FIRE and MUD Contents

Watershed Disturbance and Lahars on the East Side of Mount Pinatubo During the mid-June 1991 Eruptions

By Jon J. Major,1 Richard J. Janda,1 3 and Arturo S. Daag2

1U.S. Geological Survey.

2Philippine Institute of Volcanology and Seismology.

3Deceased.


ABSTRACT

Densely populated alluvial fans on the east side of Mount Pinatubo were affected by widespread lahars triggered during the June 15, 1991, eruptions. The lahars were triggered by heavy rainfall. However, the lahars were not generated because of uncommonly heavy precipitation, but rather because of radical alteration of watershed hydrology by volcanic deposits in conjunction with heavy rainfall. Fine-grained fall and surge deposits related to eruptions that preceded the climactic-phase activity damaged vegetation, reduced the infiltration capacity of hillslope surfaces, and smoothed the natural-scale hillslope roughness. These effects led to enhanced overland flow that instigated hill-slope and channel erosion, triggered minor slope failures, and initiated the peak-discharge lahars. Sediment mobilized by rilling and shallow landsliding of the mantle of pumice tephra deposited by the climactic-phase eruption contributed to lahars interbedded with pyroclastic valley fill and to pumice-bearing recessional flow that followed peak-discharge.

The peak-discharge lahars that flooded fan channels in Tarlac and Pampanga Provinces varied in rheology, magnitude, and timing. Below fanheads, lahars were dominantly hyperconcentrated streamflow or flow transitional to hyperconcentrated streamflow; above fanheads, lahars were dominantly debris flow. Along the mountain front from the Gumain River to the Sacobia River, peak discharge preceded deposition of broadly distributed plinian pumice fall; between the Sacobia River and the O'Donnell River, peak discharge largely followed pumice fall. Variations in rheology, magnitude, and timing of peak flows reflect the interplay of variations in rainfall intensity and timing, variations in the degree of watershed disturbance, and characteristic response times of the watersheds. Interplay of these factors affects the amount of water that moves from hillslopes to channels to fanheads, how rapidly it moves, and how easily sediment is entrained and transported.

Damage to habitation and infrastructure on the alluvial fans resulted primarily from lateral bank erosion and from aggradation of mainstem channels that induced backflooding of tributary channels. Overbank flooding by lahars was generally localized and of secondary importance except along distal alluvial plains, where primarily agricultural land was inundated.

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

INTRODUCTION

Heavy rainfall during the explosive eruptions of Mount Pinatubo in mid-June 1991 produced lahars that damaged habitation, infrastructure, and arable land on densely populated alluvial fans that flank the eastern sector of the volcano and the surrounding terrain. These lahars had sediment and fluid-dynamic characteristics very similar to lahars that have formed as a result of explosive eruptions at snow-clad volcanoes, for example during the 1989-90 eruptions of Redoubt Volcano, Alaska (Alaska Volcano Observatory, 1990; Brantley, 1990; Dorava and Meyer, 1994), the 1980 and 1982 eruptions of Mount St. Helens (Janda and others, 1981; Waitt and others, 1983; Pierson and Scott, 1985; Major and Voight, 1986; K.M. Scott, 1988), and during the 1971 eruption of Hudson Volcano, Chile (Best, 1992), despite having markedly different initiation mechanisms. However, unlike lahars generated during eruptions of snow-clad volcanoes, where interaction among volcanic products and snow and ice (for example, Major and Newhall, 1989; Trabant and others, 1994) can greatly increase the volume of water typically delivered to watershed channels, the lahars at Mount Pinatubo were generated by commonplace, albeit heavy, tropical rainfall. Although a typhoon crossed central Luzon island coincident with the June 15 eruption, that typhoon was not especially intense (Oswalt and others, this volume). The lahars concurrent with the June 15 eruptions of Mount Pinatubo were not the result of an uncommon meteorological event coincident with the eruption but were, as we will argue, the result of a fundamental alteration of watershed hydrology that acted in conjunction with heavy rainfall.

The purposes of this paper are to document and discuss the effects of, and the deposits of, lahars that occurred on the east side of Mount Pinatubo, in Tarlac and Pampanga Provinces (fig. 1), before, during, and shortly after the climactic June 15, 1991, eruption, and to explore the central cause of these lahars. In particular, we discuss the sedimentology and behavior of lahars generated in several watersheds, the relation between the lahars and tephras deposited by the mid-June eruptions, and the geomorphological consequences of those lahars.

Figure 1. Drainage network around Mount Pinatubo. Cited outcrops are indicated. Villages in parentheses were destroyed during the eruption.

Annual monsoon rains have mobilized vast amounts of sediment and generated numerous lahars that have greatly modified the east-side alluvial fans since the June 1991 eruptions (Pierson and others, this volume; K.M. Scott and others, this volume; Dolan, 1993). This paper discusses lahar deposits and channel conditions as they existed in late June to mid-July 1991. Many deposits documented in this paper have been eroded or have been buried by as much as several tens of meters of sediment transported by subsequent lahars. Geographically, this paper will cover an area bounded on the north by the O'Donnell-Tarlac River and on the south by the Porac River (fig. 1). We do not discuss the Gumain River system. Papers by Rodolfo and others (this volume) and Umbal and Rodolfo (this volume) discuss lahars affecting basins west of Mount Pinatubo.

PHYSIOGRAPHIC SETTING

Mount Pinatubo is located in the Philippines approximately 100 km northwest of Manila in central Luzon (fig. 1). The volcano stood 1,745 m above sea level prior to the June 15 eruption; the highest point on its crater rim now stands about 1,485 m above sea level (Jones and Newhall, this volume). The volcano is surrounded by steeply dissected mountains to the west, south, and northeast. The provinces of Tarlac and Pampanga are drained by the O'Donnell-Tarlac, Sacobia-Bamban, Abacan, Pasig-Potrero, Porac, and Gumain watersheds (fig. 1). Prior to the June 1991 eruptions, the Abacan and Pasig-Potrero Rivers did not head on the flanks of the volcano; their watersheds were developed in steeply dissected terrain that fronted Mount Pinatubo. Above 1,000 m altitude, watersheds had hillslope angles that generally ranged from nearly 20° to more than 45°; channel gradients commonly were about 6° to 18°. Between altitudes of about 1,000 m and 200 m, channel gradients flattened to about 1°. Below 200 m, broad, coalescing alluvial fans having low gradients (fig. 2) that range from about 0.02 m/m at fan heads to less than 0.0002 m/m along distal alluvial plains dominate the topography (Pierson and others, 1992; this volume). Stream channels generally were well incised above fanheads; on the alluvial fans, channels were only mildly confined, except where artificially manipulated, by river banks less than a few meters high. Drainage basin characteristics above fanheads are provided in table 1.

Figure 2. Preeruption longitudinal profiles of major river channels east of Mount Pinatubo that conveyed lahars during the June 15, 1991, eruption. Horizontal distance is measured from the channel head.

Table 1. Characteristics of drainage basins above fanheads east of Mount Pinatubo.

TERMINOLOGY

The term lahar can mean many things to many people; the literature is replete with confusing, and sometimes conflicting, usage. Participants of a recent Geological Society of America Penrose Conference proposed the following usage (Smith and Fritz, 1989): "...lahar: a general term for a rapidly flowing mixture of rock debris and water (other than normal streamflow) from a volcano. A lahar is an event; it can refer to one or more discrete processes but does not refer to a deposit." This usage allows a single event to encompass a variety of rheological behaviors. In the following discussion we shall emphasize the effects, behavior, and consequences of two types of highly concentrated suspensions of sediment and water: debris flows, defined as flowing mixtures of sediment and water that commonly consist of a broad distribution of grain sizes and have sediment concentrations that usually exceed about 60 percent by volume, and hyperconcentrated flows, defined as flowing mixtures of sediment and water commonly composed of a narrower distribution of solid particles, dominantly sand sized, which commonly have sediment concentrations that range between 20 and 60 percent by volume (for example, Beverage and Culbertson, 1964; Pierson and Scott, 1985; Smith and Lowe, 1991; Pierson and others, this volume). Sedimentologic characteristics of deposits inferred to have resulted from these processes are illustrated in figure 3.

Figure 3. Deposit facies associated with lahars (from Pierson and Scott, 1985). Debris-flow deposits commonly are poorly sorted and have broad distributions of grain size; the mean grain size of the nongravel fraction commonly ranges from about 1 mm to more than 4 mm (very coarse sand to pebbles). Hyperconcentrated streamflow deposits commonly are massive to slightly laminated and have a narrower distribution of grain sizes; the mean grain size commonly ranges from about 0.25 mm to 0.5 mm (medium to coarse sand).

Wentworth sediment-size terms (silt, sand, gravel) are used to describe the textures of deposits. Division of the Wentworth size terms into finer designations is well established (Folk, 1974). Pyroclastic size terms (ash, lapilli, bombs), on the other hand, represent broad categories of grain size that are insufficiently divided into finer designations (Fisher and Schminke, 1984). The use of Wentworth sediment-size terms may irritate some readers, especially when the terms are applied to pyroclastic deposits. However, the Wentworth terminology accurately describes the texture of sediments, is unambiguous, and holds no genetic connotations.

THE EXPLOSIVE ERUPTIONS AND THEIR DEPOSITS

CHRONOLOGY OF EVENTS AND CHARACTERISTICS OF DEPOSITS

The first in a series of powerful explosions occurred on June 12, 1991 (Philippine Volcano Observatory Team, 1991; Hoblitt, Wolfe, and others, this volume). Explosive pulses during the early part of this phase of activity produced pyroclastic flows that traveled up to 6 km from the old summit on the north, northwest, and west flanks of the volcano. Small-volume pyroclastic flows also descended the south and east flanks of the volcano. Sandy tephra fell mainly north to southwest of the volcano; little if any tephra fell to the east during these explosions (Hoblitt, Wolfe, and others, this volume). Eruptions on the afternoon of June 14 marked the end of a phase of strong vertical eruptions and the beginning of activity dominated by laterally directed blasts (PVO Team, 1991; Hoblitt and others, 1991; Hoblitt, Wolfe, and others, this volume; Wolfe and Hoblitt, this volume). Several of these blasts produced various pyroclastic density currents that traveled as much as 10 km from the vent, and heavy fall that blanketed many sectors of the surrounding terrain (Hoblitt, Wolfe, and others, this volume; W.E. Scott and others, 1991; this volume). Explosive activity culminated about 1400 (all times listed are local time) on June 15 with a sustained pumice-rich, climactic eruption that persisted into the evening (Hoblitt and others, 1991; Hoblitt, Wolfe, and others, this volume). Following the climactic phase of activity, voluminous ash emission continued, and slowly decreased, over a period of several weeks.

Explosive activity on June 15 produced voluminous deposits of tephra fall and of various pyroclastic density currents (W.E. Scott and others, 1991; this volume; Paladio-Melosantos and others, this volume). Pyroclastic currents swept all sectors to a maximum distance of 16 km from the vent. Pyroclastic-flow deposits accumulated chiefly between 5 and 15 km of the vent and broadly buried upland channels to depths ranging from a few centimeters to more than 200 m; large areas are buried to an average thickness of 30 to 50 m (W.E. Scott and others, 1991; this volume). On the east flank of Mount Pinatubo, mainstem valleys in the headwaters of the O'Donnell, Sacobia, and Pasig River watersheds are deeply filled (table 1; W.E. Scott and others, this volume, figs. 1 and 2). Valleys northeast and southeast of the volcano contain relatively minor pyroclastic-flow deposits (W.E. Scott and others, 1991).

Tephra-fall and surge deposits associated with preclimactic eruptive activity are chiefly fine grained (fig. 4). East of the volcano these deposits consist primarily of normally graded fine sand to silt (Hoblitt, Wolfe, and others, this volume). The climactic phase of the eruption, on the other hand, produced a dominantly coarse, granular pumice fall (W.E. Scott and others, 1991; Paladio-Melosantos and others, this volume). East of the volcano this fall deposit consists chiefly of a normally graded bed of pumiceous and minor lithic small pebbles, granules, and sand, depending on proximity to the cone (fig. 4). Tephra-fall thickness east of the volcano ranges from about 1 to 5 cm distributed broadly across the distal alluvial fans to about 50 cm on slopes in upland areas (see figs. 3, 4 of Paladio-Melosantos and others, this volume). Pumiceous pyroclastic flows generated by the climactic eruption are synchronous with, or lie stratigraphically above, the pumice fall (W.E. Scott and others, this volume).

Figure 4. Cumulative grain-size distributions of the fine-grained preplinian and plinian tephra-fall deposits east of Mount Pinatubo. Data from R.P. Hoblitt (written commun., 1993) and Paladio-Melosantos and others (this volume).

WATERSHED DISTURBANCE

Assorted volcanic processes disturbed watersheds in the vicinity of Mount Pinatubo to varying degrees. Pyroclastic-flow deposits are more extensive in watersheds west of Mount Pinatubo than in watersheds to the east (W.E. Scott and others, this volume). Minor volumes of pyroclastic debris were deposited in the upper O'Donnell and Sacobia watersheds on June 11 (Hoblitt, Wolfe, and others, this volume). Activity on June 12 and 13 did not affect east-side watersheds. Small explosions and low fountaining of ejecta on the afternoon of June 14 dispersed pyroclastic density currents into the O'Donnell basin. Many subsequent explosions leading to the climactic activity dispersed pyroclastic density currents radially and deposited sediment from pyroclastic surges in nearly all watersheds.

Thicknesses of surge deposits in catchments are not well known. Proximal surge deposits, within 10 km of the old summit, are as much as a meter thick; individual beds of distal surge deposits typically range from nearly 0 to a few centimeters thick (Hoblitt, Wolfe, and others, this volume).

Tephra fall was relatively evenly distributed across all sectors. The first broadly distributed, fine-grained fallout to affect the east side of Mount Pinatubo was associated with an eruption at about 1500 on June 14 (Hoblitt, Wolfe, and others, this volume). Fine-grained tephra associated with preclimactic eruptions on the morning of June 15 was also broadly distributed to the east and obscured views of the volcano from early morning until after the climactic eruption.

Pyroclastic flows associated with the climactic-phase eruption extended down several mainstem channels on the east side, deeply filling the upper valleys (W.E. Scott and others, this volume). Plinian pumice fall on the east side is commonly 10 to 50 cm thick in the upper basins (Paladio-Melosantos and others, this volume). We will show, however, that peak discharge of several of the lahars that inundated channels on the east side of Mount Pinatubo occurred before deposition of the plinian pumice fall and the extensive pyroclastic valley fill. The pyroclastic-surge deposits that accumulated in the 24-h period from the afternoon of June 14 until the beginning of the climactic eruption on the afternoon of June 15 were responsible for altering watershed hydrology and establishing conditions conducive to the generation of lahars once rainfall commenced.

RAINFALL

The climate of central Luzon is characterized by distinct wet and dry seasons. The climate generally is dry when dominated by easterly flow of the North Pacific trade winds. Between May and October, however, the moist, tropical southwest monsoon dominates (Pierson and others, this volume), and many typhoons commonly pass across, or near, the Philippines and draw in even more moist air (Pierson and others, this volume). More detailed discussions of the central Luzon climate are found in Pierson and others (1992, this volume), Rodolfo (1991), Rodolfo and others (this volume), and Umbal and Rodolfo (this volume).

On June 11 and 12, 1991, a small tropical depression over the Pacific Ocean east of the Philippines developed into Typhoon Yunya (Oswalt and others, this volume). By 0800 on June 14, the typhoon, still at sea southeast of central Luzon, packed winds that peaked at 195 km/hr; nevertheless, Yunya was considered only a modest-sized typhoon (Oswalt and others, this volume). Twenty-four hours later, the typhoon went ashore about 90 km east of Clark Air Base (fig. 5). The typhoon tracked northwesterly and rapidly lost organization as it passed over land and began interacting with the eruptive column from Mount Pinatubo (Oswalt and others, this volume). By 1200 on June 15 it passed within 60 km northeast of Clark, and by about 1700 the typhoon, then downgraded to a tropical storm, passed Baguio, 150 km northeast of Mount Pinatubo. Yunya exited the Philippines approximately 185 km northwest of Clark at about 2000.

Figure 5. Path of Typhoon Yunya (after Oswalt and others, this volume). Circles with dots indicate tropical depression; unfilled circles indicate tropical storm; and filled circles indicate typhoon. Date and local time are given as 15/0800H, meaning 0800 on June 15, 1991.

Rainfall from Typhoon Yunya triggered several lahars at Mount Pinatubo; however, the amounts and intensities of rainfall produced by the typhoon in upland areas are unknown. Measurements at Clark indicated that 150 mm of rain fell between the mornings of June 15 and June 17, but the onset and intensities of the rainfall are not known (Maj. W. Nichols, U.S. Air Force, oral commun., 1991). Eyewitnesses report that rain did not begin to fall at Clark until late-morning on June 15 and that heavy rainfall did not begin until around 1400 (R.P Hoblitt, U.S. Geological Survey, oral commun., 1993). Distributing this rainfall evenly over 2 days yields a minimum rainfall intensity of 75 mm per day. Daily rainfall of this magnitude is not uncommon during the southwest monsoon. Rainfall data collected at Clark from June 20 to September 9, 1991, show that daily rainfall exceeded 70 mm once, and on 6 days exceeded 60 mm (U.S. Air Force, unpub. data, 1991). Despite ambiguities of onset and intensity, we infer that the rainfall associated with the typhoon was not an exceptional meteorological event.

Rainfall in central Luzon is characterized by high spatial and temporal variability. Records from Magalang and San Fernando (fig. 1) illustrate the variability of rainfall across distal reaches of alluvial fans east of Mount Pinatubo from June 12 to 17, 1991 (table 2). Between June 14 and 16, precipitation at these stations ranged from 8.2 to 42.0 mm per day. Two-day moving averages of the rainfall collected at the gages between June 14 and 17 range from 13.8 to 25.1 mm per day, far less than the estimated 75 mm per day at Clark during the same period. Precipitation on lowland alluvial fans has little bearing, however, on precipitation responsible for triggering lahars in upland catchments, owing to orographic and rain-shadow effects. This has been illustrated clearly elsewhere, particularly in tropical climates in which the geomorphic setting is characterized by rugged topography (for example, Giambelluca and others, 1984; MacArthur and others, 1992). A study of rainfall-triggered lahars at Mayon volcano, southern Luzon, showed that orographic effects enhanced the rainfall accumulation on the slopes of the volcano commonly more than two, and sometimes as much as five, times the accumulation recorded on lowland alluvial fans only 12 km away (Rodolfo and Arguden, 1991). That study also showed that rainfall intensity, and, particularly, the short-term intensity of rainfall that accumulates during the most intense 10 min, were important factors in triggering lahars at Mayon.

Intensities, durations, and amounts of precipitation in the upper basins of the major river systems that drain Mount Pinatubo probably exceeded those recorded on the distal fans. Intense convective storm cells commonly formed in close spatial and temporal association with localized volcanic heat sources prior to the June 15 eruptions (Pierson and others, 1992; Oswalt and others, this volume), and it is likely that similar locally developed storm cells were induced by pyroclastic debris in catchment headwaters and by the large convecting plume of ash discharging from the volcano on June 15. As an example of the intensity of rain that can fall in the upper catchments, we cite a storm that occurred on July 9, 1991. A rain gage installed at approximately 1,060 m in altitude near Mount Cuadrado in the Gumain River watershed (fig. 1) recorded 25.4 mm of rain 13 min after the onset of a local storm and 50.8 mm of rainfall after 60 min (R.J. Janda, U.S. Geological Survey, unpub. data, 1991). Four rain gages within or close to the east-side watersheds, installed subsequent to the June 1991 eruptions, show large variations in measured daily rainfall during the rainy season (Pierson and others, this volume). In fact, high-intensity rainfall was sufficiently localized in July, August, and September 1991 that lahars commonly occurred when little or no rainfall was recorded at one or more of the rain gages (Pierson and others, this volume). Although we have no information on rainfall in the upper basins during the June 15 eruptions, we infer that the rainfall intensity was least 75 mm/day and probably substantially greater.

Table 2. Daily precipitation (in millimeters) measured at stations in the vicinity of Mount Pinatubo, Pampanga Province.

THE LAHARS: CHARACTERISTICS, TIMING, AND GEOMORPHIC CONSEQUENCES

Owing to variations in the volcanic processes that affected watersheds, to variations in the degree of disturbance by those processes, and to variations in rainfall intensity and timing, the lahars that inundated channels on the eastern flank of Mount Pinatubo varied in rheology, magnitude, and timing. The downstream geomorphic consequences of the lahars are directly related to characteristics and timing of the lahars, which in turn are directly linked to initiation mechanisms, sediment characteristics, and watershed hydrology.

O'DONNELL-TARLAC RIVER SYSTEM

Deposits in the lower O'Donnell-Tarlac system record passage of several lahars having characteristics of hyperconcentrated flow or of flow undergoing transition from debris flow to hyperconcentrated flow (fig. 6; table 3). At an irrigation dam near Maniknik (location O1, fig. 1), three thin, sharply defined deposits of pebbly to pebbly-muddy sand support pumice and (or) lithic pebbles. The basal deposit contains almost exclusively pumice pebbles, whereas the upper two deposits are dominantly lithic rich. The middle unit ranges widely in thickness, is discontinuous, and exhibits soft sediment deformation, indicative that it was not sufficiently consolidated before the overlying unit was deposited. The sequence of lahar deposits overlies a discontinuous deposit of June 15 pumice fall, which in turn lies directly on preeruption alluvium.

Figure 6. Stratigraphic sections of the mid-June lahar and related deposits along river channels east of Mount Pinatubo. Values (45 km) indicate the distance from the former summit of the volcano; code names (PP1) indicate cited outcrops (see fig. 1). Sample numbers refer to data in table 3. The line between sections correlates the position of the plinian pumice fall.

Table 3. Characteristics of mid-June 1991 lahar deposits on the eastern sector of Mount Pinatubo.

By the time the lahars reached Tarlac, 8 km farther downstream (location O2, fig. 1), they had dropped nearly all of their gravel in the channel and were evolving toward hyperconcentrated and (or) sediment-laden streamflow (fig. 6; table 3). Here, flood-plain deposits of slightly pebbly to slightly pebbly-muddy sand support few clasts larger than 3 cm diameter, although scattered pumice cobbles as large as 20 cm were present on the surface. These deposits exhibit a greater variety of sedimentary textures than those upstream, ranging from massive and poorly sorted to well laminated and trough crossbedded. Fall deposits were not observed within this section.

Precisely when lahars flooded the lower O'Donnell River is ambiguous. Residents near Maniknik report that the first notable flow arrived between the evening of June 15 and early on June 16. Residents 15 km upstream report arrival of the first flow at 1900 on June 15 and subsequent flows on June 18. All eyewitness reports indicate several rises in stage having durations of 1 to 2 h, with intervening recessions of nearly 1 h. All accounts consistently report that multiple waves of flow swept the channel within a span of several hours and that the first flow did not reach the lower O'Donnell River until well after the onset of climactic eruptive activity.

SACOBIA-BAMBAN RIVER SYSTEM

Exposures in the Sacobia-Bamban valley reveal passage of several lahars prior to, during, and well after onset of the climactic phase of eruptive activity. The deposits suggest that flows progressively diluted downvalley and that the peak-discharge lahar prior to the onset of climactic activity either did not extend far onto the alluvial fan, deposited its sediment load entirely within the active channel, or diluted to sediment-laden streamflow that passed easily through the distal fan environment.

Incision in valley fill adjacent to Clark Air Base (location S1, fig. 1) reveals multiple debris-flow deposits beneath climactic-phase pyroclastic-flow deposits. Exposures also show that debris flows occurred before and after deposition of the pumice fall (fig. 6). The basal debris-flow deposit, of which 40 cm was exposed in early July 1991, consisted of a sandy gravel that contained dispersed lithic pebbles as large as 4 cm (sample S1D, fig. 6, table 3). By late July, channel incision exposed more than 4 m of that basal deposit. Deeper in the section, it consisted of dark, lithic boulders dispersed in a lithic sand matrix (K.M. Scott, U.S. Geological Survey, written commun., 1993).

The basal debris-flow deposit is separated from an overlying debris-flow deposit by more than a centimeter of gray laminated silt that represents fall and surge deposits associated with eruptions that occurred during the 24 h preceding the climactic eruption (Hoblitt, Wolfe, and others, this volume; W.E. Scott and others, this volume). Until mid-morning of June 15, little widespread tephra had fallen to the east; however, density currents easily transported fine-grained sediments along confined channels to distances in excess of 10 km (Hoblitt, Wolfe, and others, this volume). Judging by the thickness of silt that overlies the basal debris-flow deposit, we conclude that the debris flow occurred before mid-morning of June 15. R.P. Hoblitt (U.S. Geological Survey, oral commun., 1991) observed a hot lahar flowing along the Sacobia River near Clark Air Base on the evening of June 14, and we believe the basal deposit in this section records that lahar.

A debris flow that occurred between the evening of June 14 and early afternoon on June 15 deposited sediment that lies between silt layers below the pumice fall. Because this deposit lies between a relatively thick silt at its base and only a couple of millimeters of silt at its top, we speculate that it records flow along the valley during the late morning to early afternoon on June 15. This sandy gravel deposit, which is slightly finer grained than the upper part of the underlying debris-flow deposit and which locally thickens and truncates along channels eroded into the underlying deposit, has sedimentary characteristics that suggest that it resulted from a debris flow that may have been evolving toward hyperconcentrated streamflow (figs. 3, 6; samples S1A,B,C, table 3).

A several-meter-thick sequence of pumiceous pyroclastic-flow deposits overlies the pumice fall in the lower Sacobia valley. Within that sequence of typically loose, silty sediment is a deposit of well-indurated, poorly sorted, ungraded, pumiceous silty sand that contains dispersed inflated pumice pebbles around which are localized zones of fines-depleted sand. This deposit fills shallow channels incised into underlying debris and truncates locally over minor topographic rises. These deposits are found near the distal end of the pyroclastic fill in the Sacobia valley. We find it unlikely that the pyroclastic flows at this distance would be particularly erosive, especially on such low channel gradients. Indeed, a 3-m-thick pyroclastic-flow deposit overlies easily erodible pumice fall at this site. We conclude that water eroded the channels observed within the sequence and that the well-indurated deposit that fills those channels resulted from shallow slurries of reworked pyroclastic debris. Clearly, significant quantities of water were moving off hillslopes and through valleys during the emplacement of the pyroclastic valley fill.

Peak discharge inundated the flood plain on the lower Sacobia River near Dolores (Mabalacat Municipality) (location S2, fig. 1) during early morning to early afternoon on June 15. The flood-plain deposit from that flow consists of massive slightly pebbly sand that is moderately laminated at its base (fig. 6). It underlies pumice fall and overlies several millimeters of gray silt. Stratigraphic position of the peak-flow deposit between fall deposits of June 15 eruptions constrains the timing of the event. Deposit textures suggest the peak flow wave gradually diluted downvalley; here it was composed of a relatively dilute leading edge followed by a hyperconcentrated wave.

Lahars from the Sacobia River valley passed from a naturally confined channel along the mountain front and across the alluvial fan into the artificially contained Bamban River on the alluvial plain. Although lahars generally remained within the containment structures along the upper reaches of the plain, they broadly inundated arable land on the distal plains. Deposits of two lahars that inundated the flood plain are exposed near San Francisco (location B1, fig. 1), nearly 40 km from the volcano's former summit. Here, deposits consist of generally thin beds (<50 cm thick) of lithic pebbly to slightly pebbly sand (fig. 6; table 3) inferred to be deposited by hyperconcentrated flows or by flows transitional to hyperconcentrated flow. These deposits overlie pumice-rich alluvium that mantles preeruption lithic alluvium.

Conspicuously absent from the Bamban River exposure are the pumice fall and prepumice-fall lahar deposits correlative with the lithic-rich peak-flow deposits in the Sacobia valley. Beyond the containment dikes, an 8-mm-thick pumice-fall deposit is preserved. Eyewitnesses report that the pumice-rich alluvium at the base of the section was deposited on the evening of June 15 and that superposed sediments were deposited later that evening, about 2300 h. Streamflow that deposited the pumice alluvium possibly eroded and reworked the pumice fall within the confined reach. Absence of prepumice-fall deposits on the flood plain may indicate that the earlier peak-discharge lahar, recorded in the Sacobia valley, did not get this far down valley (unlikely), that it deposited its sediment load broadly within the active channel, or that it transformed completely to sediment-laden streamflow that passed easily through the distal fan environment and onto the alluvial plains, leaving little recognizable sediment.

Flow velocity of at least one lahar was estimated from the geometry of a mudline preserved on the concrete piers of the San Francisco bridge. The mudline extended nearly 3 m above pier footings on the upstream side of the bridge and nearly 2 m above footings on the downstream side. Eyewitnesses report that a pumice-rich flood scoured the channel slightly, exposing the bridge footings, and that a later flow left the mudline. From the mudline geometry, we infer that at least one of the later lahars had an instantaneous minimum wave velocity of about 4 to 4.5 m/s (see Major and Iverson, 1993). Mean flow velocity and mean discharge are not known, however.

ABACAN RIVER SYSTEM

Stratigraphic relations among deposits of pyroclastic flows, ash-cloud surges, lahars, and tephra fall in the upper Abacan watershed record sustained rainfall runoff and active formation of lahars concurrent with emplacement of the pyroclastic valley fill. Despite stratigraphic complexities, deposits near a gap in the divide that separates the Sacobia River valley from the Abacan watershed (location A1, figs. 1, 7) indicate that pyroclastic activity associated with the June 15 eruption was sustained for many hours. Near-surface pyroclastic deposits, which represent the latter phases of that sustained activity as well as secondary pyroclastic flows (W.E. Scott and others, this volume), lie stratigraphically above the pumice fall but below the silty fall of waning-phase eruptive activity. Debris-flow deposits, composed of well-indurated, poorly sorted, pumiceous, muddy sandy gravels that have vesiculated matrices (table 3), are found interbedded with pyroclastic deposits, filling depressions and thinning over topographic rises. One extensive debris-flow deposit mantled the deposits of all but the last pyroclastic flow to spill through the gap. Morphology, stratigraphy, and sedimentology of deposits in the upper Abacan watershed show that several lahars were concurrent with pyroclastic activity. However, none of these lahar deposits record the peak flow that passed down the Abacan River channel, which occurred prior to the deposition of the pyroclastic valley fill.

Figure 7. Aerial view of the upper and lower gap in the watershed divide between the Sacobia and Abacan River valleys. The view is looking from the Abacan watershed toward the Sacobia valley. Note the thick pyroclastic fill that spilled through the gap. Photograph by R.P. Hoblitt, March 18, 1992.

Near Sapangbato (fig. 1), adjacent to Clark Air Base, two tributaries join to form the mainstem Abacan River. Stratigraphic relations indicate that lahar deposition along the northern tributary occurred primarily after the pumice fall; however, isolated areas from which channel alluvium was stripped to bedrock and subsequently replaced by pumice fall attest to an important phase of prepumice-fall activity. These stratigraphic relations contrast sharply with those observed in the southern tributary channel. In that channel the peak-discharge lahar deposit contains almost no pumice and is overlain by pumice fall. In both channels, pumice-bearing alluvium was deposited after the pumice fall. Below the confluence (location A2, fig. 1) channel- and peak-flow-facies deposits reflect variable character and timing of flows (fig. 6). Channel-facies deposits consist primarily of laminated to crossbedded, lithic, sandy- to sandy-gravel alluvium that contains frothy pumice fragments (fig. 6; table 3). Gray silt mantles some of these channel deposits that record dominantly fluvial deposition during a period of sustained flow that followed the pumice fall. Pumice-rich alluvium mantled only the valley floor. The peak-flow deposit, on the other hand, was preserved on the valley side more than 6 m above the channel bed (figs. 6, 8). That deposit consisted of massive lithic sand underneath pumice fall and waning-phase silty fall. Near the channel bed, however, a lithic debris-flow deposit that contains minor angular fragments of pumice overlies preeruption deposits. This deposit does not overlie pumice fall nor was pumice fall found on its surface. There was no obvious contact in the intervening zone between the deposit at the limit of peak flow and the deposit near the channel bed. However, the valley margin had been disturbed by later flow. Sustained recessional flow possibly removed whatever pumice fall may have mantled the lower deposit.

Figure 8. Stratigraphic sequence of deposits exposed in Abacan River valley at Sapangbato (location A2), Pampanga Province. View is looking downflow and represents a half cross section of the valley; deposits on the right of the figure are situated in the middle of the valley. See figure 6 and text for descriptive details. Horizontal dimensions not to scale.

Although houses upstream were inundated locally by lahars, severe channel erosion occurred along the Abacan River in Sapangbato (fig. 9). Sections of the channel were scoured to bedrock, suggesting that at least a meter or more of alluvium had been removed, and several buildings were undercut, indicative of several meters of lateral bank erosion (fig. 10). Neither exhumed bedrock nor freshly exposed bank faces in Sapangbato were covered by pumice fall. These observations, combined with the absence of pumice fall on channel-facies deposits, suggest that at least some of the erosion occurred during the period of sustained recessional flow that followed the peak-flow lahar. Above Sapangbato, particularly in the northern tributary of the Abacan River, pumice fall lies directly on some bedrock from which alluvium has been stripped. That stratigraphic relationship and the absence of prepumice-fall deposits in that channel suggest that water passed along the northern tributary before deposition of the pumice fall and scavenged sediment from the channel. This flow may have combined with flow coming down the southern tributary and formed the peak-flow lahar whose deposits were preserved at the distal end of Sapangbato. In Sapangbato, thick channel fill and an absence of pumice fall mantling eroded channel banks suggest that channel aggradation during recessional flow contributed to extensive lateral bank erosion. Increased sediment supply and rapid aggradation lead to smoothing of the channel bed (Dietrich and others, 1989), which in turn alters bedforms that can promote flow diversion and lead to bank erosion. From the observed stratigraphic relations and from general hydraulic responses of bedload-dominated rivers to increased sediment supply (for example, Richards, 1982), we speculate that bed sediment was scoured and mobilized mainly by the peak flow and that lateral bank erosion was accomplished primarily by the recessional flow.

Figure 9. Homes along the northern tributary of the Abacan River above Sapangbato inundated by lahars after deposition of the plinian pumice fall. View is downstream. Photograph taken July 8, 1991.

Figure 10. Lateral bank erosion caused by recessional phase of the June 15 lahar on Abacan River in Sapangbato, Pampanga Province. Photograph taken July 6, 1991.

Flood-plain sediments along the Abacan River 3 km below Sapangbato (location A3, fig. 1) record peak-flow deposition by a hyperconcentrated flow (figs. 6, 11; table 3). A massive to mildly laminated, slightly pebbly sand that contains lithic and pumice pebbles, deposited approximately 4 m above the channel bed, lies between gray silt deposits and is overlain by pumice fall. These stratigraphic relations show that peak flow passed this reach between the evening of June 14 and early afternoon on June 15. Personnel at Clark report hearing thunderous noises along the Abacan valley around 1400 on June 15 (R.P. Hoblitt, U.S. Geological Survey, oral commun., 1993). Peak-flow depth along this reach was about 2 m above the channel bed (at the time of investigation in late June 1991 the channel contained an estimated 3 m of fill), and the flood-plain deposit lies about 2 m above that limit. The position of the deposit, which lies at a sharp bend in the channel, suggests a minimum peak-flow velocity of about 6 m/s if the implied potential energy resulted from the complete conversion of the kinetic energy of the flow.

Figure 11. Stratigraphic sequence of deposits exposed in the Abacan River channel adjacent to Clark Air Base (location A3). The section view is oblique to the downflow direction. See figure 6 and text for descriptive details. Horizontal dimensions not to scale.

Channel deposits along this reach of the Abacan River consisted of primarily massive, pebbly to distinctly laminated, slightly pebbly lithic sands (fig. 6; table 3) that record hyperconcentrated flow and subsequent sediment-laden streamflow. Two hyperconcentrated-flow deposits underlie the pumice fall; streamflow deposits overlie the pumice fall.

Stratigraphic relations among deposits at the confluence of the Taug and Abacan Rivers, in Angeles City, record the passage of at least three waves of flow prior to deposition of the pumice fall (fig. 6). On the flood plain, poorly sorted, lithic sandy gravel overlies preeruption volcaniclastic deposits and is sharply overlain by massive sand and inversely graded pebbly sand. This sequence of deposits lies beneath the pumice fall. The basal unit is distinctly coarser than both the flood-plain and channel deposits observed a few kilometers upstream on the Abacan River. Those upstream deposits exhibit greater similarity to the massive and inversely graded sands at this section. The basal deposit here records a debris flow along the Taug River that preceded lahars on the Abacan River. Stratigraphic relations observed on Taug River 3 km upstream from the confluence (location A4, fig. 1) reveal a generally massive, poorly sorted lithic pebbly sand having matrix-supported clasts that lies between gray silt deposits and underlies pumice fall (fig. 6). This deposit nonconformably overlies the substrate (W.E. Scott, U.S. Geological Survey, written commun., 1992), indicating that the responsible flow was erosive along the channel bottom. Postpumice-fall flows on the Taug River generally were shallow and fluvial.

Several meters to tens of meters of lateral bank erosion occurred along the Abacan River in Angeles City (fig. 12). Buildings erected on terraces several meters above the channel collapsed after their foundations were undercut. Similar to the erosion documented in Sapangbato, the extensive lateral erosion in Angeles City occurred primarily during recessional flow after deposition of the pumice fall. Eroded banks and associated talus are not draped with pumice fall. Scour around piers and impacts from boulders and other debris caused structural failure of nearly all bridges across the river in Angeles City. The Northern Expressway bridge downstream from Angeles City escaped collapse; however, channel aggradation nearly buried that bridge.

Figure 12. View of Abacan River at Angeles City. Note erosion of left channel bank. Photograph taken June 25, 1991.

The most areally extensive deposition by lahars in the Abacan watershed occurred in distributary channels across the alluvial plain beyond Angeles City (fig. 13). By the time peak flow reached the bridge between Anao and Mexico (location A5, fig. 1), some 40 km river distance from the former summit of Mount Pinatubo, it had been significantly diluted by loss of sediment and it succeeded the pumice fall. There, pumice fall lies on flood-plain soil and beneath gray silt (fig. 6). Laminated sand overlies the silt (fig. 6; table 3). The top of that sand deposit is more than 2 m above the channel bed and about 1.2 m below the peak-flow limit preserved on fence posts.

Figure 13. Downstream aerial view of Abacan River below Angeles City. Beyond Angeles City, the lahar branched into many distributary channels, spilled overbank, and inundated parts of the alluvial plain, as seen on the right side of photograph. Photograph taken June 25, 1991.

In the village of Mexico, several kilometers farther downstream, peak flow was about 1 to 2 m deep according to eyewitnesses, and it deposited a laminated very fine sand and silt that overlies gray silt and pumice fall (location A6, fig. 6; table 3). This deposit, as well as that near Anao, shows clearly that along the alluvial plains the peak-discharge lahar associated with the June 15 eruption attenuated, slowed, and diluted.

PASIG-POTRERO RIVER SYSTEM

At the confluence of the primary tributaries that join to form the mainstem Pasig River (location PP1, fig. 1), a lahar deposit is interbedded with tephra fall and underlies thick pyroclastic fill (fig. 6; W.E. Scott, U.S. Geological Survey, written commun., 1992). At this location a muddy silt rests directly on an old road surface and on the former channel bed. In the former channel, this muddy silt is overlain by a generally thin, poorly sorted, lithic pebbly sand that lies directly below several centimeters of pumice fall. The pumice fall is overlain by several meters of pyroclastic debris as well as laharic debris from late-season monsoon-triggered flows. That the pebbly sand deposited below the pumiceous tephra, interpreted to be the deposit of a debris flow, is not found on terraces above the former channel indicates that the primary lahar was confined to the canyon at this location.

Deposits near Mancatian (location PP2, fig. 1) confirm that the primary lahar on the Pasig-Potrero River was a debris flow (fig. 6; table 3). Here, a massive, poorly sorted sandy pebble gravel was deposited on the flood plain more than 4 m above the channel bed. Pebbles are composed of both fresh and weathered lithics and of rare pumice. This deposit overlies silt that in turn rests on preeruption flood plain alluvium. The debris-flow deposit is overlain by pumice fall. These stratigraphic relations as well as an absence of silt on top of the debris-flow deposit suggest that peak discharge in the middle reach of the Pasig-Potrero River occurred in the early afternoon on June 15. Residents of Bacolor, about 13 km downstream, reported a sudden rise of stage at 1500. Within the main channel, and partly on the flood plain, the draping of preeruption volcaniclastic deposits and the pumice fall with a veneer of pumice sand, pebbles, and cobbles transported by subsequent water floods indicates that multiple waves of elevated discharge passed through the channel subsequent to the onset of the June 15 eruptions (fig. 14).

Figure 14. Stratigraphic sequence of deposits exposed on the right-bank flood plain of the Pasig-Potrero River at Mancatian (location PP2). See figure 6 and text for descriptive details. Horizontal dimensions not to scale.

The debris flow and early recessional-phase flow scoured nearly 1 m of alluvium from the channel and flood plain in the Mancatian reach, exhuming the underlying, relatively indurated, volcaniclastic deposits prior to deposition of the pumice fall (fig. 14). The minimum depth of the peak flow was about 4.5 m if we assume that the channel bed was not significantly lowered. The peak flow was sufficiently powerful to transport meter-sized boulders, to collapse the concrete bridge that crossed the river, and to transport slabs of that bridge, more than several meters long, nearly 2 km downvalley.

An aerial reconnaissance on July 3, 1991, revealed that the Pasig-Potrero channel was eroded along a several-kilometer reach. Several nickpoints were developed on the main stem of the river, and tributary valleys above Mancatian were freshly incised. The aerial investigation confirmed the passage of multiple flows. Consistent with the stratigraphic evidence, the peak discharge occurred before the pumice fall, and subsequent elevated discharges passed through the system during the late stages of, and after, pumice-fall deposition.

At the Santa Barbara bridge near Bacolor (location PP3, fig. 1), peak discharge is recorded by an ungraded, poorly sorted, lithic sandy pebble-gravel debris-flow deposit that lies directly on the formerly vegetated flood plain. This deposit, which is finer grained than that at Mancatian (table 3), is overlain by pumice fall. Pumice boulders are scattered on top of the fall deposit and locally on top of the debris-flow deposit where the fall has been eroded (fig. 6). This stratigraphic succession lies approximately 1.75 m above the former channel bed and records passage of at least two flows. At least one, and perhaps several, wave(s) of streamflow having a lower stage than the primary lahar reworked deposits within and near the channel and deposited the pumice clasts on top of the pumice fall.

Flow lines on the bridge piers attest to the differences in character and stage between the peak and recessional flows in this channel. An upper mudline reached nearly 3 m above the channel bed on the upstream face of the pier and about 2.4 m on the downstream face. A lower trimline, nearly 1.5 m above the channel bed on both the upstream and downstream faces of the pier, reflected the limit to which dilute flow(s) removed the muddy veneer. Differences in mudline elevation suggest that the instantaneous velocity of the peak discharge was about 3 to 3.5 m/s.

The peak flow along the Pasig-Potrero River spilled over a containment dike on the northeast side of the channel inundating arable land about 0.5 km upstream from Bacolor. Below Bacolor, aggradation of the mainstem channel and flow upstream along low-gradient distributary channels blocked drainage canals leading from rice fields to the river. As a result, parts of Bacolor and the surrounding area were inundated by backflooding of tributary canals. Although this lahar did not directly inundate large sections of inhabited or arable lands, it wreaked considerable havoc indirectly by inducing backflooding of channels along the alluvial plains.

PORAC RIVER SYSTEM

Residents of Porac report that silty tephra fall on June 15 was followed by a lahar at 1330, which was followed shortly thereafter by pumice fall (W.E. Scott, U.S. Geological Survey, written commun., 1992). Deposits in Porac and those a few kilometers upstream near Dolores (Porac Municipality) are consistent with those reports and further show that both debris flow and hyperconcentrated flow passed along the Porac River. Near Dolores (location P2, fig. 1), older flood-plain deposits are mantled by a few centimeters of silty tephra fall, which is overlain by a thin, massive to slightly inversely graded sand and a pebbly muddy sand (W.E. Scott, U.S. Geological Survey, written commun., 1992). These deposits are covered by pumice fall. The pumice fall is overlain by poorly sorted coarse sand containing conspicuous lithic pebbles and massive to laminated sand that contains scattered lithic pebbles. The postpumice-fall deposits represent a debris flow followed by more dilute recessional flow that passed through the channel on June 30 (W.E. Scott, U.S. Geological Survey, written commun., 1992).

Complex stratigraphy in the upper Porac River basin (location P1, fig. 1) indicates contemporaneous emplacement of pyroclastic flows and at least one lahar and also shows that some lahars were generated well after tephra-mantled hillslopes were incised by overland flow. Deposits related to the peak-discharge lahar, which occurred prior to the deposition of the pumice fall, were not observed, however. Several meters of multiple postpumice-fall pyroclastic-flow deposits fill the upper valley. At least one debris-flow deposit is interbedded with this fill, another mantles the fill, and channel fill inset against the pyroclastic deposits has the characteristics of a debris-flow deposit. An extensive network of rills eroded into the tephra mantle scores the surrounding hillsides (fig. 15). These rills, some as deep as a few tens of centimeters, extend downslope to the valley fill where they are partly filled by debris-flow deposits. Only a few major rills cut across the surface of the valley fill; minor rills are truncated. These relations show that debris was mobilized along the channel after the formation of the rill network on the hillslopes and that overland flow had diminished before deposition of the valley fill was complete.

Figure 15. Network of rills established on tephra-mantled hillslopes by rain that fell during the June 15 eruption. Photograph taken June 25, 1991.

Aerial reconnaissance on July 3 revealed that most of the low-order channels in the Porac watershed were scoured, reminiscent of erosion observed along lahar-impacted steepland channels following the 1985 eruption of Nevado del Ruiz Volcano, Colombia (Pierson and others, 1990). Although the sediment mobilized by the rilling of the hillslope tephra mantle contributed to downstream channel aggradation, it was not the source of sediment for the peak-discharge lahar. We suspect that the lithic-rich peak flow, which preceded deposition of the pumice fall, resulted from erosion and mobilization of channel sediments along the low-order channels.

DISCUSSION

The geomorphological responses of watersheds on the east side of Mount Pinatubo and the resulting timing and magnitudes of lahars reflect volcanic as well as climatic effects. Widespread lahars and floods of a magnitude and scale similar to, or slightly larger than, those of June 1991 last occurred during an eruptive period about 500 years ago (Newhall and others, this volume). Although isolated large, sediment-routing floods and debris flows undoubtedly occurred during the five-century repose, the watersheds generally have routed precipitation from hillslopes to channels and out onto alluvial fans without generating extensive laharic flows. Precipitation in June 1991 was apparently not uncommon, yet widespread debris flows and floods were generated, many prior to the plinian pumice fall. Clearly there is a link with renewed eruptive activity. Unlike lahars generated during eruptions of snow-clad volcanoes, where interaction among volcanic products and snow and ice can greatly increase the amount of water typically delivered to watershed channels, at Mount Pinatubo we cannot invoke an unusual increase in the volume of water delivered to the watersheds as the principle basis for generating these lahars. The lahars were triggered not solely by heavy precipitation but by substantive changes in watershed hydrology, which altered the manner and the rate that water was delivered from the hillslopes to the channels, in concert with coincidental heavy rains. Both conditions were necessary; neither was sufficient.

Volcanic events can profoundly alter the hydrologic regime of watersheds in two primary ways: (1) they can damage, strip, or remove vegetation and (2) they can deposit fine-grained, low-permeability sediment that commonly mantles the topography (for example, Segerstrom, 1950; Waldron, 1967; Kadomura and others, 1983; Collins and Dunne, 1986). Damage and removal of vegetation reduces interception of precipitation by the canopy, so more water falls unimpeded onto the landscape. Deposition of low-permeability sediment buries ground litter, which reduces infiltration capacity of the substrate and also reduces the scale of hillslope roughness by smoothing the landscape. Infiltration rates can be reduced by more than an order of magnitude following tephra deposition (Leavesley and others, 1989). Each of these types of disturbance promotes increased overland flow, which delivers greater amounts of water from hillslopes to channels at a rate faster than would normally occur under preeruption conditions, and, through commonly associated rilling (for example, Collins and Dunne, 1986) and related shallow landsliding, enhances sediment delivery. Volcanically disturbed watersheds are therefore subject to larger peak flows and to greater volumes of runoff than would occur under preeruption conditions, and opportunities for sediment erosion and transport are substantially increased (Leavesley and others, 1989). Enhanced overland flow resulting from accumulated tephra fall in watersheds around Parícutin volcano, Mexico (Segerstrom, 1950), Irazú volcano, Costa Rica (Ulate and Corrales, 1966; Waldron, 1967), and Usu volcano, Japan (Kadomura and others, 1983), triggered several debris flows, even during relatively light rainfall.

Although the climactic-phase activity produced the most dramatic landscape disturbance, the hydrology of watersheds at Mount Pinatubo was severely altered, and substantial overland flow was generated, before the onset of the climactic eruption. Stratigraphic relations show clearly that many of the peak-discharge lahars were generated prior to deposition of the widely distributed pumice fall related to the climactic eruption. Catchment headwaters were impacted by several pyroclastic surges, associated tephra fall, and minor pyroclastic flows in the days, and particularly in the 24 h, preceding the climactic eruption. These volcanic events damaged, removed, or buried vegetation and deposited generally thin, but variable, layers of chiefly fine-grained sediments, dominated by fine sand and silt, across the landscape. The occurrence of widespread, preclimactic lahars illustrates the delicate sensitivity of watershed hydrology to disturbance by even relatively thin, fine-grained volcanic deposits.

Significant rainfall runoff persisted into the phase of climactic eruptive activity, then rapidly diminished. An extensive rill network and numerous shallow landslides were observed in the mantle of pumice tephra on the hillslopes (figs. 15, 16), yet valley fill truncates many rills, and by late July 1991, the thick pyroclastic debris filling several valleys still lacked an integrated drainage system (K.M. Scott and others, this volume). The fact that only major rills cut across valley-bottom deposits indicates that most of the observed rill network was developed in the short time between deposition of the pumice tephra during the climactic eruption and deposition of the thick valley fill. Despite the fact that deposits from the climactic eruption completely destroyed vegetation in catchment headwaters and severely disrupted watershed hydrology, the magnitude of overland flow had diminished before final deposition of the valley fill; otherwise, a more extensive drainage network would have developed on that fill. An extensive, integrated drainage network rapidly developed on the thick valley fill once the seasonal rains began in earnest (K.M. Scott and others, this volume). The reduction of overland flow likely reflects migration of Typhoon Yunya and a decrease in precipitation intensity. The sediment mobilized from these rills, and from the shallow slides, accounts for the numerous debris-flow deposits observed interbedded with the pyroclastic valley fill, but these were not the sources of sediment for the peak-discharge lahars.

Figure 16. Shallow landslides in pumice-tephra mantle. Photograph taken June 25, 1991.

Along the mountain front from the Porac River to the Sacobia River, peak discharge preceded deposition of the broadly distributed pumice fall (table 4). Between the Sacobia River and the O'Donnell River, peak discharge largely followed pumice fall; however, we were unable to examine proximal areas. The timing of peak discharge in each watershed probably reflects localization of variable-intensity eruption-induced rainfall, northward tracking of Typhoon Yunya, and perhaps variations in the intensity to which catchment headwaters were disturbed prior to the climactic activity. The characteristic response time of each watershed, the time required for the whole watershed to contribute to steady-state discharge at some outlet point, is probably a second-order effect influencing the variation in the timing of peak discharge. Transit times for overland flow off hillslopes and for channel flow to the chosen outlet point (see Dunne and Leopold, 1978; Dunne and Dietrich, 1980) reveal gross similarity of estimated response times (table 1). Thus, under equivalent conditions, a parcel of fluid from the farthest point in the basin should reach a comparable outlet point in a roughly equivalent amount of time for each watershed, except for the larger O'Donnell watershed, where the response time is slightly longer.

Table 4. Summary of characteristics and timing of peak-discharge lahars on the east side of Mount Pinatubo.

Peak flow in the Abacan River valley may have been affected by the emplacement of fill in the Sacobia River valley. We have established that peak discharge along the Abacan River and its tributaries occurred before deposition of the pumice fall and that channel erosion probably was associated with peak flow, at least in the upper basin. The position of peak-flow deposits along valley walls in Sapangbato (more than 6 m above the channel floor), the magnitude of channel erosion by peak flow, and the thickness of subsequent channel deposits attest to large, and sustained, discharges of water and sediment moving through this valley. Yet, the Abacan River does not head on the flanks of Mount Pinatubo, and its watershed above Sapangbato has an area of only about 14 km2. The magnitude and physical effects of the peak flow suggest a larger source of water than is probably deliverable by runoff within that part of the watershed. To estimate a possible peak discharge by rainfall runoff, we used a unit hydrograph developed for the upper Abacan watershed by the U.S. Army Corps of Engineers (S.L. Stockton, USCOE, written commun., 1993) and an assumed design storm of 150 mm of rainfall over 6 hours. In the calculation of runoff from that storm (Dunne and Leopold, 1978, p. 385-387) we allow a 1-h hydrograph lag but do not allow runoff removal by infiltration. We estimate that the watershed above Sapangbato was capable of delivering a peak water discharge of about 20 m3/s from such a storm. Even if we allow peak flow to have a sediment concentration of about 66 volume percent, the peak discharge deliverable from such a storm in the upper Abacan watershed would be only about 60 m3/s, much less than the peak discharge that is suggested by channel conditions. Thus, the upper Abacan watershed was either subject to a more intense storm or there was contribution of water from another source. By late July 1991, flow from the Sacobia River valley, whose watershed encompasses some 30 km2 above Clark Air Base, was diverted completely into the Abacan watershed (K.M. Scott and others, this volume). We hypothesize that the peak discharge on the upper Abacan River, which preceded the pumice fall, may have resulted from temporary diversion of flow from the Sacobia River valley to the Abacan River valley. If so, temporary diversion of flow must have occurred before, or during, the earliest phases of the onset of climactic eruptive activity. W.E. Scott and others (this volume) have shown that within 5 km of the volcano, deposition of the thick pyroclastic valley fill and the pumice fall were synchronous; beyond about 5 km, pumice fall preceded emplacement of the pyroclastic fill. If our hypothesis regarding the origin of peak discharge on the upper Abacan River is correct, it indicates that some 40 to 60 m of fill must have rapidly choked the Sacobia River valley and allowed water to spill across the drainage divide before pumice fall began along the mountain front. Thick fluvial deposits (more than 2 m) in the Abacan River channel that do not underlie pumice fall indicate sustained elevated discharge, which suggests that flow from the Sacobia watershed may have been actively diverted for several hours. Pyroclastic-flow deposits that cap the fill in the Sacobia River valley indicate that pyroclastic activity also was sustained for several hours and outlasted any diversion of water, because no well-defined channel was incised on that fill or through the watershed divide by July 8, 1991. The primary drawback to our hypothesis on the cause of peak flow in the Abacan River valley is that peak-flow deposits contain little pumice. It is difficult to reconcile substantial water sweeping across extensive pyroclastic valley fill without entraining much pumice, unless a part of that valley fill was not pumiceous. W.E. Scott and others (this volume) have noted zones of lithic debris in at least the upper sections of the valley

Despite their minuscule magnitude in comparison with subsequent lahars, the mid-June lahars seriously disturbed channels downstream. Overall, little direct damage resulted from lahars spilling overbank and flooding populous areas, although localized overbank flooding along fan channels inundated homes and on distal alluvial plains inundated arable land, particularly along the Bamban and Abacan Rivers. Instead, lahars generally aggraded channels, and this aggradation triggered lateral bank erosion. Along the alluvial plains, aggradation of mainstem channels led to backflooding of tributary channels.

Hydrograph peaks attenuated and slowed, and flows diluted as they traveled downvalley (table 4). The best example is illustrated by flow along the Abacan River. The hydrograph peak at Sapangbato was nearly 6 m above the present channel bed, and it passed before the pumice fall was completely deposited. Three kilometers downstream, peak-flow depth was no more than 2 m above the present channel bed and probably no more than 5 m above the original channel bed. By the time the peak traveled nearly 20 km farther downvalley, it had been significantly diluted by loss of sediment, its depth was little more than 3 m, and it succeeded the pumice fall. Along the distal plains, peak flow was less than 2 m deep and transported only very fine sand and silt. Above Angeles City, where peak flow preceded pumice fall, we have estimated peak-flow velocity to be about 6 m/s. Eyewitnesses report that peak flow reached the town of Mexico at approximately 1600 on June 15. If we assume that widespread pumice fall began in the Pampanga Province around 1400, the arrival time of peak flow in Mexico yields a mean velocity of less than 4 m/s below Angeles City. Shunting of flow into several distributary channels and dilution of flow through loss of sediment probably contributed to the attenuation and slowing of peak flow below Angeles City.

CONCLUSIONS

Lahars were generated in all major watersheds east of Mount Pinatubo as a result of the June 1991 eruptions. Although these lahars were triggered by heavy rainfall associated with Typhoon Yunya, and perhaps with localized convective storm cells, the rainfall that triggered these lahars was not uncommon for this tropical climate. Heavy rainfall alone was not responsible for generating the lahars. Another necessary condition for the formation of these lahars was alteration of watershed hydrology by volcanic deposits. Deposits associated with pyroclastic density currents and chiefly fine-grained tephra fall that disturbed catchment headwaters prior to the climactic phase of eruptive activity on June 15 played a key role in the formation of the peak-discharge lahars that impacted channels along the mountain front from the Porac River to the Sacobia River. Those volcanic deposits destroyed or damaged vegetation, reduced infiltration capacity of the substrate, and smoothed the natural-scale roughness of the hillslopes. As a result, overland flow increased, and a higher percentage of the rain that fell on the landscape was delivered from the hillslopes to the channels at a rate faster than usual. The rapid delivery of large volumes of water from hillslopes to channels triggered hillslope and bank erosion, channel scour, and possibly minor slope failures, which led to lithic-rich debris flows and more dilute hyperconcentrated flows. Unconsolidated channel sediment probably was the dominant source of sediment for these lahars because many channel banks were composed of pumiceous debris (Newhall and others, this volume). Pyroclastic flows, ash-cloud surges, and pumice fall deposited during the climactic eruptive activity completely destroyed vegetation and produced the most dramatic landscape disturbance in catchment headwaters. Sustained rainfall during this activity led to extensive, shallow slope failures and to extraordinary overland flow that developed an extensive rill network on the pumice-tephra mantle. Sediment mobilized from these rills and shallow landslides generated several pumice-rich lahars that are found interbedded with the pyroclastic valley fill and contributed to the postpumice-fall lahars on the O'Donnell River and to the pumice-bearing recessional-flow deposits observed in several channels. Truncation of rills by valley fill and the lack of an integrated drainage on that valley fill show that excessive overland flow had diminished before the caldera formed and pyroclastic flows ended. Although the climactic eruptive activity produced the most dramatic and widespread landscape disturbance and severely disrupted watershed hydrology, stratigraphic constraints on the timing of peak discharge in several eastern watersheds illustrate the delicate sensitivity of watershed hydrology to disturbance by even relatively thin, fine-grained volcanic deposits.

The timing of peak discharge with respect to the onset of the climactic phase of activity on June 15 varied in a northerly direction across watersheds and downstream within watersheds. From the Porac River to the Sacobia River, peak discharge along the mountain front occurred before the plinian pumice fall; between the Sacobia River and O'Donnell River, peak discharge largely followed the pumice fall. On the alluvial plains, peak flows generally followed the pumice fall. The timing of these lahars probably reflects localization of variable-intensity, eruption-induced rainfall, northward movement of Typhoon Yunya, and perhaps variations in the degree of disturbance of catchment headwaters by preclimactic eruptive activity. Characteristic response times of watersheds do not appear to differ significantly, and their influence on the variation in lahar timing is probably more subtle than is the influence of variations in rainfall. The interplay of these various factors affects the amount of water that moves from hillslopes to channels to fanheads, how rapidly it moves, and its ability to entrain and transport sediment.

Peak flows commonly attenuated, slowed, and diluted as they moved across the alluvial fans and onto the alluvial plains. Along the mountain front, peak flows were as deep as 6 m or more and had instantaneous velocities of about 6 m/s. In distal reaches, peak flows commonly were less than 1 to 3 m deep, and mean peak-flow velocities were less than 4 m/s. The sedimentology of downstream deposits suggests that many lahars were rapidly transformed to hyperconcentrated and normal streamflows beyond the mountain front; however, some of the lahars remained as debris flows across the alluvial fans.

The mid-June lahars damaged inhabited areas on the densely populated alluvial fans of Tarlac and Pampanga Provinces primarily by triggering lateral bank erosion that undermined buildings and by aggrading mainstem channels that led to backflooding of tributary channels. Although some areas were directly buried by lahars, burial was generally of secondary importance to damage triggered by bank erosion and by induced backflooding, except along distal alluvial plains where primarily agricultural land was inundated. As devastating as these lahars were, they had far less social and economic impact than subsequent lahars triggered by the seasonal monsoon rains.

ACKNOWLEDGMENTS

Reconnaissance fieldwork upon which this report is based was accomplished between late June and mid-July 1991, when J.J. Major and R.J. Janda visited the Philippines under the auspices of the USGS-USAID Volcano Disaster Assistance Program. The emphasis of the fieldwork was to establish the nature and magnitude of watershed disturbance and the character of downstream deposits in order to provide a first-approximation forecast of subsequent hydrologic hazards. Under very trying circumstances, Dick Janda was a master at honing in upon, and elucidating, the critical essence of outcrop and landscape relations regardless of stratigraphic and morphologic complexity. His enthusiasm, critical eye, and firm belief that colleagues should have the opportunity to experience such trying situations led to the development of this paper. He will be sorely missed. Logistical support for the study was provided by the U.S. Air Force, U.S. Navy, and the Pinatubo Lahar Task Force. We thank W.E. Scott, K.M. Scott, and R.P. Hoblitt for several enlightening discussions of volcanic and hydrologic events at Mount Pinatubo, and Tom Dunne for discussions concerning watershed hydrology. We thank K.M. Scott, C.G. Newhall, G.A. Smith, Kelin Whipple, and K.S. Rodolfo for critical reviews of an early draft.

REFERENCES CITED

Alaska Volcano Observatory, 1990, The 1989-1990 eruption of Redoubt Volcano: Eos, Transactions, American Geophysical Union, v. 71, no. 7, p. 265-275.

Best, J.L., 1992, Sedimentology and event timing of a catastrophic volcaniclastic mass flow, Volcan Hudson, Southern Chile: Bulletin of Volcanology, v. 54, p. 299-318.

Beverage, J.P., and Culberston, J.K., 1964, Hyperconcentrations of suspended sediment: American Society of Civil Engineers, Journal of Hydraulics Division, v. 90, HY6, p. 117-128.

Brantley, S.R., ed., 1990, The eruption of Redoubt Volcano, Alaska, December 14, 1989-August 31, 1990: U.S. Geological Survey Circular 1061, 33 p.

Collins, B.D., and Dunne, Thomas, 1986, Erosion of tephra from the 1980 eruption of Mount St. Helens: Geological Society of America Bulletin, v. 97, p. 896-905.

Dietrich, W.E., Kirchner, J.W., Ikeda, Hiroshi, and Iseya, Fujiko, 1989, Sediment supply and the development of the coarse surface layer in gravel-bedded rivers: Nature, v. 340, p. 215-217.

Dolan, M.T., 1993, Buried in mud: A look at the sedimentological and morphological characteristics of lahars on Pinatubo Volcano's eastern slopes: Eos, Transactions, American Geophysical Union, v. 74, no. 43, p. 671.

Dorava, J.M., and Meyer, D.F., 1994, Hydrologic hazards in the lower Drift River basin associated with the 1989-90 eruptions of Redoubt Volcano, Alaska: Journal of Volcanology and Geothermal Research, v. 62, p. 387-407.

Dunne, Thomas, and Dietrich, W.E., 1980, Experimental investigation of Horton overland flow on tropical hillslopes. 2. Hydraulic characteristics and hillslope hydrographs: Zeitschrift für Geomorphologie, Supplement-Band 35, p. 60-80.

Dunne, Thomas, and Leopold, L.B., 1978, Water in environmental planning: W.H. Freeman, 818 p.

Fisher, R.V., and Schminke, H.U., 1984, Pyroclastic rocks: Springer-Verlag, 472 p.

Folk, R.L., 1974, Petrology of sedimentary rocks: Austin, Tex., Hemphill Publishing Co., 182 p.

Giambelluca, T.W., Lau, L.S., Fok, Y.S., and Schroeder, T.A., 1984, Rainfall frequency study for Oahu: U.S. Army Corps of Engineers, Honolulu district, report R-73, 34 p.

Hoblitt, R.P., Wolfe, E.W., Lockhart, A.B., Ewert, J.E., Murray, T.L., Harlow, D.H., Mori, J., Daag, A.S., and Tubianosa, B.S., 1991, 1991 eruptive behavior of Mount Pinatubo: Eos, Transactions, American Geophysical Union, v. 72, no. 4, p. 61.

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

Janda, R.J., Scott, K.M., Nolan, K.M., and Martinson, H.A., 1981, Lahar movement, effects, and deposits, in Lipman, P.W., and Mullineaux, eds., The 1980 eruptions of Mount St. Helens, Washington: U.S. Geological Survey Professional Paper 1250, p. 461-478.

Jones, J.W., and Newhall C.G., this volume, Preeruption and posteruption digital terrain models of Mount Pinatudo.

Kadomura, Hiroshi, Imagawa, Tashiaki, and Yamamoto, Hiroshi, 1983, Eruption-induced rapid erosion and mass movements on Usu volcano, Hokkaido: Zeitschrift für Geomorphologie, Supplement Band 46, p. 123-142.

Leavesley, G.H., Lusby, G.C., and Lichty, R.W., 1989, Infiltration and erosion characteristics of selected tephra deposits from the 1980 eruption of Mount St. Helens, Washington, USA: Hydrological Sciences Journal, v. 34, p. 339-353.

MacArthur, R.C., Hamilton, D.L., Harvey, M.D., and Kekaula, H.W., 1992, Analyses of special hazards and flooding problems in tropical island environments: Proceedings of American Society of Civil Engineers Hydraulic Engineering Water Forum '92, Baltimore, Maryland, August 2-6, p. 1061-1066.

Major, J.J., and Iverson, R.M., 1993, Is the dynamic behavior of a debris flow recorded by its deposit?: Eos, Transactions, American Geophysical Union, v. 74, no. 43, p. 315.

Major, J.J., and Newhall, C.G., 1989, Snow and ice perturbation during historical volcanic eruptions and the formation of lahars and floods--a global review: Bulletin of Volcanology, v. 52, p. 1-27.

Major, J.J., and Voight, Barry, 1986, Sedimentology and clast orientations of the May 18, 1980 southwest flank lahars, Mount St. Helens, Washington: Journal of Sedimentary Petrology, v. 56, p. 691-705.

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

Oswalt, J.S., O'Hara, J.F., and Nichols, William, 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., and Delos Reyes, P.J., this volume, Tephra falls of the 1991 eruptions of Mount Pinatubo.

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 post-eruption hot lahars on the east side of Mount Pinatubo, July-October, 1991.

Pierson, T.C., Janda, R.J., Thouret, J.C., and Borrero, C.A., 1990, Perturbation and melting of snow and ice by the 13 November 1985 eruption of Nevado del Ruiz, Colombia, and consequent mobilization, flow and deposition of lahars: Journal of Volcanology and Geothermal Research, v. 41, p. 17-66.

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 Investigation Report 92-4039, 35 p.

Pierson, T.C., and Scott, K.M., 1985, Downstream dilution of a lahar: Transition from debris flow to hyperconcentrated streamflow: Water Resources Research, v. 21, no. 10, p. 1511-1524.

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

Richards, Keith, 1982, Rivers: Form and process in alluvial channels: Methuen, 361 p.

Rodolfo, K.S., 1991, Climatic, volcaniclastic, and geomorphic controls on the differential timing of lahars on the east and west sides of Mount Pinatubo during and after its June 1991 eruption: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 62.

Rodolfo, K.S., and Arguden, A.T., 1991, Rain-lahar generation and sediment-delivery systems at Mayon Volcano, Philippines, in Fisher, R.V., and Smith, G.A., eds., Sedimentation in volcanic settings: SEPM Special Publication No. 45, p. 71-87.

Rodolfo, K.S., Umbal, J.V., Alonso, R.A., Remotigue, M.C., Paladio, M.L., Salvador, J.H.G., Evangelista, D., and Miller, Y., this volume, Two years of lahars on the western flank of Mount Pinatubo: Initiation, flow processes, deposits, and attendant geomorphic and hydraulic changes.

Scott, K.M., 1988, Origins, behavior, and sedimentology of lahars and lahar-runout flows in the Toutle-Cowlitz River system: U.S. Geological Survey Professional Paper 1447-A, 74 p.

Scott, K.M., Janda, R.J., de la Cruz, E., Gabinete, E., Eto, I., Isada, M., Sexon, M., and Hadley, K.H., this volume, Channel and sedimentation responses to large volumes of 1991 volcanic deposits on the east flank of Mount Pinatubo.

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: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 61-62.

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

Segerstrom, Kenneth, 1950, Erosion studies at Parícutin, State of Michoacan, Mexico: U.S. Geological Survey Bulletin 965-A, 164 p.

Smith, G.A. and Fritz, W.J., 1989, Volcanic influences on terrestrial sedimentation: Geology, v. 17, p. 375-376.

Smith, G.A., and Lowe, D.R., 1991, Lahars: Volcano-hydrologic events and deposition in the debris flow-hyperconcentrated flow continuum, in Fisher, R.V., and Smith, G.A., eds., Sedimentation in volcanic settings: SEPM Special Publication No. 45, p. 59-70.

Trabant, D.C., Waitt, R.B., and Major, J.J., 1994, Disruption of Drift glacier and origin of floods during the 1989-90 eruptions of Redoubt Volcano, Alaska: Journal of Volcanology and Geothermal Research, v. 62, p. 369-385.

Ulate, C.A., and Corrales, M.F., 1966, Mud floods related to the Irazú volcano eruptions: American Society of Civil Engineers Journal of Hydraulics Division, v. 92, no. HY6, p. 117-129.

Umbal, J.V., and Rodolfo, K.S., this volume, The 1991 lahars of southwestern Mount Pinatubo and evolution of the lahar-dammed Mapanuepe Lake.

Waitt, R.B., Pierson, T.C., MacLeod, N.S., Janda, R.J., Voight, B., and Holcomb, R.T., 1983, Eruption-triggered avalanche, flood, and lahar at Mount St. Helens--effects of winter snowpack: Science, v. 221, p. 1394-1396.

Waldron, H.H., 1967, Debris flow and erosion control problems caused by the ash eruptions of Irazú volcano, Costa Rica: U.S. Geological Survey Bulletin 1241-I, 37 p.

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 <https://pubs.usgs.gov/pinatubo/major/>
Contact: Chris Newhall
Last updated 06.11.99