of lithic-rich facies form lag on eroding slopes, but in situ breccia forms unit about 2- to 3-m thick at top of section (L). (Photograph by W.E. Scott, no. WES-92-9-15, February 26, 1992.)
Distal areas adjacent to pyroclastic-flow deposits contain massive to bedded ash deposits that originated by surge and fall from ash clouds elutriated from moving pumiceous pyroclastic flows. These ash-cloud deposits are typically tens of centimeters thick and are light gray to light pinkish gray; grain size of individual beds ranges from sand with scattered fine lapilli to silt with various mixtures between these endmembers (fig. 16). Stratification is highly variable--from faint to well defined, from plane-parallel lamination to low-angle cross lamination. Such deposits mantle valley slopes and uplands adjacent to pyroclastic-flow deposits (fig. 1) and are locally intercalated with massive pyroclastic-flow deposits (see "Relative Timing of Events of the Climactic Eruption of June 15"). As the valleys filled with pyroclastic-flow deposits, ash-cloud deposits of earlier flows were buried by subsequent flows (fig. 16). Ash-cloud deposits are typically restricted to within a few hundred meters of pyroclastic-flow deposits but may extend 1 km or more away on surfaces between channelized deposits that lead downstream from major ponds, such as between Balin Baquero and Bucao River tributary valleys in the west sector (fig. 1).
Figure 16. Massive to weakly stratified ash-cloud deposits (ac) lying above tephra layers C and B, which lie on former unimproved road surface and below massive pyroclastic-flow deposits (pfl) near mouth of Pasig River canyon (site 12, fig. 1). Total thickness of 1991 deposits is about 1 m. This section is portrayed diagrammatically in figure 18 (section C). View is downstream. Photograph by W.E. Scott, no. WES-92-88-5, February 25, 1992.)
Pyroclastic-flow deposits of June 15 are mantled by a variable thickness of younger strata that originated by several processes.
Figure 17. Postclimactic deposits overlying June 15 massive pyroclastic-flow deposits in Sacobia River valley about 0.5 km south of area shown in figure 5 (site 6, fig. 1). Fine-grained, laminated ash-fall deposits of layer D lie to right of trowel, which is 27 cm long. Coarser grained bedded sediments above layer D were emplaced by fall and base surges generated by secondary (rootless) phreatic explosions in hot pumiceous pyroclastic-flow deposits. (Photograph by W.E. Scott, no. WES-92-19-16, March 13, 1992.)
The climactic eruption of June 15 occurred under extremely poor viewing conditions, owing to the arrival of Typhoon Yunya (Oswalt and others, this volume). Satellite observations provide information on behavior of the stratospheric portion of the eruption column (Koyaguchi and Tokuno, 1993; Koyaguchi, this volume; Lynch and Stephens, this volume), but detailed observations of pyroclastic flows, surges, and other events were limited (Hoblitt, Wolfe, and others, this volume; Sabit and others, this volume). Hence, the stratigraphic record provides key evidence for understanding the eruption's course. Deposits of June 12 to early afternoon June 15 are clearly differentiated from deposits of the climax on the basis of lithology and character (Hoblitt, Wolfe, and others, this volume; Pallister and others, this volume). A variably complete sequence of these deposits mantled much of the landscape within 20 km of the vent by early afternoon of June 15, when the sustained plinian eruption commenced. In distal areas, except along the southwest-trending andesitic scoria-fall deposit of June 12 (layer A of Koyaguchi and Tokuno, 1993; and Paladio-Melosantos and others, this volume), preclimactic deposits are dominantly sand size and finer (layer B of Koyaguchi and Tokuno, 1993; and Paladio-Melosantos and others, this volume); in medial areas they comprise a sequence of numerous graded beds with coarse-grained bases. In contrast, the June 15 pumice-fall deposit (layer C of Paladio-Melosantos and others, this volume) is coarse grained, forming a normally graded bed of dacitic pumice lapilli, ash, and minor lithic debris. The pyroclastic-flow deposits of the climax, as discussed in previous sections, are distinguished by their great thickness and uniform dacitic composition. These climactic June 15 deposits are capped by a variable thickness of one or more types of younger sediments (see "Strata that Overlie Pyroclastic-Flow Deposits").
Stratigraphic relations between layer C and the pyroclastic-flow deposits convey information about the timing and dynamics of the climactic eruption. Below, we summarize data from numerous medial and distal localities that cover all sectors of the volcano. Little information comes from proximal areas, owing to the absence of deposits. In general, stratigraphic sections in distal areas consist of pyroclastic-flow and related deposits overlying or interbedded with layer C, whereas sections in medial areas are composed entirely of pyroclastic-flow deposits with no trace of layer C.
Stratigraphic relations at the mouth of the Pasig River canyon (fig. 1, site 12) are typical of those in distal areas (fig. 18). The sections are situated on a vegetated terrace about 100 m south of and 10 to 20 m above the pyroclastic-flow margin (section A), on or near the former valley floor as exposed in what is now a terrace at the edge of the recently downcut channel (section B), and at the margin of pyroclastic-flow deposits on a former road surface along the north valley wall (section C). Each section has a basal layer about 5 to 10 cm thick of fine-grained ash-fall deposits (layer B) related to pyroclastic surges of June 14 and 15 (Hoblitt, Wolfe, and others, this volume). A thin (20 cm) clayey lahar deposit is interbedded in the basal deposits in section B.
Figure 18. Stratigraphic relations of pumice-fall and pyroclastic-flow deposits at mouth of Pasig River canyon (site 12, fig. 1). Figure 15 is photograph of section C. Correlation lines joining columns separate tephra layers D and B from deposits of climactic eruption. Deposits of June 15 climactic eruption: ac, ash-cloud deposits; pfl, pumiceous pyroclastic-flow deposits (solid-symbol pattern); C, tephra layer C (open-symbol pattern); a, ash-rich beds within pumice-fall deposits. Unit a is inferred to be distal equivalent of deposits of pyroclastic flows that terminated upvalley from sections. Note different depth scales for each column.
The lower, coarse-grained portion of normally graded layer C forms the base of climactic deposits in all three sections. The normal grading of layer C is interrupted by thin beds of silt- and fine-sand-sized ash that may reflect fallout from ash clouds of pyroclastic flows. We infer that pyroclastic flows terminated upvalley but their ash clouds continued downvalley to this site. Total thickness of layer C at sections A and C is about 20 to 25 cm, including the finer grained beds. At section C, layer C was followed by ash-cloud deposits and a pyroclastic-flow deposit, whereas section A lay above the reach of pyroclastic flows and received only minor ash-cloud deposits. At section B, in contrast, only about 10 cm of layer C, including an upper finer grained ash bed, accumulated before burial by about 4 m of pyroclastic-flow deposits. Another 5 to 10 cm of layer C accumulated before emplacement of three more pyroclastic-flow deposits. In summary, (1) some of layer C preceded pyroclastic flows reaching this site, (2) ash-rich beds in layer C record pyroclastic flows that terminated farther up valley, (3) the first pyroclastic flow reached this site after about half of layer C had accumulated, (4) layer C continued to accumulate while pyroclastic flows did not reach the site, and then (5) several more pyroclastic flows reached the site before activity waned, and a thin layer of fine ash (layer D of Paladio-Melosantos and others, this volume) accumulated.
This basic sequence is repeated in other distal areas (fig. 19, sections 18, 26), but at one locality on the Maloma-Marella drainage divide (section 17) there is evidence of pyroclastic flows occurring earlier in the sequence. Here, a total of about 10 to 15 cm of coarse-grained layer C is divided into four parts of subequal thickness by three cohesive beds that we interpret as ash-cloud facies of pyroclastic flows. In nearby areas beyond the pyroclastic-flow margin, layer C is 30 cm or more thick; thus, the pyroclastic flows that produced the ash-cloud deposits must have occurred during the early part of the pumice fall.
Figure 19. Stratigraphic relations of 1991 pyroclastic-fall and pyroclastic-flow deposits in distal areas. Sections are located on figure 1; symbols same as on figure 18. Unit pfl with lined-symbol pattern (section 25) is lithic-rich facies of climactic pyroclastic-flow deposits. Pyroclastic-flow deposits at top of sections 13 and 17 are several meters thick. Unit B in sections 18 and 26 may include fall deposits of June 12-13 eruptions; unit A is the scoria-fall deposit of June 12.
In other distal areas, pyroclastic-flow deposits or their ash-cloud facies overlie all or part of layer C (fig. 19, sections 2, 3, 4, 8, 13, 25, 27). In some of these sections (2, 8, 13, 25, and 27), layer C is thin compared with its thickness in nearby areas that were not swept by pyroclastic flows or their ash clouds. We don't know if this is due to erosion of an initially thicker layer C by subsequent flows, or if flows occurred concurrently with the later part of the pumice fall and the upper part of layer C was incorporated in the flows. Some distal sites, therefore, give evidence that pyroclastic flows were being produced concurrently with the pumice fall, but emplacement of much of the distal pyroclastic-flow deposits may have occurred late in the pumice fall. Such evidence does not preclude pyroclastic flows being generated early in the pumice fall, because we presume that early flows would have encountered the greatest resistance to flow in the narrow, sinuous, preeruption canyons and terminated upvalley. But as filling progressed, resistance to flow would have decreased, and subsequent flows could extend farther downvalley.
The absence of layer C in medial areas, either below thick sequences of flow deposits in valleys or below thin stratified pyroclastic-flow deposits in uplands, is consistent with but not unambiguous evidence for pyroclastic flows coinciding with pumice fall. Nearly continuous or pulsating flows would incorporate falling tephra. If distinct pyroclastic flows occurred, succeeding flows could sweep away any tephra that fell in the likely brief (minutes?) time intervals between flows. Coarse pyroclasts might even become embedded in the inflated, gas-charged tops of recently deposited units. In areas of stratified pyroclastic-flow deposits, turbulent flow conditions would ensure removal of loose pumice-fall deposits.
Whether or not initial pumice-fall deposits mantled the surface of medial areas before the arrival of the first pumiceous pyroclastic flows is difficult to assess. Our only evidence is negative. We have inspected six 1-km2 medial sites in which pyroclastic-flow deposits are relatively thin (<15 m) and highly dissected and have found no trace of layer C, even in relatively protected areas behind ridges where the flows killed trees but did not remove them. The bases of pyroclastic-flow deposits lie either on a differentially eroded sequence of preclimactic fall and surge deposits or at some level in the soil or subsoil. Fine-grained surge beds of layer B were relatively resistant to erosion, but once they were removed the underlying coarse-grained layer A was typically entirely stripped away. These scenarios suggest that layer C would have suffered a similar fate. As of this writing, erosion had not yet exposed bases of the thickest sequences of ponded pyroclastic-flow deposits in medial areas, which may have provided sites for preserving early pumice-fall deposits. The presence of at least part of layer C below thick pyroclastic-flow deposits in section 25 (fig. 19) in the near-distal Maraunot area suggests that some pumice-fall deposit will probably be found as erosion exposes more widely the base of the pyroclastic-flow deposits. In summary, we cannot yet determine the relative timing of the beginning of pyroclastic-flow activity with regard to the start of pumice fall, but strong stratigraphic evidence demonstrates that pumiceous pyroclastic flows were generated during much of the pumice fall.
Our preliminary investigations of pyroclastic-flow and related deposits of the June 15 climactic eruption of Pinatubo volcano have led us to several key findings that are summarized in the context of the following model (fig. 20).
Figure 20. Schematic model for events of June 15 climax as viewed along section from O¹Donnell valley (on left) across Mount Pinatubo and into upper Gumain valley (on right); see discussion of model in text. A, Conditions that prevailed during most of climactic eruption including plinian column, parts of which were collapsing to form pyroclastic flows. Single vent is shown in position of June 7-15 dome. Long arrows illustrate low-level air circulation into the convectively rising column. B, Conditions as caldera is foundering and single vent is destroyed; geometry of caldera is inferred. Large dots (not to scale) represent production of lithic clasts that were left near vent as lithic-rich facies. C, Postclimactic conditions as ash plumes rise from several vents in caldera floor. Lithic-rich facies (thickness exaggerated) locally mantle pumiceous pyroclastic-flow deposits (thickness approximately to scale) in proximal and medial areas.
Stratigraphic relations described in the previous section are consistent with broadly synchronous emplacement of pyroclastic-flow and pumice-fall deposits during much of the climactic eruption. Pyroclastic flows must therefore have been spawned either by (1) gravitational collapse of dense portions of a sustained plinian column that was simultaneously producing the pumice fall (fig. 20A) or by (2) alternating collapse and convective rise of the column, which is not depicted in figure 20. Hourly satellite observations during the climactic phase on June 15 (Koyaguchi and Tokuno, 1993) do not show changes in the stratospheric portion of the column directly above the vent that might be interpreted as reflecting extended periods of column collapse. Likewise, the normal grading of layer C (Paladio-Melosantos and others, this volume) suggests that the eruption column was not interrupted by prolonged periods of collapse. If the column had repeatedly collapsed and reformed over long time intervals, layer C should consist of alternating fine- and coarse-grained beds. Thus, if the column was alternating between collapse and convective rise, periods of collapse must have been brief (minutes?). Either origin for pyroclastic flows is different from that inferred initially by Scott and others (1991), who, from preliminary investigations prior to appreciable incision of the deposits, thought that pyroclastic flows largely followed deposition of layer C, as is consistent with classic models of late-stage column collapse (for example, Sparks and Wilson, 1976). The possibility that pyroclastic flows were generated simultaneously with the pumice fall should provide a challenge for modelers of eruption columns.
Broadly synchronous emplacement of fall and flow deposits for several hours and stratigraphic relations between stratified and massive pyroclastic-flow deposits imply that major fills of massive pyroclastic-flow deposits accumulated gradually. Regardless of whether flows were closely spaced discrete events or a quasisteady flow, we infer that multiple, relatively thin (tens of centimeters to less than a few meters?) beds of pyroclastic-flow material came to rest to form a thickening fill (progressive aggradation in the model of Branney and Kokelaar, 1992), even though little stratigraphic evidence for such emplacement is readily visible in the massive deposits themselves. The compositional and textural homogeneity of all flow deposits makes contact recognition difficult. In addition, the inflated gas-charged character of deposits would have encouraged erosion and mixing by subsequent units, further obscuring contacts. Other than pumice-concentration zones and contacts with overlying lithic-rich facies, the only abrupt contacts were formed in areas where substantial time elapsed between flow units to allow fine ash to settle out and the surface to firm up. Apparently only distal areas and some medial uplands met this requirement.
As noted in the discussion of massive pyroclastic-flow deposits, the content of fine ash (silt and clay size) in June 15 pyroclastic-flow deposits, 2 to 18 wt%, is lower than that of certain other comparable pumiceous pyroclastic-flow deposits (table 2). Such a low content of fine ash might partly explain the relatively low mobility of Pinatubo pyroclastic flows. Unless such fines depletion is a sampling artifact, why is fine ash apparently underrepresented in the Pinatubo pyroclastic-flow deposits? Perhaps a typical proportion of fine ash was not produced by fragmentation processes at Pinatubo, but such would be inconsistent with the high volatile content of the magma (Rutherford and Devine, this volume). Rather, this high volatile content ensured great fragmentation, high explosivity, and mixing with air, which together facilitated segregation of fine ash in the eruption column. Likewise, the great inflation, turbulence, and mobility inferred for pyroclastic flows in proximal areas promoted fine-ash segregation (fig. 20A). Such processes also would have produced relatively low emplacement temperatures of pyroclastic-flow deposits, as indicated by the absence of welding. The fine ash from pyroclastic flows could then have been swept by low-level circulation of air into the rapidly convecting column (for example, Fisher, 1979). Once entrained, fine ash would be taken to high altitudes, from which it could be dispersed far downwind.
The sudden and apparently short-lived production of abundant lithic clasts late in the emplacement of pyroclastic-flow deposits suggests that vent conditions changed markedly near the end of the eruption. Low lithic contents in the pumice-fall deposit (Paladio-Melosantos and others, this volume) and pumiceous pyroclastic-flow deposits suggest that little vent erosion occurred during most of the eruption. Drawing on the model of Druitt and Sparks (1984), we hypothesize that a large amount of lithic debris was created as the summit dome foundered (fig. 20B). Collapse would have disrupted the vent and forced fragmenting pyroclastic material to escape through growing fractures and faults in the old summit dome. Large amounts of dome rock would have been entrained during this process. Some of the dense lithic material settled through the flows near the vent, as predicted in various models for formation of co-ignimbrite lag deposits (Druitt and Sparks, 1982; Walker, 1985), and some traveled farther to form lithic-rich pyroclastic-flow deposits. In contrast to the radial distribution of pumiceous pyroclastic-flow deposits, lithic debris from the disintegrating summit was preferentially funneled through the deep canyons of major drainages that headed on the former summit dome (fig. 1). Reduced height of pyroclastic-flow formation in the eruption column, or perhaps a transition to low pyroclastic fountains issuing from numerous vents in the foundering caldera, may have localized the lithic-rich facies. In addition, the flows would have been denser than the earlier pumiceous pyroclastic flows and thus strongly channeled by topography.
The stratigraphic position of lithic-rich facies at or near the top of the sequence of pumiceous pyroclastic-flow deposits suggests that once caldera collapse began, pyroclastic-flow activity quickly waned, and the most intense phase of the eruption soon ended. Perhaps foundering rubble created a plug that largely sealed the vent and greatly reduced eruption rate. Williams and Self (1983) inferred that such a mechanism terminated the plinian eruption of Santa María, Guatemala, in 1902. Thus, unlike the Druitt and Sparks model (1984), caldera collapse at Pinatubo did not lead to subsequent emplacement of even greater volumes of pyroclastic-flow deposits. Instead, the eruptive activity waned over a period of weeks as ash issued from several vents in the caldera floor (fig. 20C). These ash emissions deposited layer D over broad areas southwest and northeast of the volcano (Paladio-Melosantos and others, this volume).
We infer that caldera collapse occurred during all or part of the intense, 6-h episode of large, regional, volcano-tectonic earthquakes (29 events between mb 4.5 and 5.7) on June 15 that began about 1530 (Mori, White, and others, this volume; Wolfe and Hoblitt, this volume). These events account for about 50% of the total energy released by earthquakes in the Pinatubo region during the 7-week period that began with this episode (fig. 21). Most of the energy release in this episode is represented by the two largest events, mb 5.5 (1841) and 5.7 (1915); the location of the second is at or very near the volcano. B.C. Bautista and others (this volume) conclude that the earthquakes of this episode, as well as most of the 7-week period, occurred on regional tectonic faults that were reactivated by the stress change induced by magma withdrawal. The earthquake episode began about 2 h into the climactic eruption (fig. 21); about 1 h later the high-amplitude phases of the infrasonic record ended (Tahira and others, this volume) and high-frequency, volcano-tectonic earthquakes began to dominate the seismic record (Power and others, this volume). Caldera collapse may have occurred at about this time (1630), thus emplacing the lithic-rich facies, ending the 3-h plinian eruption and initiating a 6-h period of diminishing activity as observed by geophysical techniques. The only primary deposit emplaced during this 6-h period would be the lower part of fine-grained, tephra layer D (Paladio-Melosantos and others, this volume).
Figure 21. Plot of cumulative seismic energy release for sequence of large earthquakes that occurred between about 1530 on June 15 and 0700 on June 16, 1991. Magnitudes were converted to energy release by using equation in Mori, White, and others (this volume), who estimated that total seismic energy release in the 7-week period beginning at 1530 on June 15 was 6.3x1013 J. Dark-gray band, 3-h period of high-amplitude infrasonic record (Tahira and others, this volume) and domination of seismic record by low-frequency earthquakes (Power and others, this volume). Crosshatched and light-gray bands, 6-h period of decreasing amplitude of barometric (Oswalt and others, this volume) and infrasonic records; boundary between these bands corresponds to marked drop in amplitude of Cubi Point barometric record.
Alternately, as permitted by stratigraphic relations (fig. 19), the change at 1630 might represent a shift to a low, persistently collapsing eruption column and continued generation of pyroclastic flows after pumice fall had ceased. Barograph records from Clark Air Base and Cubi Point (Oswalt and others, this volume) indicate a period of declining atmospheric-pressure disturbances after about 1630; the record from Cubi Point shows an additional drop in amplitude at about 1830, just as two large earthquakes dominate a period of conspicuous seismic-energy release (fig. 21). Perhaps caldera collapse occurred at about this time as the crust was responding markedly to the withdrawal of magma. Could these large tectonic earthquakes have even triggered collapse of the caldera? We have few constraints on such speculations, but if collapse occurred at about 1830, the eruption may have progressed through (1) a 3-h-long, high-intensity, plinian phase that emplaced the pumice-fall deposit and some pyroclastic-flow deposits, (2) a phase as long as 2 h characterized by continued generation of pyroclastic flows and ending with caldera collapse and emplacement of the lithic-rich facies of pyroclastic-flow deposits; accompanying tephra fall would be of co-ignimbrite origin and produce fine-grained deposits, and (3) a phase lasting several hours of declining infrasonic and barographic signals accompanied by ash emission from the caldera and continued accumulation of fine-grained tephra layer D. Thus, the emplacement of the 3.7 to 5.3 km3 (dense-rock equivalent) of pumice-fall deposits and pumiceous pyroclastic-flow deposits (tables 4 and 5) occurred in 3 to 5 h; an average volume eruption rate of 2x105 to 5x105 m3/s (mass eruption rate of 5-12x108kg/s).
Table 4. Volume relations (in rock or magma volume) among climactic eruptive products and topographic change caused by caldera collapse.
Table 5. Comparison of climactic 1991 Pinatubo eruption and Fierstein and Hildreth's (1992) reconstruction of episode I of 1912 Novarupta eruption, Alaska.
We know little about the caldera structure. But in figures 20B and C we infer a funnel-shaped structure (for example, Aramaki, 1984) because of the relatively small diameter of the caldera (2.5 km) compared to estimates of depth of the top of the magma chamber (6 km, Mori, Eberhart-Phillips, and Harlow, this volume; 7 to 8 km, Rutherford and Devine, this volume; 10 km, Pallister and others, this volume). One implication of caldera formation occurring while emplacement of pyroclastic flows was ending, as described above, is that thickly (hundreds of meters) ponded pyroclastic-flow deposits are probably not present in the caldera. Such a condition is consistent with the relatively low crater lake temperatures (Campita and others, this volume) and lack of widespread fumaroles, pumice, and secondary phreatic explosion craters on the caldera floor.
Topographic changes induced by caldera collapse roughly balance the volume of erupted products (table 4). The estimate of the topographic volume lost by caldera collapse (2.5 km3) is less than the combined volume, in dense-rock equivalent (or DRE), of pyroclastic-flow deposits (2.1-3.3 km3) and the pumice-fall deposit (1.6-2.0 km3; Paladio-Melosantos and others, this volume), which totals 3.7-5.3 km3. But, the rock mass that collapsed into the vent must have increased in volume, owing to dilation. The funnel-shaped collapse as depicted in figure 20C (cone with radius=1.25 km; height=10 km, to top of magma chamber) has a volume of about 16 km3. An estimate of 10% dilation (Aramaki, 1984) would account for an additional 1.6 km3 of volume loss and roughly balance erupted volume.
Four other explosive eruptions of this century emplaced bulk volumes of silicic pyroclastic deposits similar to or greater than Pinatubo's estimated 8.4 to 10.4 km3--1902 Santa María, Guatemala, 7.8 km3 (Fierstein and Nathenson, 1992) or 20.2 km3 (Williams and Self, 1983); 1912 Novarupta, Alaska, 28+-4 km3 (Fierstein and Hildreth, 1992); 1932 Volcán Quizapu, Chile, 9.5 km3 (Hildreth and Drake, 1992), and 1991 Cerro Hudson, Chile, 7.6 km3 (Scasso and others, 1994). Of these large-volume eruptions, the products of Santa María, Quizapu, and Hudson were virtually all plinian pumice-fall deposits, with only minor pyroclastic-flow deposits. The Novarupta eruption was most like the Pinatubo eruption in that both produced appreciable proportions of flow deposits. Moreover, the simultaneous emplacement of fall and flow deposits at Pinatubo is similar to events of episode I (phases A and B) of the Novarupta eruption as reconstructed by Fierstein and Hildreth (1992). Parallels between these events and their deposits are striking (table 5). Significant contrasts between the two events include (1) pronounced compositional changes, greater volume of products, and a more complex conclusion (episodes II and III) of the Novarupta eruption and (2) caldera collapse around the vent and emplacement of lithic-rich facies at Pinatubo late in the pyroclastic-flow sequence.
Much of the fieldwork upon which this study is based would not have been possible without the logistical and helicopter support provided by the U.S. Navy, U.S. Marine Corps, U.S. Air Force, and the Philippine Air Force. Numerous U.S. Geological Survey and Philippine Institute of Volcanology and Seismology colleagues shared important ideas, data, and feedback, of whom only a few can be thanked here: Glenda Besana and Jesse Daligdig worked on the initial volume estimates, John Pallister cooperated in much of the fieldwork in 1992 and contributed many insightful observations and suggestions, and Marvin Couchman performed the density determinations and most of the grain-size analyses. Karl Eriksen and Monty Pearson of the Portland District U.S.Army Corps of Engineers shared their volume estimates. Richard Waitt, Brian Hausback, Chris Newhall, and Ed Wolfe reviewed the manuscript and made many helpful suggestions.
Aramaki, Shigeo, 1984, Formation of Aira caldera, southern Kyushu, ~22,000 years ago: Journal of Geophysical Research, v. 89, p. 8485-8501.
Bautista, B.C., Bautista, M.L.P., Stein, R.S., Barcelona, E.S., Punongbayan, R.S., Laguerta, E.P., Rasdas, A.R., Ambubuyog, G., and Amin, E.Q., this volume, Relation of regional and local structures to Mount Pinatubo activity.
Booth, B., 1973, The Granadilla pumice deposit of southern Tenerife, Canary Islands: Proceedings of the Geological Association of London, v. 84, p. 353-370.
Branney, M.J., and Kokelaar, Peter, 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.
Campita, N.R., Daag, A.S., Newhall, C.G., Rowe, G.L., Solidum, R., this volume, Evolution of a small caldera lake at Mount Pinatubo.
Cas, R.A.F., and Wright, J.V., 1987, Volcanic successions, modern and ancient: London, Allen and Unwin Ltd., 528 p.
Cole, P.D., and Scarpati, C., 1993, A facies interpretation of the eruption and emplacement of the upper part of the Neapolitan Yellow Tuff, Campi Flegrei, southern Italy: Bulletin of Volcanology, v. 55, p. 311-326.
Druitt, T.H., and Bacon, C.R., 1986, Lithic breccia and ignimbrite erupted during the collapse of Crater Lake caldera, Oregon: Journal of Volcanology and Geothermal Research, v. 29, p. 1-32.
Druitt, T.H., and Sparks, R.S.J., 1982, A proximal ignimbrite breccia facies on Santorini, Greece: Journal of Volcanology and Geothermal Research, v. 13, p. 147-171.
------1984, On the formation of calderas during ignimbrite eruptions: Nature, v. 310, p. 679-681.
Fierstein, Judy, and Hildreth, Wes, 1992, The plinian eruptions of 1912 at Novarupta, Katmai National Park, Alaska: Bulletin of Volcanology, v. 54, p. 646-684.
Fierstein, Judy, and Nathenson, Manuel, 1992, Another look at the calculation of fallout tephra volumes: Bulletin of Volcanology, v. 54, p. 156-167.
Fisher, R.V., 1979, Models for pyroclastic surges and pyroclastic flows: Journal of Volcanology and Geothermal Research, v. 6, p. 305-318.
Folk, R.L., 1980, Petrology of sedimentary rocks: Austin, Hemphills, 184 p.
Heming, R.F., and Carmichael, I.S.E., 1973, High-temperature pumice flows from the Rabaul caldera, Papua New Guinea: Contributions of Mineralogy and Petrology, v. 38, p. 1-20.
Hildreth, Wes, and Drake, R.E., 1992, Volcán Quizapu, Chilean Andes: Bulletin of Volcanology, v. 54, p. 93-125.
Hoblitt, R.P., 1986, Observations of the eruptions of July 22 and August 7, 1980, at Mount St. Helens, Washington: U.S. Geological Survey Professional Paper 1335, 44 p.
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.
Inman, D.L., 1952, Measures for describing the size distribution of sediments: Journal of Sedimentary Petrology, v. 22, p. 125-145.
Koyaguchi, T., this volume, Volume estimation of tephra-fall deposits from the June 15, 1991, eruption of Mount Pinatubo by theoretical and geological methods.
Koyaguchi, Takehiro, and Tokuno, Masami, 1993, Origin of the giant eruption cloud of Pinatubo, June 15, 1991: Journal of Volcanology and Geothermal Research, v. 55, p. 85-96.
Lynch, J.S., and Stephens, G., this volume, Mount Pinatubo: A satellite perspective of the June 1991 eruptions.
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, Jim, White, R.A., and others, this volume, Volcanic earthquakes following the 1991 climactic eruption of Mount Pinatubo: Strong seismicity during a waning phase.
Oswalt, J.S., Nichols, W., and O'Hara, J.F., this volume, Meteorological observations of the 1991 Mount Pinatubo eruption.
Paladio-Melosantos, M.L., Solidum, R.U., Scott, W.E., Quiambao, R.B., Umbal, J.V., Rodolfo, K.S., Tubianosa, B.S., Delos Reyes, P.J., and Ruelo, H.R., this volume, Tephra falls of the 1991 eruptions of Mount Pinatubo.
Pallister, J.S., Hoblitt, R.P., Meeker, G.P., Knight, R.J., and Siems, D.F., this volume, Magma mixing at Mount Pinatubo: Petrographic and chemical evidence from the 1991 deposits.
Pierson, T.C., Daag, A.S., Delos Reyes, P.J., Regalado, M.T.M., Solidum, R.U., and Tubianosa, B.S., this volume, Flow and deposition of posteruption hot lahars on the east side of Mount Pinatubo, July-October 1991.
Pierson, T.C., Janda, R.J., Umbal, J.V., and Daag, A.S., 1992, Immediate and long-term hazards from lahars and excess sedimentation in rivers draining Mount Pinatubo, Philippines: U.S. Geological Survey Water-Resources Investigations Report 92-4039, 35 p.
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.
Riehle, J.R., 1973, Calculated compaction profiles of rhyolitic ash-flow tuffs: Geological Society of America Bulletin, v. 84, p. 2193-2216.
Riehle, J.R., Miller, T.F., McGimsey, R.G., and Keith, T.E.C., 1992, A compaction profile from the 1912 ash-flow sheet, Katmai National Park, Alaska [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 43, p. 636.
Rowley, P.D., MacLeod, N.S., Kuntz, M.A., and Kaplan, A.M., 1985, Proximal bedded deposits related to pyroclastic flows of May 18, 1980, at Mount St. Helens, Washington: Geological Society of America Bulletin, v. 96, p. 1373-1383.
Rutherford, M.J., and Devine, J.D., this volume, Preeruption pressure-temperature conditions and volatiles in the 1991 dacitic magma of Mount Pinatubo.
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.
Scasso, R.A., Corbella, Hugo, and Tiberi, Pedro, 1994, Sedimentological analysis of the tephra from the 12-15 August 1991 eruption of Hudson volcano: Bulletin of Volcanology, v. 56, p. 121-132.
Scott, W.E., Hoblitt, R.P., Daligdig, J.A., Besana, G., and Tubianosa, B.S., 1991, 15 June 1991 pyroclastic deposits at Mount Pinatubo, Philippines [abs]: Eos, Transactions, American Geophysical Union, v. 72, p. 61-62.
Sheridan, M.F., 1971, Particle-size characteristics of pyroclastic tuffs: Journal of Geophysical Research, v. 76, p. 5627-5634.
Smith, R.L., 1960, Ash flows: Geological Society of America Bulletin, v. 71, p. 795-842.
Sparks, R.S.J., 1976, Grain-size variations in ignimbrites and implications for the transport of pyroclastic flows: Sedimentology, v. 23, p. 147-188.
Sparks, R.S.J., Self, S., and Walker, G.P.L., 1973, Products of ignimbrite eruptions: Geology, v. 1, p. 115-118.
Sparks, R.S.J., and Wilson, Lionel, 1976, A model for the formation of ignimbrite by gravitational column collapse: Journal of the Geological Society of London, v. 132, p. 441-452.
Tahira, M., Nomura, M., Sawada, Y., and Kamo, K., this volume, Infrasonic and acoustic-gravity waves generated by the Mount Pinatubo eruption of June 15, 1991.
Torres, R.C., Self, S., and Martinez, M.L., this volume, Secondary pyroclastic flows from the June 15, 1991, ignimbrite of Mount Pinatubo.
Umbal, J.V., and Rodolfo, K.S., this volume, The 1991 lahars of southwestern Mount Pinatubo and evolution of the lahar-dammed Mapanuepe Lake.
Valentine, G.A., 1987, Stratified flow in pyroclastic surges: Bulletin of Volcanology, v. 49, p. 616-630.
Walker, G.P.L., 1981, Generation and dispersal of fine ash and dust by volcanic eruptions: Journal of Volcanology and Geothermal Research, v. 11, p. 81-92.
------1983, Ignimbrite types and ignimbrite problems: Journal of Volcanology and Geothermal Research, v. 17, p. 65-88.
------1985, Origin of coarse lithic breccias near ignimbrite source vents: Journal of Volcanology and Geothermal Research, v. 25, p. 157-171.
Walker, G.P.L., Wilson, C.J.N., and Froggatt, P.C., 1981, An ignimbrite veneer deposit: The trail-marker of a pyroclastic flow: Journal of Volcanology and Geothermal Research, v. 9, p. 409-421.
Williams, S.N., and Self, Stephen, 1983, The October 1902 plinian eruption of Santa María volcano, Guatemala: Journal of Volcanology and Geothermal Research, v. 16, p. 33-56.
Wilson, C.J.N., and Walker, G.P.L., 1982, Ignimbrite depositional facies: The anatomy of a pyroclastic flow: Journal of the Geological Society of London, v. 139, p. 581-592.
Wilson, Lionel, and Head, J.W., 1981, Morphology and rheology of pyroclastic flows and their deposits, and guidelines for future observations, 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. 513-524.
Wolfe, E.W. and Hoblitt, R.P., this volume, Overview of the eruptions.
FIRE and MUD Contents
PHIVOLCS | University of Washington Press | U.S.Geological Survey
This page is <http://pubs.usgs.gov/pinatubo/wescott/>
Contact: Chris Newhall
Last updated 06.11.99