FIRE and MUD Contents

Channel and Sedimentation Responses to Large Volumes of 1991 Volcanic Deposits on the East Flank of Mount Pinatubo

By Kevin M. Scott,1 Richard J. Janda,1 Edwin G. de la Cruz,2 Elmer Gabinete,2 Ismael Eto,2 Manuel Isada,2 Manuel Sexon,2 and Kevin C. Hadley1

1 U.S. Geological Survey.

2 Philippine Institute of Volcanology and Seismology.


ABSTRACT

With the beginning of the 1991 rainy season in late July, devastatingly high sediment yields began from the east flank of Mount Pinatubo. Sediment was derived from the deposits of pyroclastic flows, which filled valleys locally to depths of over 100 m during the mid-June eruption, and from airfall deposits that mantled upland slopes. This paper focuses on early sedimentation responses and on the characteristics of the flows and deposits that are important for immediate control measures. Long-term responses were forecast for individual drainages on the basis of the type and thickness of eruptive products and on bedrock characteristics.

A critical factor in the immediate planning of mitigation measures is a trend toward posteruption fill in fanhead areas, indicated by measured channel cross sections that record high rates of upstream deposition. With the consequent potential for channel avulsions near the fanhead, the entire arc of active fans is at risk from future flows. This trend is validated by the pattern of sedimentation after previous eruptions.

At Mount Pinatubo, the flows causing widespread damage on populated fan surfaces are noncohesive (low in clay-size sediment). The flows are dominantly in the range of hyperconcentrated flow (containing 20 to 60 percent sediment by volume), and sand-size sediment is the dominant constituent. Some large flows leave the mountain front as debris flows that, because of their noncohesive character, are rapidly diluted to erosive hyperconcentrated flow. Consequently, mitigation plans must assess both the loss of channel capacity and a high potential for levee erosion and breaching.

A related consideration is the texture of fan deposits available for levee construction. Noncohesive deposits of identical hyperconcentrated flows associated with previous eruptions dominate the sediment composing the fans; they are the most accessible levee-construction materials but are, however, extremely erodible. Coarser, less-erodible materials for levee facing or wing dikes are less readily available.

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

Between 5 and 6 km3 of eruptive products were emplaced in watersheds surrounding Mount Pinatubo during the eruption climaxing June 15, 1991 (see W.E. Scott and others, this volume). Most of this volume was emplaced as thick valley fills of pyroclastic-flow deposits. Part of a smaller volume of airfall deposits on Luzon, estimated as approximately 1 km3 (Paladio-Melosantos and others, this volume), mantled upland areas. The potential for destructive redistribution of this material by sediment-laden, water-mobilized flows into habitated areas was immediately clear; how, where, and over what timeframe it would occur was not.

The greatest concern was for the downstream, highly populated areas on the east side of the volcano (fig. 1). The risks, however, would probably vary with factors that would become clear only after onset of the rainy season. Widely disparate thicknesses of pyroclastic-flow and airfall deposits occurred on watersheds that would differ greatly in their erosional and rainfall-runoff responses.

Figure 1. Drainage system of the eastern side of Mount Pinatubo. Cross-section sites are as follows: A, Bangat River at Santa Lucia (Bangat Bridge); Bangat River is a tributary of O'Donnell-Tarlac River. B, O'Donnell-Tarlac River upstream of Tarlac. C, O'Donnell-Tarlac River at Tarlac (Agana Bridge). D, O'Donnell-Tarlac River at Tarlac (downstream of site C). E, Marimla River above confluence with Sacobia-Bamban River. F, Sacobia River above Clark Air Base. G, Sacobia River near Clark Air Base. H, Sacobia-Bamban River at Bamban. I, Sacobia-Bamban River at San Francisco Bridge. J, Pasig-Potrero River at Mancatian. K, Pasig-Potrero River at Santa Barbara Bridge.Villages in parentheses were destroyed by the eruption.

Various types of sediment-laden flow transport much of the sediment from volcaniclastic terranes and have done so from Mount Pinatubo. In debris flows, sediment moves the interstitial water, rather than the reverse, as in normal streamflow, and the flows have the appearance and impact forces of flowing concrete. At Mount Pinatubo, debris flows evolve downstream into more dilute types of flow, especially hyperconcentrated flow, as described in the section on flow processes. Both debris flows and hyperconcentrated flows are commonly known by the Indonesian term "lahar." This general term is used here only when there is no evidence for a specific flow type such as debris flow or hyperconcentrated flow. All types of sediment-laden flow can produce harmful effects, and these include, in addition to loss of human life and habitation, the loss of all types of infrastructure, the loss of channel and irrigation-canal capacity, burial of agricultural lands, and siltation of aquaculture ponds.

Scour and fill are short-term erosion and deposition, respectively, commonly in channels. Degradation and aggradation are long-term erosion and deposition, respectively, in channels or on any part of the landscape. Much of the fill discussed here will ultimately prove to be aggradation.

The purpose of this contribution is to summarize the important effects of sediment transport and deposition, as they concern mitigation measures, and as they both appeared to us initially and again with the subsequent perspective of the 1991 and 1992 rainy seasons. The object is to provide planners with information critical to the engineering mitigation now being planned; that is, to define the type, behavior, and the depositional pattern of the sediment-conveying flows. Other long-term evaluations of the situation at Pinatubo have illustrated potential inundation areas in the absence of any engineering mitigation (Punongbayan and others, 1991; Pierson and others, l992). These studies define probable long-term trends, both areally and over time, whereas this paper summarizes other aspects of the sediment-distribution system that are important in planning present mitigation measures. Rodolfo (1991) has contrasted early responses on the east and west sides of the volcano.

DRAINAGE SYSTEM ON THE EAST SIDE OF MOUNT PINATUBO

The major drainages are shown on figure l. From north to south, the systems of greatest interest are the O'Donnell-Tarlac River system (downstream portion known as the Tarlac River), Sacobia-Bamban River system (downstream portion known as the Bamban River), the Abacan River system, and the Pasig-Potrero River system (downstream portion most commonly known as the Pasig-Potrero River). The compound systems may be referred to in part or entirety by their compound names, but upstream sites may be referenced by the names of the upstream drainages: the O'Donnell, Sacobia, or Pasig River, for example. The two major drainages farther south are the Porac and Gumain Rivers, which are now joined in a leveed drain leading to the axial, north-south drainage of central Luzon.

POSTERUPTION CHARACTERISTICS OF SLOPES AND CHANNELS

The rainfall-runoff characteristics of the landscape on the east side of the volcano were radically changed by the slope-mantling airfall deposits. Rilling in the deposits began almost immediately, and rills and small gulleys rapidly cut through the airfall and into the preexisting ground surface. Shallow, lahar-conveying channels gradually migrated upstream on the valley-fill deposits of the pyroclastic flows.

Primary pyroclastic flow deposits filled east-flank valleys to depths of as much as 140 m (R.S. Punongbayan, oral commun., 1992). Before rainy-season onset, little drainage integration had occurred on the fill surfaces, and many channels flowed into phreatic explosion pits. Explosion pits were first caused by infiltrating surface drainage, but they subsequently guided drainage development over the short term. A form of beaded drainage (fig. 2) was observed in the first overflights and aerial photographs. This drainage was largely independent of the preexisting pattern, other than in general direction. Some drainage followed primary flow lineations preserved on deposit surfaces of the pyroclastic flows.

Figure 2. Drainage pattern on surface of deposits of pyroclastic flows. Note beaded drainage pattern connecting phreatic explosion pits. The latter are here aligned along primary flow lines on the surface of the deposits. Drainage is from upper left to lower right. Because cloud cover obscured similar features on the east flank of Mount Pinatubo before they were destroyed by erosion, this photograph is from the Marella River drainage, southwest of Mount Pinatubo. Photograph is approximately 0.5 km in width and was taken July 5, 1991, by the U.S. Air Force.

Significant flooding and some lahars had extended to eastern populated areas in response to the storm shortly before the eruption, coincident with it, and with later storms that predated the rainy season. The timing of these events relative to the tephra fall and pyroclastic flows was complex (see Janda and others, 1991; Major and others, this volume). These events, however, were not the precursors of major trends.

The most obvious early channel response at bridge crossings nearest the mountain front was scour. Scour was observed as a response to the syneruptive lahars throughout the upper lowland course of the Pasig-Potrero River (J.J. Major, written commun., 1992). Scour also affected the fanhead area of the Sacobia-Bamban River. Scour of 2 to 5 m occurred between July 15 and 22 opposite the site of Clark Air Base, before most of the headwater drainage was recaptured on July 25 (see section on channel diversions). Another episode of 6 to 7 m of scour occurred there in late August and early September (Pierson and others, 1992). Bridge-pier footings were visibly exposed to a depth of about 2.5 m during mid-July at the site on the Sacobia-Bamban River labeled I on figure 1.

Scour proved to be a misleading response in light of the devastating volumes of fill that began in upland and many lowland areas with the onset of the 1991 rainy season in mid-July. Some remaining bridge piers were buttressed against erosion but would be mainly subject to unremitting burial. The longer term trends, which appeared as the rainy season progressed, are discussed below.

GEOMORPHIC SETTING

Although the vegetative cover is normally dense, the eastern side of Mount Pinatubo (preeruption altitude of 1,745 m) is in many ways similar to the dissected uplands and alluvial fans of semiarid regions (see Ely and Baker, 1990). The principles of flood hydrology on semiarid alluvial fans are closely applicable. The factor in common to both settings is rare periods of intense sediment yield. Terrace levels and alluvial fills record previous eruptive periods at Mount Pinatubo, just as the same features record the alluviation by rare, catastrophic flows in semiarid areas. Incised channels on alluvial fans east of Pinatubo reveal sequences of flow deposits from previous eruptions. This is a critical observation in assessing the distribution of future flows and possible structural measures for their control.

Upland terrane, whether consisting of bedrock or unconsolidated deposits, will be referred to here as "upland," and areas below the mountain front (bedrock-alluvium contact) will be distinguished as "lowland." The lowland is a piedmont formed of coalesced alluvial fans. Well-defined fanhead areas separate the upland portions of drainage systems from their lowland extensions (fig. 1). Implications of the alluvial fan environment include the fanwide distribution of flood risk, beginning at the apex of the arc in the fanhead area. Fanhead areas are characterized by incised channels, known as fanhead trenches, that record low-flow channel degradation between major flows. Rapid fill at the fanhead can direct large flows to any part of an active fan. During a sedimentation episode as profound as that beginning at Mount Pinatubo, repeated cycles of trench filling, avulsion, degradation, and repeated filling will expectedly distribute sediment across much of a fan surface.

CLIMATE

The tropical climate is dominated by several air masses during the year, the most important of which at Mount Pinatubo is the southwest monsoon that typically begins in late May or early June and lasts until September or October (see Umbal and Rodolfo, this volume). The 1991 rainy season began in July, but the 1992 rainy season arrived in late June. Mean annual precipitation at Clark Air Base, altitude 146 m, is 1,950 mm. On the basis of long-term trends, 57 percent of the precipitation occurs in July, August, and September. Precipitation is substantially greater at higher altitudes. The Clark site represents a typical middle altitude location, between the Sacobia and Abacan Rivers, near the mountain front on the east side of the volcano.

Typhoons are rare early in the year, but their frequency increases gradually through April and May. They are frequent from June through December and are an especially characteristic threat from September to November, a period that is late in the rainy season and during the "transition season" following the main rainy season. Detailed long-term weather records are available from the location of Clark Air Base (U.S. Air Force, 1991).

FLOW TYPE AND DEPOSIT SEDIMENTOLOGY

Granular debris flows are the initial process by which volcanic detritus is redistributed from uplands to lowlands, primarily during rainy seasons. The flows are dominated by sand-size sediment (0.0625 to 2.0 mm); they are uniformly low in clay-size sediment (finer than 0.004 mm) and commonly contain less than 1 or 2 percent of that size fraction. The flows are distinct in both origin and behavior from debris flows with more than 3 to 5 percent of clay, and the two types can be distinguished as noncohesive and cohesive, respectively (Scott and others, 1992).

Both cohesive and noncohesive flows commonly recur at many volcanoes, but noncohesive flows dominate at Pinatubo. A significant behavioral distinction is that noncohesive flows transform from debris flows to hyperconcentrated flow (20 to 60 percent sediment by volume) to normal streamflow (less than 20 percent sediment) as they move downstream. Figure 3 illustrates three textural fields, based on data from Mount St. Helens, through which cumulative curves undergo downward inflection as these changes in the flow occur. Textures of three hyperconcentrated flows from the Pasig-Potrero River plot within the field for hyperconcentrated flows at Mount St. Helens.

Figure 3. Cumulative curves illustrating textures (particle sizes and their distribution) of three deposits of hyperconcentrated flow in the Pasig-Potrero River (curves 1, 2, and 3), compared with a typical deposit of noncohesive debris flow at Mount St. Helens. Also shown are fields containing cumulative curves of deposits of (A) noncohesive debris flow, (B) hyperconcentrated flow, and (C) normal, sediment-laden streamflow, all based on data from Mount St. Helens. phi, a measure of particle size (see Folk, 1980).

A distinctive "transition facies" records the downstream dilution from noncohesive debris flow to hyperconcentrated flow (Scott, 1988, fig. 37); many examples of this facies were seen in the 1991-92 deposits of the Abacan and Sacobia-Bamban Rivers. Although nearly all hyperconcentrated flows in the Cascade Range evolved from upstream debris flows, that common origin there and the evidence from the transition facies do not mean that origin was the rule, or was even common, for hyperconcentrated flows during the 1991 and 1992 rainy seasons at Pinatubo. Observations of 1992 deposit sequences reveal many deposits of hyperconcentrated flows that may not have entrained enough sediment to reach the level of sediment concentration for rheological debris flow (fig. 4).

Figure 4. Deposits of 1992 hyperconcentrated flows in the Sacobia River inset against deposits from an older eruptive episode. Arrows mark the surface of the 1992 deposits; about 15 m of 1992 deposits occur in vertical section above the river. Location is 1 km upstream from the fanhead of the Sacobia River, November 15, 1992.

Regardless of their origin as primary hyperconcentrated flows or as the runouts of debris flows, flows with hyperconcentrations of sediment volumetrically dominate the flow system and depositional record downstream. Noncohesive debris flows and hyperconcentrated flows at Mount Pinatubo are highly erosive, on the basis of their observed ability to scour older deposits and to erode levees. Likewise, the sandy deposits of noncohesive flows are significantly more erodible than those either coarser or with more clay. Gravel-size fractions (coarser than 2 mm), which would increase erosion resistance of levee materials, are mainly deposited in upland channels as the noncohesive flows lose competence and transform to more dilute flow types. Thus, mainly erodible, sand-dominated deposits are locally available for levee construction at downstream sites, where the confinement of flow is most needed.

The deposits of hyperconcentrated flow are distinctive: a dominant size mode in the sand range, massive or poorly stratified beds typically 0.5 to 2.0 m thick, and sorting (a measure of size dispersion) intermediate between that of debris flow deposits and the sorting of deposits of normal streamflow. The loss of normal-density, gravel-size clasts shows that the flow has either lost, or never possessed, the strength necessary to suspend such clasts.

The textures of typical Pinatubo hyperconcentrated flows are shown in figure 3. Note that the downward inflection of the cumulative curves, and thus the transformation in flow type, is reflected primarily by loss of coarser fractions. A measure of sorting, graphic deviation (sigmaG, measured in phi units; Folk, 1980), typically varies from 1phi to about 2phi in hyperconcentrated flows. An exact range cannot be specified because of the strong influence of fine-sediment content on shear strength.

Hyperconcentrated flow deposits in the eastern drainages of Mount Pinatubo are dominated by sand-size phenocrysts from the pyroclastic flows with an admixture of mineral grains from older deposits. Pumice clasts are present but are volumetrically minor; most are preserved in coarse deposits near the source. Cumulative curves of deposits show that Pinatubo flows contain slightly more fine sediment that their St. Helens analogs (fig. 3). Most fine sediment at Mount Pinatubo, including devitrified glass and comminuted mineral and vitric material, is flushed through the fan environment to be deposited in axial lowland drainages and deltaic mudflats. This mineralogic fractionation is a partial analog of an example in the geologic record (Cather and Folk, 1991).

FIRST APPROXIMATION OF WATERSHED RESPONSES

Between the time of the climactic eruption, June 15, 1991, and the then-unknown time of rainy-season onset, it was necessary to characterize and prioritize how each drainage basin would respond to increased runoff. Resources were few, access problems were large because of destroyed bridges, and the rationale for moving additional groups into evacuation camps was weak unless the threat was immediate. The east-side watersheds were grouped by one of us (Janda) into three categories of response (table 1). This grouping, slightly modified from the initial version, was intended to be a first approximation, and, as such, it was successful. The categories were:

  1. Watersheds in which the sedimentation response would be immediate and persistent. Each was characterized by thick, fine-grained airfall distributed relatively uniformly on an erodible terrane. Pyroclastic flows entered some headwaters, but sediment yields from erosion of their deposits would not dominate the systems.
  2. Watersheds in which the flow response could be delayed but would also be large and persistent. Each drainage was characterized by large deposits of pyroclastic flows, abundant airfall deposits, and an erodible upland terrane. Thick deposits of pyroclastic flows created the potential for large sediment yields and massive downstream aggradation once drainage was integrated across the deposit surfaces.
  3. Watersheds in which the response would be immediate but less severe and of shorter duration than in the preceding categories. This response would be a function of the relative paucity of pyroclastic flows entering the drainages as well as the resistant underlying bedrock.

Table 1. Sedimentation response of watersheds on east side of Mount Pinatubo, as based on thickness and type of eruptive products and erodibility of terrane.

[See text for discussion of A-C]


Drainage

Upland drainage area (km2)

Slope
(m/m)

A. Immediate, persistent response1

 

 

     Bangat (or Bangut) River (major tributary of O'Donnell-Tarlac River)

51.6

0.04

     Marimla River (tributary of Sacobia-Bamban River)

65.1

.04

B. Delayed, massive, persistent response

 

 

     O'Donnell-Tarlac River

111.2

.08

     Sacobia-Bamban River

243.5

.06

     Abacan River (before stream piracy of July 25, 1991)

2,347.5

.04

     Pasig-Potrero River

23.9

.04

C. Immediate but less-severe response

 

 

     Abacan River (after stream piracy of July 25, 1991)

2,314.2

.04

     Porac River

29.9

.06

     North Fork Gumain River

47.0

.07

     Middle Fork Gumain River

 

 

     West Fork Gumain River

117.1 (both forks)

.07


1Category A is similar to category C, but it was assumed that response would be greater in category A because of larger tephra volumes.

2Most of 33.3 km2 diverted from Abacan River back to Sacobia River July 25, 1991.

3Excludes Taug River (Sapang Bayo Creek).

DEVELOPMENT OF DRAINAGE PATTERN AND FLOW-MARGIN LAKES

With the onset of the l99l rainy season in July, overflights showed an almost immediate transition from sheet wash and shallow, lahar-conveying channels on the valley-fill deposits to channels rapidly working their way upstream, almost exclusively by headcutting. Headcuts in the pyroclastic-flow deposits in the Sacobia River valley were estimated at greater than 10 m in height.

The longitudinal rates of headcutting, when seen during brief intervals of clear weather, were well in excess of 10 m per hour in some cases. Drainage of adjacent uplands was already integrated with the surfaces of the flow deposits, so that the process accelerated as the contributing drainage area increased exponentially.

The effects of secondary pyroclastic flows weeks, months, and more than a year later were an unanticipated circumstance (see Torres and others, this volume). The largest was reported in mid-August 1991 (U.S. Department of Defense, unclassified data, 1991) on the west side of Mount Pinatubo. That flow averaged almost a kilometer in width and extended at least 9 km. The largest 1991 flow on the east side occurred in the upper Pasig-Potrero River in early September and was probably more than 5 km in length. Another of similar size occurred July 13, 1992, in the same drainage. The thicknesses of these secondary flows were as much as several tens of meters. Triggering mechanisms are discussed by Torres and others (this volume).

Because the secondary pyroclastic flows occurred early in both rainy seasons, their occurrence triggered numerous lahars. An effect of the deposits of both flow types was the damming of flow from lateral tributaries. Several small lakes formed as a result of runoff gradually accumulating behind flow deposits emplaced in mid-June 1991. A typical example is shown in figure 5. This lake had discharged by July 21, 1991, and the blockage is seen in figure 5 (taken July 7) to be eroding both from upstream headcutting from a large explosion pit in deposits in the Sacobia River and by wave erosion or saturation-induced slumping of the blockage face.

Figure 5. Lake in tributary of Sacobia River dammed by pyroclastic flow deposits in Sacobia River (light area trending southwest-northeast at lower right). Note slumping on both sides of the blockage. Length of lake is approximately 0.4 km. Photograph taken July 7, 1991, by the U.S. Air Force.

The most significant blockage on the eastern side of Mount Pinatubo was that on the south side of the Pasig-Potrero River just 3 km upstream from the fanhead. A shallow lake was first observed there on July 24, 1991, and its clear water indicated recent impoundment. When next seen on August 1 the water was highly discolored by algal growth. The original blockage was a primary pyroclastic flow, lahars, or both. When the lake discharged in September 1991, the large secondary pyroclastic flow noted above had raised the blockage considerably, and the lake volume was greatly increased. The lake reformed, due to blockage by the 1992 secondary pyroclastic flow and associated lahars, and that lake discharged during a period of intense rainfall in late August 1992 (see Arboleda and Martinez, this volume).

CHANNEL DIVERSIONS

So thick were the fills of pyroclastic flows in some upland areas that, in many cases on the west side of the volcano (Rodolfo, 1991) and several cases on the east side, watersheds were filled and drainage divides were crossed. Drainage could potentially shift from the Sacobia-Bamban River to the Pasig-Potrero River, from the Sacobia-Bamban to the Abacan River, and from the Pasig-Potrero to the Abacan and Porac Rivers. Reversals of any diversions were clearly also possible but would be less likely once an integrated channel system was established. The distribution of even more extensive deposits from previous eruptive periods shows that similar diversions, of even larger scope, occurred during those periods.

The best documented channel shift of this type was drainage from the Sacobia-Bamban River system entering the Abacan River about July 15, 1991, and then mainly reverting to its preeruption course on July 25. As first observed on July 6 by J.J. Major and R.J. Janda (Major, written commun., 1992), the potential for diversion was clear but it had not yet occurred. A small channel that was headcutting from the Abacan River through the drainage divide had the potential to intersect the yet-embryonic drainage of the Sacobia-Bamban River. However, study of deposits in the Abacan drainage led Major and others (this volume) to suggest that a significant volume of flow may have entered the Abacan watershed from the Sacobia-Bambam River prior to the main pyroclastic flows of June 15.

Most of the upper Sacobia-Bamban River drainage (33.3 km2; table 1) was diverted to the Abacan River after July 6. Long periods, hours to days, of lahar surges were generated by relatively gentle rains (up to 6 to 8 cm per day on the eastern, lee side of the volcano). These flows traveled through Angeles City, reaching as far as the town of Mexico, 43 km downstream. As recorded by acoustic flow monitors (fig. 6), the rapid headcutting of the pyroclastic-flow deposits in the Sacobia-Bamban River then triggered recapture of most of the upper part of the river's drainage.

Figure 6. Avulsion of flow in upper Sacobia River from flow routed down Abacan River to flow routed normally to downstream Sacobia River, as recorded by acoustic flow monitors. A, Record of acoustic flow monitor recording flow in Abacan River on July 25, 1991. B, Record of acoustic flow monitor recording flow in Sacobia River on July 25, 1991. Arrows record time of completed channel diversion, 2100 hours, July 25, 1991.

The dramatic shift in acoustic signal at 2100 hours on July 25 records the exact time of the capture (fig. 6). Almost continuous flow of lahars then began in the Sacobia-Bamban River. The importance and at least short-term irreversibility of the capture were not confirmed until ground observations early on July 26. Figure 6 illustrates the suddenness of the diversion. For a detailed discussion of flow sensing with acoustic flow monitors, see Marcial and others (this volume).

A subordinate, southern part of the Sacobia-Bamban River continued drainage to the Abacan, but this pattern ceased in response to a secondary pyroclastic flow in the Sacobia-Bamban system in April 1992. All flow from the upper Sacobia-Bamban drainage was subsequently directed to the downstream part of that system.

CHANNEL CROSS SECTIONS

Repeated surveys of channel cross sections at bridge locations were critical in evaluating the locations and magnitudes of early channel changes. These data record the trends in erosion and deposition and thus yield important data for the planning of countermeasures. Selected examples of these surveys are shown in figures 7, 8, and 9. The significance of the changes is discussed and interpreted in the following sections by major drainage, from north to south. Survey sites and drainages are shown on figure 1. Burial of the section reference point at one site (Sacobia-Bamban River at Bamban) required a change from the local datum to mean sea level (MSL).

Figure 7. Cross-section changes in the O'Donnell-Tarlac River system, including the tributary Bangat River. Sites are as follows (locations on fig. 1): A, Bangat River at Santa Lucia, 1991; B, O'Donnell-Tarlac River upstream of Tarlac, 1992; C-1, O'Donnell-Tarlac River at Tarlac, 1991; C-2, same as site C-1, 1992; D-1, O'Donnell-Tarlac River downstream of site C, 1991; D-2, same as site D-1, 1991-92. Note variation in horizontel scale, sites B, D-1, and D-2.

Figure 8. Cross-section changes in the Sacobia-Bamban River system, including the tributary Marimla River. Sites are as follows (locations on fig. 1): E, Marimla River above confluence with Sacobia-Bamban River, 1991; F, Sacobia River above Clark Air Base, 1992; G, Sacobia River near Clark Air Base, 1992; H-1, Sacobia-Bamban River at Bamban, 1991; H-2, same as site H-1, 1992; I-1, Sacobia-Bamban River at San Francisco Bridge, 1991; I-2, same as site I-1, 1992.
Sites E, F, G
Sites H-1, H-2, I-1
Site I-2

Figure 9. Cross-section changes in the Pasig-Potrero River system. Sites are as follows (locations on fig. 1): J-1, Pasig-Potrero River at Mancatian, 1991; J-2, same as site J-1, 1992; K-1, Pasig-Potrero River at Santa Barbara Bridge, 1991; K-2, same as site K-1, 1992.
Sites J-1, J-2, K-1
Site K-2

O'DONNELL-TARLAC RIVER SYSTEM (INCLUDING TRIBUTARY BANGAT RIVER)

Bangat River at Santa Lucia (site A).--Although the Bangat River does not head near the crater, airfall and subordinate pyroclastic-flow deposits, emplaced on erodible bedrock (table 1), were rapidly flushed into downstream reaches. Nearly all channel capacity at the bridge site was lost with the 3 m of fill that occurred by August 23, 1991, one month into the first rainy season (site A, fig. 7). The magnitude of this response was expected (table 1).

O'Donnell-Tarlac River upstream of Tarlac (site B).--This site was surveyed only in 1992 (site B, fig. 7). Little change occurred early in the rainy season, but a uniform 1.5 m of fill occurred between August 13 and 20. The immediately following scour may have been natural, in response to dilute recession flows. Because scour this uniform across a channel is so atypical, it was more probably due to channel dredging. The channel level on October 14, near the end of the rainy season, was similar to that near the start of the rainy season.

O'Donnell-Tarlac River at Tarlac (where it is known as Tarlac River) (site C).--Gradual net scour occurred here, probably due to nearby channel dredging, during the 1991 rainy season (site C-1, fig. 7). By the early part of the 1992 season (site C-2, fig. 7), a uniform meter of fill had occurred, followed by almost 3 m of scour probably in response to dredging. In spite of the disturbance of the natural pattern by engineering work, the general lack of extensive fill both here and at site B shows that early channel response in the downstream part of this drainage was relatively benign compared to more southerly east-side drainages. However, the extensive fill observed upstream in the O'Donnell-Tarlac River can be expected to extend to these downstream reaches.

O'Donnell-Tarlac River at Tarlac (downstream of site C, also where drainage is known as Tarlac River) (site D).--The events at site C are confirmed by the same pattern of scour and fill at this location (site D; D-1 and D-2, fig. 7). Timing and causes of the channel changes are also similar, but changes are less pronounced, probably indicating a diminution of effects in a downstream direction. As at site C, fill will occur in the future, the amount dependent on engineering works and the effects of tributaries to increase sediment conveyance through these reaches.

SACOBIA-BAMBAN RIVER SYSTEM (INCLUDING TRIBUTARY MARIMLA RIVER)

Marimla River above confluence with Sacobia-Bamban River (site E).--Fill occurred rapidly at the start of the 1991 rainy season, and most channel capacity was lost by August 17 (site E, fig. 8). The rapidity of the fill reflects the higher base level and backwater effects caused by the extreme aggradation in the Sacobia-Bamban River. The confluence of the two streams is immediately downstream.

Sacobia River above Clark Air Base (site F).--Although not evident in the 1992 cross sections (site F, fig. 8), extensive fill at this site is documented by estimated levels of the fill surface late in 1992 compared with levels observed during the 1991 rainy season. At least 10 m, and locally as much as 15 to 20 m, of fill has occurred during this period at and near this location. The 2 to 5 m of scour seen in the cross sections as occurring during late July 1992 is a perturbation in the longer trend, evidenced in the subsequent fill at the following downstream site.

Sacobia River near Clark Air Base (site G).--The devastating volume of fill beginning in the upper part of the river system is illustrated by change at this site. Slightly over 10 m of 1992 fill is recorded at one side of the channel (site G, fig. 8). Approximately the same amount was observed late in 1991. Times of the 1992 surveys indicate that much of the fill occurred during major storm periods in late August and early September 1992. Similarity of the level of the low-water channel in late 1992 to the channel level prior to those runoff periods is mainly due to cutting by recession flows.

Sacobia-Bamban River at Bamban (site H).--Fill at this site of a major highway bridge began with the eruption, with approximately 10 m occurring before the first survey on July 23, 1991 (H-1, fig. 8); the bridge shown in the 1991 survey was originally at a height of 20 m; it failed from inundation by sediment-laden flows in late August 1991.

A new bridge was constructed above the remaining supports of the old bridge, and the new, higher bridge failed in the same way during the late August 1992 storm period. Almost 5 m of fill occurred across the channel during that 3-day period (H-2, fig. 8). Unfortunately, benchmarks for the two destroyed bridges cannot be compared, and the 1991 and 1992 surveys cannot be tied to the same datum. At a minimum, 25 m of fill has occurred locally at this site since the eruption. By the end of the 1992 rainy season, scour related to downstream channel maintenance had exposed the tops of the supports of the second bridge.

Sacobia-Bamban River at San Francisco Bridge (site I).--Scour at this site, 9.2 km downstream from site H, is in marked contrast to the upstream fill. The late 1991 scour is due to removal of sediment from the channel. Fill early in the 1991 (I-1, fig. 8) and 1992 (I-2, fig. 8) rainy seasons is the dominant natural process but is clearly of less magnitude than that upstream. This decrease is explained mainly by upstream losses of flow diverted from the leveed channel. Even so, the total cross-sectional area of fill at a given point in both the main channel and the diversions generally decreases downstream.

The channel downstream from Clark Air Base will be especially prone to diversion and levee breaks. This risk is due to the tendency toward radial drainage on the fan surface and to the trend toward fill in the engineered channels.

PASIG-POTRERO RIVER SYSTEM

Pasig-Potrero River at Mancatian (site J).--The 15 m of scour recorded here during August 1991 (J-1, fig. 9) is due to natural processes and channel maintenance. Scour began with flows associated with the eruption (see Major and others, this volume), and the Mancatian bridge was destroyed shortly thereafter. The early flows were almost certainly more dilute and thus more erosive than those later in the June-September period, reflecting the gradual integration of the drainage system on the vast pyroclastic flow deposits upstream and the consequent increase in sediment conveyance. With this general increase in sediment concentration of the flows, unremitting fill occurred between September 4 and November 21, 1991. This behavior coincided generally with the expected watershed response (table 1).

The 1991 filling continued unabated throughout most of the 1992 rainy season (J-2, fig. 9). Mining of large volumes of sand from the channel immediately downstream of Mancatian occurred during 1992 and slowed the rate of fill there. Channel dredging downstream also acted to slow the rate of 1992 filling, especially late in the rainy season as recorded in the surveys on September 1 and November 4, 1992. Runoff was also relatively low following the early September storm period noted in the section on flow-margin lakes.

The channel reaches from above Mancatian to the vicinity of Bacolor (fig. 1) are prone to channel diversions and levee breaks for the same reasons cited above in the case of the Sacobia-Bamban River below Clark Air Base.

Pasig-Potrero River at Santa Barbara Bridge (site K).--This site, located far from the mountain front and near the axial drainage of central Luzon, illustrates the fate of large leveed channels at the distal segments of river systems. This wide channel filled very rapidly early in the 1991 rainy season (K-1, fig. 9), and the bridge would have been destroyed but for loss of flow from the channel upstream.

This situation continued in 1992 (K-2, fig. 9). Large flows left the leveed channel system a short distance downstream of Mancatian and caused substantial damage in several small lowland communities. The channel at the Santa Barbara Bridge was maintained with sufficient capacity to retain the bridge, but most flows during major storm periods broke out of the leveed channel upstream. The channel at the bridge and for a considerable distance upstream is perched on fill at a level several meters above the surrounding countryside, so that losses from the channel are continuous and cause local flooding of areas adjacent to the channel.

LATERAL EROSION AT CROSS-SECTION SITES

Lateral or bank erosion is not addressed by the cross-section data because the survey sites are located predominantly at bridge locations, which are stable, bedrock-side-sloped sites or are massively protected against lateral erosion. Generally, however, more bridges fail from lateral erosion than from any other cause, but this commonly occurs because of lateral erosion of a pier footing. At nearly all cross-section sites at bridges or bridge sites, aggradation precludes either scour or lateral erosion of pier footings. Lateral change seen in the successive cross sections is relatively insignificant and represents the vagaries of intrachannel flow movement. However, rapidly aggrading channels are generally associated with high rates of lateral erosion, and observations away from the cross-section sites confirm this association. The hyperconcentrated flows conveying most of the volcaniclastic sediment from the uplands clearly can erode unprotected levees very rapidly.

CONCLUSIONS

Fluvial redistribution of the extensive eruptive products of the l99l eruption will continue to pose widespread risk to lowland areas surrounding Mount Pinatubo. The trends in channel filling, depositional site, and the types of flows are important factors in assessing the engineering alternatives for sediment-control measures. The most critical observations for immediate mitigation plans can be summarized in the following two categories:

1. Factors related to the alluvial-fan environment. In distributing the flood risk on active alluvial fans in semiarid areas, modern hydrologic practice assumes that risk extends across the entire surface of the fan. In all probability, this assumption extends to the tropical analogs of alluvial fans that surround Mount Pinatubo. The populated lowlands around the volcano are largely underlain by fluvially redistributed volcaniclastic sediment from eruptions prior to 1991.

Trends in channel cross sections show, after local 1991 syneruptive scour, large and continuing amounts of fill concentrated (and probably increasingly concentrated) in fanhead and upper fan areas. At least 25 m of fill is documented in a major river channel (Sacobia-Bamban River at Bamban). This trend will probably continue, and fanhead channels will fill and channel avulsion will occur. Repetitions of periods of aggradation and avulsion at the fanhead can distribute sediment across any part of an active fan. Thus, a levee system conveying flow from the mountain front is likely to be subject to burial or bypass.

These observations do not imply that existing fan channels with significant capacity will not continue to convey flows, over an intermediate time span of several years, to midfan locations where temporary levee systems can then route flow to sediment-storage sites. They do imply, however, that these channels are likely to be filled during the present aggradational cycle.

Any analysis of the situation must be viewed as a first approximation, with successive refinements to follow; thus, any mitigation plan will view effective sediment control as a "moving target." Control must be a running battle because the only permanent solution--huge sediment-retention structures tied to bedrock at the mountain front, with channels conveying the remaining flow downstream--is unlikely to be economically or topographically feasible at the scale required.

2. Factors related to flow type and deposit texture. The flows that convey volcaniclastic sediment to the populated lowlands around Mount Pinatubo are predominantly hyperconcentrated streamflows, and their deposits are dominated by sand-size sediment. The flows to date have been highly erosive, as observed in their ability to cut levees and by evidence seen in deposits of their ability to scour. The deposits of the flows are highly erodible because of the relative paucity of either cohesive fine material or material coarser than sand. Consequently, because the populated lowlands are widely underlain by the deposits of similar flows from previous eruptions, the availability of effectively erosion-resistant materials for levee construction is severely limited. This limitation is crucial to the economics of mitigation plans.

REFERENCES CITED

Arboleda, R.A., and Martinez, M.L., this volume, 1992 lahars in the Pasig-Potrero River system.

Cather, S.M., and Folk, R.L., 1991, Pre-diagenetic sedimentary fractionation of andesitic detritus in a semi-arid climate: an example from Eocene Datil Group, New Mexico, in Fisher, R.V., and Smith, G.A., eds., Sedimentation in volcanic settings: SEPM (Society for Sedimentary Geology) Special Publication 45, p. 211-226.

Ely, L.L., and Baker, V.R., 1990, Large floods and climate change in the southwestern United States: Hydraulics/hydrology of arid lands, Proceedings of the International Symposium, San Diego, Calif., July 30-August 2, 1990: Hydraulics Division, American Society of Civil Engineers, p. 651-656.

Folk, R.L., 1980, Petrology of sedimentary rocks: Austin, Tex., Hemphill Publishing Company, 182 p.

Janda, R.J., Major, J.J., Scott, K.M., Besana, G.M., Daligdig, J.A., and Daag, A.S., 1991, Lahars accompanying the mid-June 1991 eruptions of Mount Pinatubo, Tarlac and Pampanga Provinces, The Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 62.

Major, J.J., Janda, R.J., and Daag, A.S., this volume, Watershed disturbance and lahars on the east side of Mount Pinatubo during the mid-June 1991 eruptions.

Marcial, S.S., Melosantos, A.A., Hadley, K.C., LaHusen, R.G., and Marso, J.N., this volume, Instrumental lahar monitoring at Mount Pinatubo.

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.

Pierson, T.C., Janda, R.J., Umbal, J.E., and Daag, A.S., 1992, Immediate and long-term hazards from lahars and excess sedimentation in rivers draining Mt. Pinatubo, Philippines: U.S. Geological Survey Water-Resources Investigations Report 92-4039, 35 p.

Punongbayan, R.S., Besana, G.M., Daligdig, J.A., Daag, A.S., and Rimando, R.E., 1991, Mudflow hazard map: Philippine Institute of Volcanology and Seismology, 1 sheet.

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 eruptions [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 62.

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

Scott, K.M., Pringle, P.T., and Vallance, J.W., 1992, Sedimentology, behavior, and hazards of debris flows at Mount Rainier, Washington: U.S. Geological Survey Open-File Report 90-385, 106 p.

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

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.

U.S. Air Force, 1991, Terminal forecast reference notebook: Detachment 5, 20th Weather Squadron, Clark Air Base, Republic of the Philippines, individually paginated sections.

FIRE and MUD Contents

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

This page is <https://pubs.usgs.gov/pinatubo/kmscott/>
Contact: Chris Newhall
Last updated 06.11.99