FIRE and MUD Contents

Assessment and Response to Lahar Hazard around Mount Pinatubo, 1991 to 1993

By Richard J. Janda,¹ Arturo S. Daag,² Perla J. Delos Reyes,² Christopher G. Newhall,¹ Thomas C. Pierson,¹ Raymundo S. Punongbayan,² Kelvin S. Rodolfo,³ Renato U. Solidum,² and Jesse V. Umbal^{2 3}

¹ U.S. Geological Survey.

² Philippine Institute of Volcanology and Seismology.

³ Department of Geological Sciences M/C 186, University of Illinois at Chicago, 801 W. Taylor St., Chicago, IL 60607-7059.

ABSTRACT

Lahar hazard at Pinatubo is a function of prodigious sediment yield from Pinatubo's upper and middle slopes and the sediment storage capacity in the adjoining lowlands. Both are diminishing but at mismatched rates. Sediment yields set world records during the first three posteruption years, and yields in the Balin Baquero-Bucao and Marella watersheds may do so for several years more. In general, sediment yields peaked early and are decreasing rapidly in east-side watersheds, where the volume of 1991 pyroclastic-flow deposit is relatively low, deposits and streams are confined in a few steep-walled valleys, thin ash fall from secondary explosions is common, and vegetation recovery is fast. Sediment yields peaked later and are decreasing slowly in west-side watersheds, where pyroclastic-flow deposits are more voluminous, numerous small streams drain a broad, gently-sloping, unconfined pyroclastic apron, and vegetation recovery was initially slow.

We anticipate that slightly more than 3 cubic kilometers of sediment will move from the volcano's slopes into surrounding lowlands. By late 1993, almost two-thirds of this amount had already arrived in the lowlands; most of the remaining third will be from the Balin Baquero-Bucao and Marella watersheds. Sediment yield from the Gumain watershed virtually stopped in 1992, and that in the Abacan stopped because its 1991 headwaters were recaptured by the Sacobia in April 1992; 1991-93 yield in the Sacobia-Abacan-Pasig system was about three-fourths of the expected total.

Optimism that the worst is past, especially for the east side of Pinatubo, should be tempered by three factors. First, channels in the middle reaches of alluvial fans are filled, so the threat of lahars to several populated areas remains high. On unconfined alluvial fans of the Pasig-Potrero, Marella-Santo Tomas, and, perhaps, the Sacobia Rivers, sediment must continue to spread beyond present river channels, into catch basins, if the overall problem is to diminish. Towns on these fans that are beyond the reach of the most sediment-laden flows (debris flows) could be protected from more dilute hyperconcentrated flows and floods by relatively low ring dikes if flows are allowed to spread onto surrounding agricultural land, but that solution has not, to date, been politically acceptable. Second, unusually heavy rainfall, if sustained for several days or more, would temporarily reverse the declining sediment yield and cause serious overbank lahars. Third, erosion and incorporation of sediment predating the 1991 eruptions into current lahars will add an as-yet-uncertain volume to deposits, and reworking of lahar deposits themselves will move sediment problems downstream and fill some distal reaches of channels that have not heretofore been filled.

Throughout the first 3 years of the Pinatubo crisis, assessments and warnings of lahar hazard, followed by mitigating actions, saved many lives and some property. Long-range warnings, including hazard maps and briefings, identified communities at high risk and led some residents to move transportable belongings--sometimes even houses--to safe, high ground. Immediate warnings from manned watchpoints, supplemented by rain gauges and flow sensors, alerted remaining people to flee villages when their lives were at risk. Information about the nature and magnitude of lahar hazard was also an important basis for planning projects, both sociopolitical and engineering, to help residents through this difficult time.

Despite some notable successes, not all warnings were perfect, nor were all warnings heeded. An early excess of false alarms made residents of some areas doubt all alerts; in other areas, dikes and other sediment-control structures offered a false sense of security that delayed evacuations. Some costly but futile efforts at lahar mitigation could have been avoided. In some of these instances, better scientific information and better presentation of that information could have reduced unnecessary losses; in other instances, competing political, economic, and social factors limited the acceptance of scientific information.

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

INTRODUCTION

Lahars from Mount Pinatubo have been flowing into densely populated areas of central Luzon since the major eruption of June 1991, taking a small toll of lives but causing enormous property losses and social disruption (fig. 1) (Mercado and others, this volume; C.B. Bautista, this volume). Bank erosion and thick lahar deposits have left more than 50,000 persons homeless, and flooding and isolation have affected more than 1,350,000 people in 39 towns and 4 large cities. More than 1,000 km² of prime agricultural land is affected by or at risk from lahars, flooding, and siltation. The lahar problem also delays economic recovery, as some potential investors wait until the lahar problem subsides.

Loss of life, property losses, and low investor confidence can be minimized by accurate and timely assessment of the lahar threat and corresponding prudent actions by those at risk. In this paper we examine approaches to the lahar hazard used in the 3 years following the 1991 eruption, with the goal to understand what was done well and what might yet be improved. We describe the evolution of our scientific understanding of the lahar hazard at Pinatubo, note specific warnings that were issued, and describe how those warnings were (and can still be) used to mitigate risks from Pinatubo lahars.

LAHAR HAZARD ASSESSMENT TEAMS

Preeruption and syneruption assessments of lahar hazard at Pinatubo involved scientists from the Philippine Institute of Volcanology and Seismology (PHIVOLCS), the University of Illinois at Chicago, and the U.S. Geological Survey (USGS). The Pinatubo Lahar Hazards Taskforce (PLHT) was formed under the leadership of K.S. Rodolfo to assess and warn of lahar hazards, principally on the west side of Pinatubo. Members of the team were drawn from the University of Illinois at Chicago, PHIVOLCS, the Philippine Mines and Geosciences Bureau (MGB), and the University of the Philippines' National Institute of Geological Sciences (UP-NIGS). In 1992, the PLHT became the Zambales Lahar Scientific Monitoring Group (ZLSMG). Lahar hazard assessment on the east side of Pinatubo, from 1991 to the present, was handled principally by PHIVOLCS with assistance from USGS scientists. This division of responsibility between the two groups and the two sides of the volcano began as a matter of logistical necessity and organizational autonomy; over time, the two teams may be merged into one.

Within days after the climactic eruption, PLHT organized systematic observations of lahars from a watchpoint at Dalanaoan, San Marcelino (15 km southwest of Pinatubo, along the Marella River; fig. 1B), and, while road conditions still permitted, from a watchpoint at Malumboy, Botolan (27 km northwest of Pinatubo, along the Bucao River). At the same time, topical but less systematic observations of lahars were begun on the east side of Pinatubo by PHIVOLCS and USGS scientists, who were often interrupted by the continuing eruptions. By 1992, both teams were making systematic observations and meeting to exchange insights as often as possible.

Both PHIVOLCS and PLHT have assessed lahar hazards and advised government officials on proposed mitigation measures. Their analyses have been similar, though PLHT and its successor, ZLSMG, have been more outspoken on proposed engineering countermeasures. At times, relatively minor scientific differences have arisen between the two groups, and media attention has focused more on the messengers than on constructive portrayal of the scientific differences.

Other organizations have made independent lahar hazard assessments, including the Philippine Bureau of Soils and Water Management and the U.S. Army Corps of Engineers, as noted below.

Warnings by PHIVOLCS and others are provided to the Office of the President, the National Disaster Coordinating Council (NDCC), the Regional Disaster Coordinating Council for Region III (RDCC-III), the Department of Public Works and Highways (DPWH), the Department of Social Welfare and Development (DSWD), and to Provincial, Municipal, and Barangay Disaster Coordinating Committees (PDCC's, MDCC's, and BDCC's) (fig. 2).

DEFINING THE THREAT

A lahar is a rapidly flowing mixture of volcanic rock debris and water, typically with 40-90 percent sediment by weight, and thus having a consistency ranging from muddy water to a dense slurry. We recognize two types of lahars at Pinatubo, defined in greater detail by Pierson and others (this volume). The first, debris flow, has high viscosity, notable yield strength, and sediment concentrations typically greater than 60 percent by volume. The second, hyperconcentrated flow, has moderate viscosity, low yield strength, and sediment concentrations of 20 to 60 percent by volume.

All lahars from Pinatubo are caused, directly or indirectly, by heavy, seasonal monsoon rainfall that is enhanced by rain from tropical cyclones. The rainy season can begin as early as May and continue through November; monsoonal rains normally coincide with the typhoon season and are heaviest during June through September, but early- or late-season tropical cyclones can extend the lahar season. The June 1991 eruption provided the missing ingredients for lahars: a severely disturbed landscape in which runoff would be high and an abundant supply of loose, easily erodible sediment. Most lahars of Pinatubo begin as surface runoff of rainfall; a few begin by the sudden release of standing water that has ponded on or against the margins of other deposits. The flowing water rapidly entrains loose sediment from channel beds and banks and is transformed into a sediment-rich flow with the sediment concentrations noted above.

These generalizations, based mainly on climatic data for the Pinatubo region and on experience at Mayon and at other volcanoes around the world, were understood at the time of the June 1991 eruption. In addition, we needed to learn:

What was the approximate range of velocities, discharge, sediment content, temperature, and flow behavior to be expected for the Pinatubo lahars?

How would lahars vary from one watershed to the next?

What critical amounts of rainfall would be needed to trigger lahars of various sizes?

Preliminary data about the character of Pinatubo lahars came quickly. The flows of June 15, though not observed by our scientific team, were soon reconstructed (Major and others, this volume), and additional flows were soon observed (Pierson and others, this volume; Rodolfo and others, this volume; K.M. Scott and others, this volume). A working hypothesis was sketched out for three types of Pinatubo watersheds, based on their average slopes and intensity of eruption impact (see K.M. Scott and others, this volume). Other workers used oblique aerial photographs and preeruption topographic maps to estimate the volume of new pyroclastic deposits (then estimated to be 5-7 km³, now judged to be 5.5+-0.5 km³, W.E. Scott and others, this volume), and Pierson and others (1992) estimated that between 40 and 50 percent of this debris would be transformed into lahars over the next decade. Clearly, the lahar problem would be massive and persistent.

Semiquantitative details about individual lahars were filled in during the first lahar season, comprising the period from June through November 1991. Table 1 summarizes observations of lahars in the Sacobia River; elsewhere around Pinatubo, lahars have roughly similar characteristics, scaled upward or downward according to the size of the watershed and the volume of sediment that can be entrained.

Table 1. Summary of lahar observations, Sacobia River, 1991-93.

Information concerning the critical rainfall necessary to trigger lahars of various sizes came more slowly, as we had to install a network of rain gauges and observe enough flows to draw valid conclusions. During 1991 and 1992, about 6 mm of rainfall over 30 min (0.2 mm of rain per minute) was sufficient to trigger lahars in the Marella and Sacobia watersheds, and rain at double that rate triggered medium to large lahars, especially if that rainfall was sustained for several hours or if there had been other rain in the preceding days. Figure 5 of Tuñgol and Regalado (this volume) illustrates lahar-triggering thresholds for various intensities and durations of rainfall.

Total rainfall of about 2,000 mm (station MSAC) during 1992 resulted in production of about 1 x 10⁸ m³ of lahar (water+sediment) (M.T.M. Regalado, unpub. data, 1994; Tuñgol and Regalado, this volume). For amounts of rainfall experienced in 1991 and 1992, the total volume of a lahar appeared to increase linearly with rainfall during the event (Regalado and Tan, 1992; Tuñgol and Regalado, this volume). In the Sacobia-Pasig watershed during 1992, lahar yield (V, sediment+water) was about 1 x 10³ m³ /mm of rainfall/km² of upland watershed. The same relation held for individual storms and for the lahar season as a whole. Lahar yields per millimeter of rainfall per square kilometer were generally lower in other watersheds; details of lahar yields will be discussed in the section "Long-Term Warnings: Hazards Assessment and Hazards Maps."

By 1992 and 1993, we began to detect several changes that could be projected as trends--in total sediment yield, sediment yield normalized for watershed area and rainfall, channel filling, upstream and downstream migration of the principal reaches of deposition, and other parameters. Most of these are noted and discussed later in this chapter; perhaps the most notable of these changes are the decreasing sediment yields normalized for rainfall and watershed area. Experience elsewhere suggests that this decrease will probably be exponential (Pierson and others, 1992), though perhaps not all the way back to preeruption levels (Pierson and Costa, 1994). Figure 3 contains computer-fit, manually adjusted curves for that decrease, projected back to preeruption yields of roughly 10⁵ m³/km³/yr, still high in comparison to stable watersheds.

Figure 3. General decreases in sediment accumulation from 1991 to 1993. A, Volume of sediment accumulation (approximates sediment yield), per year. Symbols and lines for lahar years 1-3, measured values; dashed lines, projections, discussed in text. Shaded area, projection of Pierson and others (1992) for Pinatubo as a whole. B, Rates of sediment accumulation normalized for watershed area. C, Rates of sediment accumulation normalized for watershed area and rainfall. Lastly, we checked to see if our perception of the threat was consistent with what the geological record told us about the lahars from past eruptions of Mount Pinatubo, and it was. The general topography of gently sloping alluvial fans around Mount Pinatubo, soils maps of central Luzon, and inspection of exposures all suggest that much of the central valley of Luzon and all of the Santo Tomas plain owe their fertile soils and sediment to lahars and related floods of Pinatubo.

PUBLIC EDUCATION AND WARNINGS

PHIVOLCS and PLHT began to provide general public education about lahars, long-range warnings about what might be expected in coming months and years, and short-range warnings about lahars expected within minutes to days.

Education about the Threat

Before June 15, 1991, residents and local leaders of the Pinatubo area had only slight familiarity with lahars that had occurred at Mayon Volcano in 1984 and at Nevado del Ruiz, Colombia, in 1985. Fortunately, a videotape that was made by the late Maurice Krafft for the International Association of Volcanology and Chemistry of the Earth's Interior (IAVCEI), largely in response to the lahar tragedy in 1985 at Nevado del Ruiz, showed graphic, frightening, but realistic images of lahars and their effects on people. We showed this video to many decisionmakers and citizens, and the result was a noticeable increase in awareness and concern.

As described by Major and others (this volume), the first lahars formed from rainfall on preclimactic eruption deposits on the afternoon of June 14, almost 24 h before the climactic eruption. Much larger lahars occurred during that eruption, triggered by rain from passing Tropical Storm Yunya. On and shortly after June 15, people learned firsthand that lahars were raging torrents that destroyed bridges, eroded river banks, and flooded fields where riverbanks were overtopped. But most people still had little idea of the magnitude or persistence of the hazard that would be with them for years to come and had no concept of the insidious sediment buildup in channels that would soon lead to many more overbank lahars and floods.

Use of the term "lahar" was vigorously promoted on June 15, 1991, and in subsequent days by two of us (K.S. Rodolfo and J.V. Umbal), who were concerned that the previously introduced term "volcanic mudflow" misrepresented the material transported (mostly sand and coarse debris rather than mud), and, for that reason, gave people a dangerously understated sense of the threat (a similar concern was raised by Voight, 1988, 1990). Another purpose of introducing the term "lahar" was pedagogical: a catchy, unfamiliar term might (and did) get special attention. Indeed, the term lahar has now received so much attention that it has become a metaphor for practically any disaster in the Philippines.

Scientists targeted three groups for special education: news reporters, public officials (including civil defense officials and engineers), and police and army personnel assigned to lahar watchposts on the slopes of Mount Pinatubo. Instruction of the news media was provided in the course of day-to-day interviews and field trips, as reporters struggled to understand lahars. Educational posters were prepared, as were, later, several videos and a primer booklet on Pinatubo lahars.

Formal press conferences on lahars were rare (perhaps, too rare); one-on-one and informal group interviews with scientists were common, as was press coverage of scientists giving briefings to emergency meetings of public officials. Most press contacts were at PHIVOLCS' Main Office in Quezon City; in Zambales with the Pinatubo Lahar Hazards Taskforce; and in San Fernando, the site of many emergency meetings for public officials. Most interviews were with the chiefs of PHIVOLCS and PLHT, though a number of PHIVOLCS and PLHT staff served as spokespersons when needed.

Concern among both volcanologists and public officials was high. Throughout late June and July 1991, one of us (R.S. Punongbayan) briefed government officials, including then-President Corazon Aquino, about both the eruption and lahars. In mid-July, another of us (K.S. Rodolfo) met with officials of the Department of Public Works and Highways (DPWH) and stressed that past eruptions at Pinatubo and similar volcanoes had been followed by such large lahars and volumes of deposited sediment that equally or more serious hazards lay ahead. A week later, another of us (R.J. Janda) met with the same group, outlined how sediment would move from the steep slopes of each watershed and fill lowland channels, and indicated how monitoring of erosion and channel filling in each watershed could provide information critical to the mitigation of the sediment problem. DPWH officials requested a ranking of barangays (villages) at risk and an estimation of the volumes of sediment that could be expected. On August 30, two of us (T.C. Pierson and K.S. Rodolfo) met with DPWH Secretary Jose P. de Jesus and his staff to further explain the lahar threat and to offer to work with design engineers on proposed engineering countermeasures.

For barangay leaders and the general public, illustrated flyers, posters, and leaflets were prepared and distributed by PHIVOLCS, the RDCC, the Philippine Information Agency, and others. One poster, "Mga Dapat Gawin Upang Maiwasan ang Pagiging Biktima ng Mudflow" ("How to Avoid Becoming a Victim of a Mudflow"), offered suggestions for what residents should do before, and when, they are warned of lahars (fig. 4). Paraphrased, the poster said:

1. Stay away from potential mudflow channels when it is raining at Mount Pinatubo and nearby hills;
2. If you live in a lowlying place, move immediately to high ground, keeping in mind that mudflows go to areas that are usually flooded during the rainy season. Keep in mind that some rivers and streams are already shallow from materials from Mount Pinatubo's eruption, so it is possible that these rivers will overflow;
3. Make your own "hill" at least 4 m high and widen the top to serve as an evacuation center;
4. Make barriers, if possible, but be aware that mudflows can be fast and strong;
5. Each group of houses should have a watchperson on a nearby hill to sound mudflow alarms;
6. Be ready with flashlights and a radio;
7. Listen for warnings and pay attention to the authorities;
8. Stay calm and don't be fooled by false news or rumors. WE WON'T HAVE TO WORRY IF WE STAY ALERT AND STAY TOGETHER.

In the absence of suitable coarse riprap to protect against erosion by lahars, the advice to "build your own high ground" was later concluded to be unwise, and the advice was withdrawn before any manmade hills were actually built.

During the second lahar season (May-November 1992), PHIVOLCS released four new educational tools. The first was a pamphlet, "A Technical Primer on Pinatubo Lahars," written to explain the general behavior of lahars to public officials, the news media, and other interested groups (Punongbayan and others, 1992a). Perhaps the most useful feature of this primer was a distinction between "malapot" (sediment-rich, viscous) and "malabnaw" (relatively dilute, less viscous) lahars, corresponding roughly to debris flows and hyperconcentrated flows, respectively. The distinction made it clear that lahars were fundamentally different from normal streamflow and much harder to control than normal floods. Another useful concept introduced in the primer was attenuation of flows, so that people would understand that warnings of "2-m-deep flow" past manned watchpoints upstream from populated areas might only be "0.5-m-deep flow" or less in distal areas.

A second educational tool, requested by participants in the May 1992 International Scientific Symposium on Mount Pinatubo, was a set of impact scenarios for three representative storms (Punongbayan and others, 1992b). In order of increasing severity, the storms were (1) normal afternoon rainshowers, (2) prolonged monsoonal rainfall ("siyam-siyam," literally, 9+9 days), and (3) intense typhoon rains. Scenarios of future lahars and their impacts were chosen from the most likely actual events, such as burial of Barangay Tabun of Mabalacat town. For example, starting on Day 5 of a "siyam-siyam," the scenario said:

In Mabalacat, sediment now fills to within m of the flat surface on which Tabun and Dolores are built . . . . [Day 6] Rain continues . . . more lahars, all rivers. Bgy. Tabun is now "Natabunan" (covered) with m of sediment. Before taking a shortcut through Tabun, lahars washed out the Bamban Bridge . . . . [Day 7] Unexpectedly great erosion of the banks of the Pasig-Potrero below Mancatian, coupled with rapid buildup of sediment in the channel just downstream, now threatens Santa Rita, Pampanga . . . . [Day 9] The Pasig-Potrero breaks out, but to the east, into Potrero, rather than into Santa Rita . . . .

The only event in this scenario that did not actually occur in 1992 was breakout on the east bank of the Pasig-Potrero; rather, flow broke over the southwest bank and buried Barangay Mitla of Porac and Barangay Balas of Bacolor (near Santa Rita). Many other events in the scenarios occurred in 1992 or 1993. Ironically, because these scenarios were so accurate, we later wondered whether we should have presented them as forecasts rather than as hypothetical scenarios. Our data were too sparse to have predicted these events with customary scientific certainty, but were we certain enough by laymen's standards? Would "forecasts" rather than "scenarios" have had a stronger, more constructive effect in preparing communities for lahars? The answer to all three questions is, in retrospect, "yes."

A third educational tool was a booklet in Tagalog titled "Ang Lahar" (The Lahar), designed principally for barangay leaders (Philippine Institute of Volcanology and Seismology, 1992b). The booklet contained nontechnical information about lahars, hazard zones, and the RDCC-PHIVOLCS lahar-warning system (the last is discussed under Short-Range Warnings).

The fourth and perhaps the most important tool was a video companion to the booklet, also titled "Ang Lahar," that provided basic information about lahars and showed graphic examples of 1991-92 damage to towns around Mount Pinatubo. Separate documentaries by Manila TV stations showed similar footage, with an emphasis on social impacts, and two new scientific videos ("Lahars of Pinatubo," by K.S. Rodolfo and H. Schaal, University of Illinois at Chicago; and "Pinatubo Volcano: Lahars and Other Volcanic Hazards," by M.T. Dolan, Michigan Technological University) were released in 1994. All of these videos have been wonderfully effective in conveying concepts that, for nonscientists, words and maps cannot convey.

Long-Range Warnings: Hazards Assessments and Hazards Maps

Information about topography, previous lahars, watershed and channel characteristics, volumes of erodible sediment, trends in current activity, and probable patterns of rainfall have been translated into long-range warnings about areas that will be at the greatest risk over coming months to years. This information has been presented in the form of hazard maps, tables of estimated sediment yield, or general statements of changing hazard, such as channel capture or upstream migration of avulsion points. Long-range warnings have influenced some decisions about relocation of towns, possible engineering countermeasures, and general emergency planning. They have also given residents a chance to move portable property--machinery, furniture, appliances, personal belongings, harvestable crops, farm animals, and even dismantled structures--out of the expected paths of future lahars. Some examples of this information are given in the following section.

Lahar Season 1, June through November 1991

On May 23, 1991, PHIVOLCS distributed a page-size map of volcanic hazards to civil defense and political leaders. Lahars (called "volcanic mudflows" on the map) were represented as hachures and later as bold lines along each of the major river valleys. Areas of potential overbank flow were not indicated because we lacked the fine-scale topographic maps needed to forecast likely points of overflow and paths of such flows across the alluvial fans. Clearly, the main preeruption channels were the most likely paths for lahars; we could not forecast whether and where lahars might fill channels and avulse.

Immediately after the June 1991 eruptions, questions about the long-term outlook for lahars at Pinatubo included:

How big is the lahar problem?

How long will the problem last?

What areas will be at greatest risk?

Over the next several years, will lahar deposits cover entire alluvial fans of Pinatubo, including all towns thereon, or just selected parts of those fans?

Between August and October 1991, PHIVOLCS and PLHT prepared the first of what was to become a series of mudflow (lahar) hazard maps. Two from PHIVOLCS were on a 1:100,000-scale base provided by the National Mapping and Resource Information Authority (NAMRIA) (Punongbayan and others, 1991a,b). The first, released in August, showed a zone judged to be "subject to mudflows" as of July 30. The second, released in October, showed lahars that had occurred as of September 15 and a significantly larger zone "subject to mudflows." The October hazard assessment was based on the lahars that had occurred to date, topography, and a subjective judgment about how widely lahars might spread during the next 1 to 5 years in the absence of engineering countermeasures. A simplified version of the October map is included as data layer A on figure 5. The October map was the first of the lahar hazard maps to be printed and widely distributed and, as such, was very influential even though the evolving crisis required later expansion of its hazard zones. Many of the evacuation sites, relocation sites, and other elements of recovery were based on that map and its subsequent revisions.

Also in August and September 1991, detailed lahar hazard maps (at a scale of 1:36,400) were issued by PLHT for the Santo Tomas and Bucao Rivers (Pinatubo Lahar Hazards Taskforce, 1991a,b). These maps were based on lahars to date, topography, a qualitative approximation of future lahar volumes, and projected channel avulsions. The map for the Santo Tomas River (dated August 23, 1991) showed lahars as of that date and four hazard zones, as follows: (a) areas subject to lahars and moderate to heavy flooding; (b) areas subject to overbank lahars and moderate to heavy flooding; (c) areas prone to minor or moderate flooding; and (d) areas subject to lahars escaping from filled irrigation canals. A September 5, 1991, map for the Bucao River used a similar zonation. In effect, these zonal categories ranked lahar hazard from highest to lowest. The relatively large map scale enabled officials and residents to judge whether their barangays, town streets, and secondary roads were in relatively high or relatively low danger.

Starting in September 1991 and continuing through the first half of 1992, the USGS, PLHT, and PHIVOLCS made semiquantitative estimates of probable volumes of lahars for the coming decade (Pierson and others, 1992). By then, it was apparent that sediment was being shed into surrounding lowlands at approximately 0.5 to 1.0 x 10⁹ m³/yr from a total upland source area of about 540 km², or at about 10⁶ m³/km² of upland watershed area per year, one order of magnitude faster than at Mount St. Helens and at the previous historical recordholder, Sakurajima Volcano, Japan (Janda and others, 1984). By analogy from Mounts Galunggung and Kelut in Indonesia, Pierson and others (1992) estimated that 40 percent of primary pyroclastic deposits on the east side of Pinatubo and 50 percent of the larger volume of primary deposits on the west side would be eroded and redeposited by lahars within roughly 10 years after the 1991 eruptions. Total sediment yield was projected to be approximately 2.5 to 3.6 km³ of the estimated 4.8 to 7.0 km³ of primary source material. Pierson and others (1992) also used aerial oblique photographs to estimate that 10 to 15 percent of most 1991 pyroclastic flow deposits had been eroded by September 10, 1991, and then used that volume of sediment as the first-year value on a sediment-delivery-rate decay curve, constructed from Mount Galunggung and Mount St. Helens data. Integration under that curve projected 1.2 to 2.5 km³ of sediment yield during the first posteruption decade (fig. 3A).

Using the additional assumption that lahars would spread in fans and deposit to an average thickness of about 2 m, the projected 2.5 km³ of sediment was translated into areas that could be covered by lahars during the coming decade (Pierson and others, 1992; also shown as data layer B on fig. 5). Potential hazards from future lahars (or from related backflooding and siltation) were mapped as being high or moderate, in relative rather than absolute terms. Whether lahars would actually impact this broad an area will depend in part on the actual advent and paths of typhoons, which deliver the most intense rains of lahar-generating duration. Preliminary copies of this 1:250,000-scale long-range hazards map were given to the Secretaries of National Defense and Public Works and Highways in mid-September 1991.

In October 1991, the Bureau of Soils and Water Management issued a GIS-based map of "Mudflow and Siltation Risk," showing a "high-risk area" subject to "moderate to severe mudflows and/or siltation" (data layer C on fig. 5), a "low risk area" subject to "low siltation," and a non-risk area (high ground) (Bureau of Soils and Water Management, 1991). The "high risk" zone of this map is much larger than that of any of the preceding maps, because it is a map of the distribution of sandy soils around Pinatubo. This map serves a useful purpose of reminding us that Pinatubo has, indeed, been the source of the fertile sandy loams throughout much of central Luzon, but it neglects to note that this sediment has been supplied by many eruptions over geologic time (>35,000 years, Newhall and others, this volume). Thus, risk appears to be exaggerated on this map. Given the relatively modest scale of the 1991 eruption in comparison to previous eruptions of Pinatubo, we do not expect the "high risk" zone of this map to be fully covered during the next decade.

One awkward role for PHIVOLCS geologists began during late 1991--responding to requests to "certify" that specific parcels of land were safe from lahars before new construction, loans, or other uses could proceed. Hazard maps were the primary guide for geologists, but, as has been indicated, several maps had been issued at various scales, and they were not understood by all applicants, not detailed enough for all purposes, and not legal documents. It was entirely appropriate for PHIVOLCS to examine sites of proposed evacuation camps and other major public projects, including one unsuitable resettlement site (Rabanes, San Marcelino) on which infrastructure was built, despite rejection of that site by PHIVOLCS, and which was destroyed by lahars in 1993. However, the task of checking hundreds of parcels has been burdensome, especially for parcels that are neither obviously safe nor obviously doomed. Risk is rarely "black or white," but present procedures in certification demand that it be considered so.

At the end of the first lahar season, PHIVOLCS and PLHT combined data from aerial photographs and field measurements and estimated 8 x 10⁸ m³ of deposited 1991 lahar sediment. This corroborated the estimate by Pierson and others (made in 1991, published in 1992) that between 5 and 10 x 10⁸ m³ of sediment would be moved into lowland areas in 1991. The estimated volumes of 1991 lahar deposits, categorized by river system, were provided to the Department of Public Works and Highways on March 24, 1992, for the purposes of design and decisions on engineering countermeasures.

On April 4, 1992, lahar hazards along the Abacan and Sacobia Rivers changed dramatically. A secondary explosion and pyroclastic flow from valley-filling 1991 deposits allowed the Sacobia River to capture drainage that had been entering the Abacan River (Martinez and others, this volume; Torres and others, this volume). Thus, the hazard along the Abacan decreased sharply, while that along the Sacobia increased. Some residents of villages along the Sacobia voiced suspicions that residents of Angeles City (along the Abacan) had caused the change, but the capture had been wholly natural, and it returned drainage to its preeruption pattern.

Lahar Season 2, May through November 1992

During 1992, lahars did not reach as far as they had in 1991, and the loci of deposition migrated up most of the alluvial fans. The reasons for these two changes are still under study; factors may include faster and greater runoff in 1991 than in 1992, channel filling and thus an increasing number of channel avulsions, and distal decreases in stream gradient. The effect of this migration was that towns like Concepcion and Bacolor (figs. 1B, 5), hard hit during 1991, were spared in 1992, while upstream barangays of the towns of Mabalacat and Porac were hard hit in 1992 and 1993.

In July 1992, in response to ever-changing conditions and a request for more specific, barangay-by-barangay assessments of lahar hazard, PHIVOLCS released a revised lahar hazard map on a large-scale (1:50,000) ozalid base that showed many, though not all, barangays at risk (Philippine Institute of Volcanology and Seismology, 1992a) (data layer D in fig. 5). On this map, hazards zones for the next 1 to 2 years were estimated on the basis of lowland topography (in which contours were a relatively coarse 20 m, with some intermediate 10-m contours), proximity to nearly filled river channels, volume of sediments to be expected, and a subjective estimation of how far and how widely overbank flows might spread. Two hundred and sixty nine barangays that fell within these newly defined hazard zones were listed, in an effort to help those unaccustomed to reading maps. At least one breathless radio announcer told listeners that all 269 barangays were at dire risk during the next rainstorm, and many people have found it difficult to understand that a hazard map with the words "subject to" is not an absolute prediction that all of the high hazard zone will get buried, but, rather, a geologist's graphic shorthand to indicate that actual flows over a period of years could but not necessarily would go anywhere within that zone.

Also on this map, hazard zones for the O'Donnell-Tarlac, Sacobia-Bamban, and Pasig-Potrero Rivers were larger than indicated in the October 1991 assessment, on the basis that 1992 lahars could have an equivalent volume to those of 1991 (8 x 10⁸ m³). Hazard zones for the Abacan and Gumain Rivers were reduced, reflecting the April 1992 capture of the upper Abacan by the Sacobia River and nearly complete erosion in 1991 of source materials in the Gumain drainage. Hazard zones for the Santo Tomas and the Bucao drainages were roughly the same as those in the August and September 1991 maps of PLHT (1991a,b). Page-size versions of these hazard maps were distributed with a corrected list of barangays at risk.

In late August 1992, the National Economic Development Authority (NEDA) and PHIVOLCS published a set of GIS-based maps of lahar hazard (Philippine Institute of Volcanology and Seismology and the National Economic Development Authority (PHIVOLCS-NEDA), 1992a). Hazard zones were modified only slightly from those of PHIVOLCS' July 1992 map (PHIVOLCS, 1992a), but the base map was substantially improved by including barangay and town boundaries. Important advantages of the GIS-based maps are that they can be revised quickly if conditions at and around the volcano change and that information about lahar hazard can be overlain by "data layers" for population, infrastructure, land use, evacuation routes, and myriad other parameters. An update of this PHIVOLCS-NEDA map, at a scale of 1:200,000, was issued on December 7, 1992 (PHIVOLCS-NEDA, 1992b).

The period after the 1992 lahar season was a time for scientists to analyze data they had gathered during 1991 and 1992 and to consider some new questions, including:

Was rainfall experienced during 1991 and 1992 below, near, or above the long-term average rainfall for these areas?

Had sediment yield at Pinatubo passed its peak and already begun to decline? (Sediment yield appeared to decline from 1991 to 1992, and we were asking whether we were past the period of peak sediment yields or just seeing an aberration due to unusually light rain.)

Unfortunately, discontinuation of rainfall monitoring at Clark Air Base in November 1991 and problems with telemetered rain gauges high on the west side of Pinatubo in 1992 prevent a direct comparison of rainfall for these 2 years. If we use rainfall at Dagupan as a proxy for that at Clark (table 2), we might conclude that rainfall on the east side during 1991 and 1992 was close to the long-term average, as was that on the west side (at Cubi Point). Rainfall high on the west slopes (fig. 8, stations BUG and QAD) was even higher than that of Cubi Point during 1991; we do not have comparable data for 1992, nor do we have any long-term average for BUG and QAD to tell whether the high rainfall in 1991 was above or below that average. Thus, we can say only that rainfall in 1991 and 1992 was not demonstrably different from long-term averages.

Field measurements during late 1992 suggest that, on the east side of Pinatubo, the volume of 1992 lahar deposits was about 40 percent that of 1991 (table 3). However, similar measurements on the west side suggest 1992 sediment volumes that were nearly equal to those of 1991. Thus, even though sediment yield on the east side appeared to have peaked and to have begun to decrease, we could not say the same for Pinatubo as a whole.

Was the difference between east and west related to rainfall? Again using Dagupan as a proxy for Clark and using Cubi Point to represent the west side, the ratio of Rainfall₁₉₉₂/Rainfall₁₉₉₁ was about 1.3 on the east and 1.0 on the west. This small, insignificant difference suggests that differences in rainfall did not account for the different sediment behavior from east to west during 1991 and 1992. Rather, we think that differences in watersheds, including differences in volume of 1991 pyroclastic debris, slopes and stream channels, and vegetation recovery, were responsible for different patterns of sediment yield. Declining yields on the east side appeared to be the start of the anticipated exponential decline in sediment yield (Pierson and others, 1992).

Table 2A. Abbreviated comparison of rainfall (in millimeters) during 1991-93.

[-, not known; ( ), average derived from 1991-93 data only]

¹ East-side lowlands are represented by Clark Air Base (1953-90) and Dagupan (1951-92) stations. Annual totals are averages of Clark Air Base and Dagupan; 24-h and monthly maxima are the maxima recorded at any single station.

² East-side highlands are represented by stations PI2, MSAC, FNG.

³ Southwest lowlands are represented by Cubi Point Naval Air Station (1955-87). 1992-94 data courtesy of Jun Tonpong, PAGASA/SBMA weather station, Cubi Point. Rainfall for 1993 was a record high since data collection began at Cubi Point in 1955.

⁴ Southwest highlands are represented by BUG, QAD; west-side highlands are represented by KAM (1993) and by average of BUG and PIE2 (1991-92). Annual values are averages of available data; monthly and 24-h values are maxima.

⁵ During Typhoon Goring, 0406 June 26 to 0308 June 27, 1993, at QAD. An even greater 24-h amount, 744 mm, was recorded from 0600 July 25 to 0600 July 26, 1994 at the same station (QAD).

Table 2B. Rainfall (in mm) at selected PHIVOLCS/USGS rain gages on and near Mount Pinatubo, 1991-93.

[-, Months before or after recording, or with seriously incomplete data. Adjacent values in parentheses are the recorded, minimum values. Data compiled by Ma. Theresa M. Regalado, PHIVOLCS]

Table 3. Volumes of source material, erosion, and lahar deposition, Mount Pinatubo watershed, 1991-93

¹ Upland area that contains, or drains through, valley-filling pyroclastic-flow deposits. In parentheses are areas of the valley-filling pyroclastic-flow deposits prior to erosion (after W.E. Scott and others, this volume).

² Volumes as estimated by W.E. Scott and others (this volume). In the O'Donnell and Sacobia-Abacan watersheds, Scott and others estimated slightly lower volumes than cited in Punongbayan and others (1994); in the Marella-Santo Tomas and O'Donnell watersheds, Scott and others inferred slightly larger volumes than estimated by Punongbayan and others. Estimates of Scott and others were made by sketching new levels of valley-filling pyroclastic-flow deposit onto preeruption topographic maps, estimating average cross-sectional thickness of 1991 deposit, and multiplying by the length of each reach.

³ Volumes are of deep channels cut during 1991-92 into 1991 pyroclastic-flow fans and are estimated by the U.S. Army Corps of Engineers (1994).

⁴ Most volumes are as estimated by PHIVOLCS and PLHT/ZLSMG lahar teams, from field measurements of accumulated sediment. Volumes for O'Donnell (1991-93) and Sacobia in 1991 are from USGS team members. All estimates have high uncertainties.

⁵ Forecasts based on figures 3A (annual sediment yield) and 3C (annual sediment yield normalized for upland watershed area and rainfall). In figure 3A, we assume that sediment yield from the Balin Baquero-Bucao watershed will decay faster than at Mount St. Helens, at rates intermediate between those of the Balin Baquero-Bucao and Sacobia-Pasig watersheds from 1991-93. Annual sediment yield will decline to preeruption levels (in the order of 10⁶ m³ /yr) in 20 yr. This assumption strongly influences the overall sediment forecast because of the large sediment contribution from the Balin Baquero-Bucao watershed. The decay rate for the Marella-Santo Tomas watershed was assumed to be intermediate between those of the Sacobia and the Balin

Baquero-Bucao watersheds. In figure 3C, we made a similar assumption, that the normalized rate of sediment yield from the Balin Baquero-Bucao, the slowest of any watershed, would reach 10 m³ _deposit /mm_rain /km² _watershed in 20 yr. Additional details are explained USGS-PHIVOLCS (1994).

⁶ To understand watershed responses and the overall east-side hazard, we must treat the Sacobia, Abacan, and Pasig-Potrero as a single watershed. For engineering measures, they must be treated separately.

⁷ Watershed areas are for conditions prior to secondary explosion in October 1993 that allowed the Pasig-Potrero to capture about 21 km² of the Sacobia watershed. After that capture, upland watershed for the Sacobia was about 25 km² and that for the Pasig-Potrero was about 44 km² . We cannot say whether or when the Sacobia might recapture its own upper watershed. Therefore, our sediment forecasts for the Sacobia-Bamban and Pasig-Potrero are given as high and low estimates that apportion the expected total sediment forecast between the two rivers in 1/3:2/3 and 2/3:1/3 ratios, based on approximate proportions of upland watershed areas in the event of capture or recapture. The high estimate for the Sacobia and the low estimate for the Pasig assume recapture by the Sacobia early in 1994; the low estimate for the Sacobia and high estimate for the Pasig assume that recapture does NOT occur. Our forecast for the Sacobia-Pasig system as a whole is unaffected by whether recapture does or does not occur, though it might be an underestimate because new, larger runoff will tend to widen narrow channels of the Pasig-Potrero that were originally eroded by the much smaller runoff.

⁸ Estimate is from U.S. Army Corps of Engineers (1994), based on photogrammetry. All other estimates of pyroclastic-flow volume by the Corps of Engineers were close to prior estimates of PHIVOLCS and those of W.E. Scott and others (this volume).

Lahar Season 3, May through November 1993

Among the questions we asked in 1993 was:

Did erosion of pre-1991 source materials, now apparent in some river valleys, add significantly to the volume of deposits to be anticipated?

Along some stream valleys, erosion had already cut through 1991 pyroclastic-flow deposits and into older deposits. Did sediment that was eroded from deposits predating 1991 add significantly to the volume of lahar deposits in 1992, above that anticipated in early studies? In the Sacobia-Abacan watershed, the volume of 1991-92 lahar deposits (270x10⁶ m³, table 3) exceeded the volume of 1991-92 erosion of the June 1991 pyroclastic fan (about 138x10⁶ m³, table 3). Could the difference represent the volume of debris predating 1991? An even greater contrast arose in the Pasig-Potrero watershed, where only 23x10⁶ m³ was estimated to have been eroded from the 1991 pyroclastic-flow fan, although about 90x10⁶ m³ of sediment was deposited downstream (table 3). If these values are extrapolated through 1993 and into the future, the ultimate volume of lahar deposits might be substantially greater than originally estimated. However, similar comparisons in the Balin Baquero-Bucao watershed show the opposite relation; that is, the volume of deposits is less than the volume eroded from 1991 pyroclastic-flow deposits (table 3). Either the uncertainties in estimating volumes are so great that these differences are within normal estimation error or differences between watersheds on the east and the west sides of the volcano have led to earlier and faster erosion of pre-1991 materials on the east than on the west. Clearly, uncertainties in estimating volumes are large, but we think that advanced erosion in east-side watersheds has also cut significantly into underlying, pre-1991 deposits. Should deposits that predate 1991 on the west side of Pinatubo ultimately be eroded to the same extent, the lahar problem in west-side watersheds might be 50 percent larger than originally estimated.

Another question we asked in 1993 was:

Was the rate of sediment deposition in 1993, by watershed and total, declining from rates in 1991 and 1992?

The volume of lahar sediment deposited on the east side of Pinatubo during 1993 was approximately 130x10⁶ m³, about 85 percent of that deposited in 1992 and one-third of that deposited in 1991 (table 3). The Pasig-Potrero sediment accumulation was anomalously high in comparison to other east-side yields, probably because the watershed's "clock" was reset by the large secondary pyroclastic flow of 1992 and because a large secondary explosion on October 6, 1993, captured a significant amount of drainage from the upper Sacobia. The volume of sediment deposited on the west side of Pinatubo during 1993 was approximately 375x10⁶ m³, about 85 percent of that in both 1991 and 1992 (table 3). Thus, sediment yields declined noticeably on the east but slowly on the west. Similar trends appear when sediment yield is normalized for watershed area (fig. 3B).

Sediment yields per unit of rainfall (table 4) also show declining trends. Despite the fact that sediment yield per millimeter of rainfall (per square kilometer of watershed area) varied considerably from one drainage to the next (some of the apparent variability is surely the result of poorly constrained volume estimates), the average sediment yield of Pinatubo clearly decreased from 1991 to 1993 (from 1.5x10⁶ m³_deposit/km²_watershed and ~460 m³_deposit/mm_rain/km²_watershed in 1991, to 0.9x10⁶ m³_deposit/km²_watershed and ~300 m³_deposit/mm_rain/km² _watershed in 1993 (table 4, figs. 3B,C).

Table 4. Sediment yields normalized for Pinatubo watershed area and rainfall, 1991-93.

[na, not applicable]

¹ Rainfall data compiled by M.T.M. Regalado, PHIVOLCS. Although data on rainfall in the Balin Baquero-Bucao watershed are especially sparse, data from the Kamanggi station in 1993 suggest that rainfall in that watershed is intermediate between that of the southwest sector (BUG) and that of the northern sector (ODON and PIE2).

² Data sources as in table 3.

³ Normalization for watershed area and rainfall is by simple division of sediment yield by those factors for the corresponding year.

⁴ Totals and averages include the Gumain, even during 1992 and 1993. Rainfall average is weighted according to watershed area.

This general trend toward decreasing sediment yield at Pinatubo has two important corollaries: that both the numbers and sizes of individual lahars, and the annual accumulations of sediment, will continue to decrease noticeably over the next several years. The precise rates at which they diminish will depend on storm events and annual rainfall, on channel captures such as that in 1993 which shifted some flow from the Sacobia watershed into the Pasig-Potrero, and on broad differences in watershed response. As regards differences in watershed response, three watersheds illustrate the range of possible behavior. The Gumain River, with a relatively small amount of 1991 pyroclastic flow-material in a generally steep watershed, was reamed out in 1991 and has produced very little sediment since 1991 (table 3, fig. 3A). Subsequent rapid recovery of vegetation in the Gumain watershed had little effect on sediment yield because source material was already exhausted. The Sacobia-Pasig watershed, with an intermediate amount of 1991 pyroclastic-flow material, large channels in a steep-walled valley, frequent secondary ashfall, and intermediate degrees of vegetation recovery, produced its peak sediment yield in 1991 and now shows a well-defined trend of decreasing sediment yield: from 1610 m³_deposit/mm_rain/km²_watershed in 1991 to 608 m³_deposit/mm_rain/km²_watershed in 1993. In contrast, sediment yield in the relatively gentle-terrain, pyroclastic-flow-rich, slowly revegetated Balin Baquero-Bucao watershed may or may not have peaked as of the end of 1993. Its future decline, though ultimately certain to occur, cannot yet be quantified. In the absence of any data to constrain its rate of decline, we assume that sediment yield will have declined to 1x10⁶ m³/yr (and to 10 m³_deposit/mm_rain /km²_watershed) after 20 years (dashed lines for Balin Baquero-Bucao, figs. 3A,C).

By using the computer-selected rates of declining sediment yield for the Sacobia-Pasig and O'Donnell Rivers, by choosing a rate of decline for the Marella that is intermediate between those of the Sacobia-Pasig and Balin Baquero-Bucao, and by assuming that further sediment yield from the Balin Baquero-Bucao watershed will decrease at the rates shown in figures 3A and 3C, we can make forecasts of eventual sediment yield for each watershed and for the volcano as a whole (last column of table 3). Though some assumptions are still required, these new forecasts are based increasingly on actual data, rather than on an assumed rate of decay that was adopted from volcanoes with much smaller eruptions than that of Pinatubo. Uncertainties are high in a forecast based upon only three annual measurements, but we think the forecasts are, nonetheless, an improvement over our original projections. The estimates are also consistent with observed differences in sediment yields from one watershed to the next. The new forecasts can serve for 1994 and will be revised as needed in subsequent years.

Interestingly, these new forecasts that are based on actual Pinatubo data are very similar to earlier forecasts that were based on an analogy with other volcanoes (Pierson and others, 1992). The only significant differences are in east-side watersheds where, apparently, incorporation of a significant volume of pre-1991 material is suggested by the abovementioned discrepancy between volume of erosion on the pyroclastic fan and volume of lahar deposits. Whether erosion of pre-1991 material on the west side will be as extensive as that on the east remains to be seen. If it is as extensive, the ultimate volume of sediment will be higher than that forecast here.

A third question we asked in 1993 was:

Was the upstream migration of sedimentation, noted in 1992, continuing?

Upstream migration of deposition continued in the Pasig-Potrero and Sacobia valleys through 1993, and avulsions above Mancatian on the Pasig-Potrero River brought lahars into that village for the first time. In contrast, some deposition in the Santo Tomas valley shifted downstream and toward the low southern margin of the Santo Tomas lahar field, largely as spillover from continuing deposition in the Marella valley. On July 1, 1993, ZLSMG released a revised, 1:50,000-scale lahar hazard map for the western Pinatubo area that identified points where lahars were likely to overtop and breach the dike that was being constructed along the south bank of the Santo Tomas River. This map accurately anticipated the site of an August 19 breach at the Western Luzon Agricultural College (WLAC). The map was modified after that breach and was modified again after lahars of October 4-7 breached the repaired dike at WLAC once again, as well as at other correctly predicted points at its eastern and western ends (Zambales Lahar Scientific Monitoring Group, 1993). Construction of dikes along the Santo Tomas River has resulted in faster, areally restricted deposition between Dalanaoan and San Marcelino proper than would have otherwise occurred, and, when breakouts have occurred, the artificially high origin points of the breakouts may have contributed to the distance traveled by breakout lahars.

All of the preceding long-range assessments of Pinatubo lahars were based on measured 1991-93 change at Pinatubo, supplemented by experience elsewhere. As a cross-check, we also asked:

Is the long-range assessment shown in the abovementioned hazard maps, based on events of just 3 posteruption years, consistent with what the geologic record tells us about entire periods of lahars following previous eruptions of Mount Pinatubo?

Pumiceous and sandy deposits in the lowlands of central Luzon are derived from lahars and related streamflow and overbank flooding. The sediment fans that surround Mount Pinatubo consist of layer upon layer of lahar and other stream deposits. Pinatubo sediments can be recognized beneath most cities and towns of Tarlac and Pangasinan Provinces, all the way northward to the Lingayen Gulf; beneath most cities and towns of Pampanga, all the way southward to Manila Bay; and beneath all towns of the Santo Tomas plain (Castillejos, San Marcelino, San Antonio, San Narciso, San Felipe) and Bucao plain (Botolan), in Zambales. Smaller parts of other provinces--Bataan, Bulacan, and Nueva Ecija--are also underlain by Pinatubo-derived sediment.

Prehistoric terraces throughout the Pinatubo area represent previous "high-stands" of sediment-rich flows. Examples include the terrace upon which Clark Air Base is built, one or more terraces immediately south of the town of Porac, several terraces along the Marella-Santo Tomas Rivers including those upon which the Aglao and the Palan evacuation camps are built, terraces along the Bucao River including that upon which San Juan and the main part of Botolan town are built, and terraces along the O'Donnell River upon which the old Crow Valley bombing targets were built (now, mostly buried) and on which Patling and Santa Lucia, Capas, are constructed.

Does this wide reach of Pinatubo sediments, areally and vertically, mean that all of these areas will be covered by lahar or flooding before the current crisis ends? We do not believe this is likely, and to draw finer distinctions, we asked:

What areas were covered after single previous eruptions of Pinatubo comparable in size to that of 1991 (specifically, after the penultimate "Buag" eruptions of 500 years ago)?

What are the ages of sediments in terraces that are still unburied, at various heights above present levels of fill, and were those sediments deposited only after much larger eruptions than that of 1991?

Which of these areas were covered by lahars, and which were covered by less lethal, but still troubling, floodwaters?

Radiocarbon ages for 25 lahar and related flood deposits (Newhall and others, this volume) help to distinguish between areas that were buried after the 500-yr B.P. Buag eruptions and areas that remained untouched after those eruptions. In general, in a clockwise direction from Tarlac Province around to Zambales Province, low-lying areas of Bamban and Capas were flooded and (or) buried by lahar and fluvial sediment of the Buag eruptive period, while slightly higher land around Mapalacsiao (Hacienda Luisita) and Clark Air Base was not covered by sediment of that period. The youngest known lahars to cross Hacienda Luisita occurred during the Crow Valley age eruptions, 5,000 to 6,000 years ago, while the youngest known lahars to cover the whole of Clark Air Base occurred during the Maraunot eruptive period, about 2,700 years ago. Some lahars did cover the Friendship area, and perhaps part of Clark Air Base, during the Buag eruptive period. Information for lahar deposits of the Pasig-Potrero River, on either side of Porac, is less well defined, but we are sure that the entire Porac area was covered by Maraunot lahars (3,000-2,500 years ago), and we also suspect that some of the famed sand- and gravel-mining of Porac might have been of Buag deposits. Terraces of the Floridablanca area, including that upon which Basa Air Base is built, are even less well known, but it appears from pyroclastic-flow deposits high in the Gumain watershed that the latest voluminous fill of the Gumain valley (and hence the most likely age of the terraces, below) was during the Pasbul eruptive period, about 9,000 years ago. However, some of the terrace deposits of Floridablanca could have been derived by flow in the Porac River, which in the recent geologic past has captured flow from the Pasig-Potrero; therefore, we would not be surprised if the youngest terraces of the Floridablanca area are the same age as those near Porac, 3,000-2500 years or younger.

Along the Santo Tomas plain, only the highest prehistoric lahar terraces remain unburied as of this writing, including a narrow terrace near Sitio Palan, and one or more broad terraces above Aglao and below Kakilingan. Between the present channel of the Santo Tomas River and Castillejos, deposits of 3,000- to 2,900-year-old lahars are just a meter above the level of 1993 fill. Upslope from Kakilingan, in Sitio Buag, it appears that Buag-age lahars filled or nearly filled some canyons but did not bury relatively high terraces. On one of these high terraces, a pre-Hispanic settlement (Dr. E. Dizon, Philippine National Museum, written commun., 1992) was buried lightly by ashfall from Buag eruptions but was not inundated by Buag lahars.

In the valley of the Bucao River, Poonbato was built on two or three prominent terraces, and downstream barangays were built on terraces overlooking the active channel of the Bucao. The main part of Poonbato and its new church, built on deposits of the Crow Valley eruptive period (6,000-5,000 yr old), were buried in 1992. A Buag deposit that formed the south bank of the Bucao River just downstream from Malumboy was buried in 1993. There is no reassurance from the geologic past for the safety of Botolan town; without dredging or other intervention, Botolan will probably be buried.

Why are some parts of Pinatubo's lahar fans relatively high and apparently safer than others? The main factor is probably the scale of the eruptions that supplied sediment for the lahars. The volume of erupted products has been generally declining in Pinatubo eruptions, from 35,000 yr ago to the present (Newhall and others, this volume). The 1991 eruption and the Buag eruptions were about one-half or one-third the size of Maraunot and Crow Valley eruptions, and about one-fifth or less of the size of the >35,000-year-old Inararo eruptions. In general, the highest terraces around Pinatubo formed after the large Maraunot and Crow Valley eruptions, when the level of sediment fill rose to those relatively high levels. However, not every terrace of those older periods can be considered safe: Poonbato is now buried and several of the highest terraces of the Santo Tomas plain are dangerously close to being overtopped. Also, we found no lahar deposit or sediment terraces of Inararo age, associated with Pinatubo's largest eruptions; some might have been completely eroded, but most are probably buried beneath younger sediment.

In summary, lahar hazard maps that have been published in the 3 years since the 1991 eruption are broadly consistent with the geologic record. Lahars in the balance of the present crisis will not cover all areas around Pinatubo that have previously been covered by lahars, but they will threaten some areas that have not yet been buried, especially in the Santo Tomas, Bucao, and Pasig-Potrero watersheds. The areas threatened will depend partly on which areas are made more, or less, vulnerable by stream captures and by human intervention. Engineered dikes and other construction works have undoubtedly decreased risk to some areas but have probably increased risks elsewhere.

Short-Range Warnings

Information about approaching typhoons, rainfall on the volcano slopes, or lahars passing upper observations points was translated into short-range warnings of lahars expected within minutes to days. Some people chose to keep living in hazardous areas because they didn't believe the warnings, had no reasonable alternative, or simply chose to rely on short-range warnings for their safety (Cola, this volume). Those who chose to rely on short-range warnings took calculated risks that warnings would be issued early enough for them to escape. In return, they were able to remain in their own homes and on their own land for as long as possible and clung to the hope that their places would not be overrun by lahars.

Lahar Season 1, June through November 1991

The first short-range warnings were issued during the urgent, last-minute preparations for a possible eruption (Punongbayan and others, this volume). Lahars had previously been discussed as a possible adjunct to that eruption, on June 13, 1991, when weather forecasters recognized that Typhoon Diding (international name, Yunya) was heading toward central Luzon, and volcanologists realized that the storm would probably generate lahars from the fresh deposits of eruptions then in progress. New warnings about lahar hazard were issued to civil defense, local officials, and military commanders on June 13 and 14 and appeared as headlines, together with the "mudflow" part of the May 23 hazard map, in the Manila Bulletin (fig. 6), the Philippine Daily Enquirer, and other Manila papers on June 14.

Figure 6. Headlines of the Manila Bulletin, June 14, 1991, warning of impending mudflows and showing the rivers along which they would likely pass.

The center of Typhoon Diding, downgraded to a tropical storm after it began crossing Luzon, passed within 100 km of Pinatubo's summit at about 1100 on June 15. Rain from Diding generated lahars from already-emplaced deposits and continued to generate lahars from fresh deposits of the climactic eruption that afternoon. Additional warnings of these lahars were issued on radio stations just before and throughout the climactic eruption, especially when reports arrived that bridges in Angeles City and in Barangay San Rafael, San Marcelino, had been destroyed by lahars. Some of these warnings originated from PHIVOLCS and the Office of Civil Defense; others were based on reports from citizens or reporters.

By the end of July 1991, the Regional Disaster Coordinating Council (RDCC Region III) deployed police and army units to ten primary (upslope) lahar watchpoints, one each on the O'Donnell, Bangut, Sacobia, Abacan, Pasig-Potrero, Porac, Gumain, Marella-Santo Tomas, Maloma, and Bucao Rivers. Initially, lahar watchers were housed in tents and made relatively crude "eyeball" estimates of the height of lahars passing their watchpoints. Information was relayed to central points at Camp Olivas, San Fernando, Pampanga, and the Zambales Provincial Disaster Coordinating Council in Iba, Zambales. Those centers raised simple alerts: Lahar Alert 1 (rain is falling on Pinatubo; get ready); Lahar Alert 2 (rain is continuing and a lahar could form; get set); and Lahar Alert 3 (a lahar has been confirmed; go to high ground).

Lahar watchers were drawn from the large personnel pools of the Philippine National Police (PNP) and Armed Forces of the Philippines (AFP) and served for about 2 weeks before being relieved by new watchers. Each group received a small amount of training before deployment, but frequent rotation limited the expertise that could be developed. During the first part of the 1991 lahar season, RDCC III in Camp Olivas tended to declare Alert 3 for all lahar-threatened areas whenever a lahar hit a single channel in Tarlac or Pampanga. This was before it became clear to everyone that the lahar-triggering rains in July 1991 were being delivered to the east side of Pinatubo by the trade winds, whereas the western slopes remained dry and lahar free (Rodolfo, 1991). An unfortunate consequence was that people in Zambales, roused too often by an Alert 3 without a corresponding lahar, became distrustful of the warnings and were largely ignoring them when the southwest monsoon finally arrived and brought lahars to Zambales. Lahar warnings on a channel-by-channel basis were started by late September 1991 after several meetings with RDCC and PDCC officials. However, people remained skeptical of warnings aired over the radio. Another difficulty with the PNP-AFP watchpoints was that many were located so far above the lahar channels that, in inclement weather, they would be shrouded by clouds and mist. Furthermore, those on hilltops were subject to lightning strikes; on one occasion, two watchers were killed by lightning drawn to their radio antenna. (For the 1992 rainy season, primary watchpoints along the Porac, Gumain, and Maloma Rivers were abandoned, and all-weather buildings were constructed at the other watchpoints.)

Secondary watchpoints were established at each major bridge or river crossing, all in populated areas. These were manned by police, generally without special training about lahars but equipped with two-way radios to learn of lahars coming their way and to relay information about those lahars as they passed each secondary watchpoint. Information flowed to and from Camp Olivas (for all of Pampanga, Tarlac, and Zambales) and the Zambales Provincial Disaster Coordinating Council at the capital in Iba (for Zambales).

PHIVOLCS and PLHT staff coordinated closely with weather forecasters at Clark Air Base and Cubi Point Naval Air Station, particularly as typhoons approached. Although the primary responsibility for typhoon warnings remained with the Philippine Atmospheric, Geophysical, and Astronomical Service Administration (PAGASA), PLHT and PHIVOLCS staff were able to give extra verbal and graphical warnings to communities threatened by lahars, on the basis of detailed weather information.

USGS, PHIVOLCS, and PLHT scientists also provided warnings of lakes impounded by pyroclastic flows or lahars that might fail catastrophically. One such warning was issued by PHIVOLCS to the RDCC on July 25, 1991, about a lake impounded in the upper Pasig-Potrero watershed. Regrettably, the warning on July 25 could not be specific and was the last official mention of this lake before it broke out on September 7, 1991, and killed several people downstream. More specific warnings were issued by PLHT about the largest such impoundment, where the Mapanuepe River was impounded by lahar deposits from the Marella River (fig. 7). On August 25, 1991, at 1700, PLHT warned PNP watchers at Dalanaoan of an impending lake breakout from Mapanuepe within the next few hours. The warning, however, was not disseminated to town officials until 2200 because the watchers' radio ran out of power. When the warning was finally disseminated, the municipal officials of San Marcelino and San Narciso were very reluctant to act on it because of previous false Alert 3 signals and because most of the people were already asleep. Fortunately, the San Marcelino police did awaken people of Barangay San Rafael before the breakout actually occurred at 0400 the next day and destroyed 8 houses (Umbal and others, 1991; Rodolfo and Umbal, 1992; Umbal and Rodolfo, this volume). The only casualty was a person who suffered a heart attack.

Figure 7. Mapanuepe Lake (left), impounded when lahars of the Marella River (lower right) blocked a tributary, the Mapanuepe River (left). Breakouts of this lake in 1991 prompted engineers to cut a channel through the ridge that lies south of (behind) the channel through which Mapanuepe Lake was still draining at the time of this photo. View is to the southwest, down the Santo Tomas River, which is the combined flow of the Marella and Mapanuepe Rivers. Much of the alluvial fan in the background (upper right) was buried by lahars in 1993.

In August 1991, after an emergency deployment of radio-telemetered rain gauges and acoustic lahar flow sensors (fig. 8) (LaHusen, 1994; Hadley and LaHusen, 1995; Marcial and others, this volume), a period of initial testing, and some initial difficulties with telephone and two-way radio links to civil defense officials, PHIVOLCS began to relay realtime instrument-based warnings to the Watch Point Center (WPC) of the RDCC at Camp Olivas, San Fernando, Pampanga. Instrumental records of flows were not yet calibrated by direct observations of those flows, nor did we know how much rain was needed to generate lahars, so warnings were limited to statements about rainfall and the occurrence of a strong signal on the instrumental flow monitors. Quiescence on the flow sensors was used to control frequent rumors of lahars. In general, WPC used PHIVOLCS' advice of lahar signals only to query the manned watchpoints; to our knowledge, WPC did not issue any lahar alert in 1991 solely on the basis of instrumental data from PHIVOLCS. One of the engineers who worked at WPC later told us that, because they didn't see the data themselves, they preferred to rely on reports from human observers.

An alternate lahar warning network, consisting of tripwires and rain gauges (Iwakiri, 1992), was installed by DPWH and operated by staff of the WPC. Tripwires and rain gauges were installed at two sites, Barangay Dolores (Mabalacat) and Barangay Sapangbato (Angeles City), and rain gauges were installed at those sites and at Barangay Dalanaoan (San Marcelino); Sitio Ugik (Botolan); Sitio Pasbul (Camias, Porac); and at the PNP Delta 5 manned watchpoint (Porac) (fig. 8), between February 1992 and April 1992. Unfortunately, the large size of these units (principally, of the 10- to 15-m-high concrete and steel mast) limited their installation to sites that were accessible by road--usually at lower elevations and far from the lahar source area. The tripwires were located so close to the towns at risk, and were broken so frequently by lahars and also by people, and were later buried so deeply by lahars, that they have been of little use. The rain gauges have been more useful, though correlation between rainfall and lahars is always questionable for rain gauges that are far from the lahar source area. Data are telemetered in realtime to RDCC-III in Camp Olivas and to the PDCC in Iba, Zambales, where they are interpreted by engineers and relayed by radio operators to manned watchpoints for confirmation. During the height of the 1992 monsoon season, however, the rain gauges at Dalanaoan and Ugik stopped transmitting data, owing to power problems that were not remedied until after the lahar season was over.

Figure 8. Sites of instruments and manned watchpoints for providing short-range warnings of lahar hazard around Pinatubo.

Lahar Season 2, May through November 1992

In 1992, to improve the technical accuracy of short-range warnings, we needed to answer the following questions:

What volumes of lahar discharge result from various amount of rainfall?

Can rainfall records and records from acoustic flow monitors (AFM's) serve as proxies, or even improvements, on direct observations from manned observation posts?

What levels of discharge are occurring, how and where is deposition occurring, and what channels are precariously full?

Is 1992 lahar activity changing relative to that of 1991, especially in any systematic way that can be of predictive value?

In June 1992, PHIVOLCS began systematic field monitoring of lahars in east-side channels, initially in the O'Donnell, Sacobia, Abacan, Pasig-Potrero, and Gumain channels and later reduced to the Sacobia and Pasig-Potrero. Staff reported peak discharges by radio, prepared simple hydrographs, recorded changes in the character of flows from start to finish, measured lahar temperatures, and took samples when practical (Arboleda and Martinez, this volume; Martinez and others, this volume). This activity was patterned after PLHT observations made on the Marella River in 1991; activity continued there and at the Bucao River in 1992 (Rodolfo and others, this volume; Umbal and Rodolfo, this volume).

Special effort was made to calibrate individual flow sensor records with observed discharge so that flow sensor data might be used as a substitute for manned observations. The first attempt at calibration, after lahar season no. 1, was for flows in the Marella watershed (Regalado and Tan, 1992). Similar work was done in the Sacobia watershed during lahar season no. 2 (Tuñgol and Regalado, this volume) and continued during lahar season no. 3 (Regalado and others, 1994). Though quantitative calibration of flow sensors in rivers other than the Sacobia is still ongoing, we can already use AFM records from all of these to distinguish the order of magnitude of discharge (for example, 10m³/s, 100 m³ /s, 1,000 m³/s, corresponding to small, medium, and large lahars). Qualitative information about the size of lahars is now included in PHIVOLCS' warnings to WPC/RDCC. In one early instrumental distinction of lahar size (August 1992), PHIVOLCS' advice of unusually strong lahar signals in the Sacobia was not used immediately by WPC/RDCC to order an evacuation, and 8 people downstream were killed by a major lahar. Since that time, somewhat greater credence has been given to instrumental indications of lahar size, but, even in the 1994 lahar season, people still find it much easier to visualize, and thus trust, a manned watchpoint's report of lahar depth (in feet) than either a quantitative or qualitative report of discharge. Few among those receiving warnings seem to be bothered that flow depth depends strongly on the channel width.

Records of flow sensors and observed discharge were also correlated with rainfall in the Sacobia watershed during 1992 and 1993. The results have allowed us to give an even earlier alert to WPC/RDCC, even before flow is detected. The thresholds for lahar generation, measured during the 1992 lahar season and described under the section "Defining the Threat," are our present (1994) basis for issuing warnings to WPC/RDCC. Because these thresholds will rise as sediment yield (normalized to rainfall) decreases, we will need to revise thresholds for warning in future years.

Also during the 1992 lahar season, we briefly explored the possibility of computer modeling of lahars to improve our understanding of flow processes and to help forecast traveltimes and downstream discharges. However, the simplest model available to us (DAMBRK, PC version) had no rigorous way of adding or subtracting sediment, and we knew from field observations that these were important processes. We could force the model to replicate some aspects of observed flows, but we concluded that, at Pinatubo, the model had no predictive value (R. Dinicola, USGS, oral commun., 1992).

A special concern arose in July 1992 when Pinatubo threatened to erupt anew. Mindful of past disasters at Kelut Volcano in Indonesia, where a crater lake was repeatedly ejected by eruptions until engineers lowered the lake level through a series of tunnels, we made rough estimates of the volume of water in the caldera lake and how much of that water might be converted to lahars were the lake to be ejected suddenly. Estimates of lake volume ranged from 5x10⁶ m³ to 8x10⁶ m³ (November 1991 air photographs), depending mainly on assumptions about the lake's average depth. However, as a lava dome grew, our concern decreased, because the lake appeared shallow, perhaps as shallow as a few meters. The eastern half of the lake gradually disappeared during this period, from alluvial filling aided in small measure by uplift of the lake floor. (Note, at the time of this writing, July 1994, the lake is several tens of meters deeper and has a substantially greater volume than it did in November 1991.)

Late 1992 and early 1993 was also a time to reestablish a battered lahar warning system. PHIVOLCS-USGS rain gauges and flow sensors required cleaning, repair, and replacement, and several needed to be moved to higher elevations in the lahar source regions to obtain more direct information about events in the source regions, longer lead times in warnings, and reduced risk of vandalism. This work was completed just before the 1993 lahar season.

Lahar Season 3, May through November 1993

The upgraded network of rain gauges and flow sensors operated successfully through the 1993 lahar season, with minimal periods of instrument failure. As expected, the threshold for lahar generation did rise slightly (Regalado and others, 1994), but not enough to change procedures for issuance of alerts. Alerts from PHIVOLCS to RDCC typically started with an alert that strong rain was falling on the upper slopes, followed by qualitative updates on the size of a resulting lahar ("small," "moderate," or "large," on the basis of the amplitude and duration of the AFM signal). Manned watchpoints provided similar information about the size of lahars, expressed as depth of the lahar (in feet). The latter reports were widely monitored by other watchpoints and local officials. Lahar watchers quickly realized that flows spread and peaks attenuate as they move downslope, so that by the time most flows reach downstream areas, reported "10-foot" lahars might be only a foot or two deep.

Some of the forward (higher elevation) manned watchpoints were discontinued in 1993, and more were discontinued in 1994, placing greater demands on PHIVOLCS instruments for initial warnings and on secondary (lower elevation) watchpoints for reports on arrival times and depths of lahars. As of this time of writing (July 1994), the formal lahar alert scheme from RDCC still retains the original 3 levels, with essentially the original meanings noted earlier. No formal distinction is made to indicate the size of an advancing lahar; rather, the depth of flow (in feet) and intensity of steaming are reported, with or without size information from PHIVOLCS, through a radio network of local observers at secondary watchpoints.

RESPONSE TO WARNINGS

The public's first and most effective response to lahar warnings was to move out of the way of the lahars. On June 14, 1991, PHIVOLCS' recommendation for general evacuation (principally for the eruption but also for lahars) was extended from 20 to 30 km in radius from Pinatubo's summit, and on June 15 it was extended again, to 40 km in radius (Punongbayan and others, this volume). During, and for weeks after, the climactic eruption, between 150,000 and 200,000 persons filled more than 300 evacuation camps (United Nations Disaster Relief Organization, UNDRO, unpub. data, 1991). Not everyone within 40 km of Pinatubo could or did evacuate; our impression, based only on anecdotal evidence, is that most of those who evacuated did so from fear of the eruption and associated earthquakes and lightning and not from the threat of lahars. But by so doing, some people also protected themselves from lahars.

After the eruption, evacuations were ordered every time Lahar Alert 3 (confirmed lahar--go to safe ground) was raised. Initially, evacuation recommendations were for all low-lying areas around the volcano; later, they were made on a river-by-river basis. Estimating the number of evacuees is complicated by having different barangays and families evacuating at different times, in some cases several times. The population of a reduced number of lahar-safe evacuation camps (~160) fluctuated seasonally, depending on rainfall. During periods of little rain and relatively low risk, the evacuee population dropped below 25,000 persons; during periods of heavy rain and strong lahars, the evacuee population swelled to more than 150,000 persons (UNDRO, unpub. data, 1992; Mercado and others, this volume). By the end of 1992, more than 50,000 persons had lost their homes entirely and another 150,000 had suffered some flooding or lahar damage (C.B. Bautista, this volume), and most of those people had been helped in evacuation centers.

As soon as the persistent character of the crisis became apparent, semipermanent and permanent resettlement sites were proposed as an alternative to evacuation camps. At about 10 previously undeveloped sites, the government built marketplaces, schools, health clinics, and buildings to attract and house manufacturing and other industry. Paved roads and utilities were, in many instances, of even higher standard than that in communities from which evacuees had come. Would-be settlers were offered low-interest loans to buy residential lots and to build houses. However, jobless, homeless families are reluctant to take on loans that they cannot repay, and for this and a variety of other reasons, displaced families have been slow to move into resettlement areas (C.B. Bautista, this volume). Virtually all of those who have moved from evacuation camps to more permanent resettlement areas have been replaced in the evacuation camps by families that are newly displaced by ongoing lahars.

Concurrently with operation of evacuation camps and development of resettlement areas, engineers from the Philippines, the United States, Switzerland, New Zealand, and Japan considered engineering measures to control the hazard itself. Politicians' reluctance to sacrifice any areas and engineers' training led them to try conventional channel maintenance measures--including construction of small "sabo dams" to trap sediment and stabilize base level, dredging of channels, and raising of dikes. The initial goal, more hopeful than realistic, was to contain the sediment within existing channels and other relatively small areas. Numerous small sabo dams were built but were promptly overrun (fig. 9). The dams were so hopelessly undersized that most scientists, and many engineers, recognized that they were little more than an employment program for local residents displaced by the eruption and visible evidence that the government was doing something rather than nothing.

Figure 9. Overtopped and breached sediment dam ("sabo dam"), in the Sacobia River just northwest of Clark Air Base. A, Aerial view, looking downstream. Small hut and person can be seen for scale. B, View from the right bank of the downstream face of the dam.

Not surprisingly, the worst problems of lahar incursions into populated areas were on unconfined alluvial fans of the Pasig-Potrero, Sacobia-Bamban, and Marella-Santo Tomas Rivers rather than in the deeper Bucao and O'Donnell valleys. Shallow channels on the alluvial fans filled quickly, and much overbank flow occurred. The dilemma for engineers quickly became apparent: the alluvial fans were not conducive to sediment storage, whereas the deeper valleys in which sediment could be trapped had relatively few people at risk.

When larger dikes were proposed to trap sediment on the fans of the Sacobia-Bamban, Pasig-Potrero, and Marella-Santo Tomas Rivers, we argued that the volumes of sediment would be so great that the only practical solution would be to let that sediment spread over relatively large "catch basins" outside current channels and in locations where sediment would naturally deposit. Most of the potential catch basins were in cultivation. In June 1992, PHIVOLCS and DPWH proposed catch basins to residents of Bacolor and Santa Rita, Pampanga (fig. 10), but the proposal was angrily rejected. No one was prepared to allow his own land to be buried.

Neither an engineer nor nature waits for debate, and, from 1991 to this time of writing, dikes on all sides of Pinatubo have been built, and all have been breached at their weakest points or points with the lowest freeboard during actual lahars. Failure at any point along a dike renders downstream portions of that dike useless (Rodolfo and others, 1993). Most failures of dikes to date have stemmed from the simple geometric difficulty of containing very large volumes of sediment within narrow confines, a faulty assumption that sediment would be deposited evenly along the full length of a diked channel (it will not), or an underestimate of the erosive power of lahars on dikes. After the 1993 lahar seasons, new, larger and stronger dikes were constructed along the outer perimeter of many areas that had been buried by lahar. In effect, catch basins were created, though they were not called by this name. These larger, widely spaced dikes are surely more realistic measures than their predecessors, but they will still be tested by remaining lahars, especially in their upper reaches where sediment fill is the fastest.

One successful engineering measure was excavation of a spillway to stabilize the level of Mapanuepe Lake, newly impounded behind lahar deposits of the Marella River, and thus to prevent overflows and potentially catastrophic breakout floods. Meetings between PLHT, DPWH engineers and consultants, Dizon Mines management and engineers, and local officials led to an engineering solution in which a permanent spillway was excavated into bedrock at the southern margin of the lake, roughly 300 m away from the present edge of the lahar dam. Excavation started on October 23, 1992, and the spillway first opened on November 20, 1992, fixing lake level at or below 121 m in altitude.

In July 1992, PHIVOLCS had preliminary discussions with DPWH and Gov. M. Cojuangco of Tarlac Province to discuss a creative, promising combination of sociopolitical and engineering measures. In that combination, the government would temporarily lease land to be used as catch basins, until the new sediment fill could be returned to agricultural production. Towns around the margins of the catch basin would be protected by dikes. The scheme recognized that it would be much easier to protect the town of Bamban if sediment were allowed to spread around it; otherwise, efforts to confine sediment within dikes of the Bamban River were likely to assure burial of Bamban. Money to lease the land and to provide other incentives for relocation would come in part from money intended for engineering works. Regrettably, decisions were made to keep building narrow, undersized sediment-containment dikes along the Bamban and other rivers. At the same time, with strong ties to family land, no strong incentives for relocation, and seemingly high and substantial new dikes in their backyards, most residents stayed in their villages, even in high-risk areas, and left only when or just before their communities were overrun by lahars. In 1992, lahars buried parts of the proposed leases in Tarlac Province before any formal leasing action was taken, and then the main path of lahars from the Sacobia shifted to the Pampanga side of the river, where government leasing of land had not been seriously considered. Thus, on both sides of the river, residents lost their life savings and livelihood and lost any opportunity that the government would lease or buy their land.

Debate continued in late 1992 about the most effective long-term measures for mitigating lahar risk. Our impression is that debate became more polarized, between those who advocated engineering structures and those who advocated that available funds be spent for resettlement and other social measures. In an ongoing saga of trial and error, each side could now point to serious problems with the other--for example, failed dikes and unpopular, half-vacant resettlement areas. In defense, each side could point to its own successes. In some instances, media coverage misrepresented positions, setting up "pro-dike" and "anti-dike" sides when, in fact, principal points were that (1) dikes might, in some instances, be viable engineering options, and (2) dikes ought to be built only if they would be large enough and strong enough to actually work as intended and not inadvertently introduce a false sense of security or an unduly long-term maintenance problem (for example, Rodolfo, 1992). These are different statements of the same conclusion, but some media groups saw them as different conclusions.

Many of the practical, philosophical, and emotional issues that arose bore a marked similarity to issues raised by dike construction that keeps the lower Mississippi River from being captured by the Atchafalaya River (McPhee, 1989). In general, dike construction made more sense to those at risk than to disinterested observers.

In August and October of 1993, as in 1992, policymakers of the Mount Pinatubo Commission were forced by serious lahars to reexamine their long-range mitigation plans. Dredging and dike construction during the 1992-93 dry season was no match for the volume of 1993 lahars, so channels on the alluvial fans soon became choked with sediment, some of the dikes were overtopped, and several towns were badly damaged. Early lahars of the 1994 rainy season have already overtopped some dikes that were constructed or raised during the 1993-94 dry season, and greatly reduced freeboard at others. In a long-term view, undersized dikes have had little effect on where lahar sediment has been deposited, and most might just as well have not been built. They have postponed inevitable burial for some communities, and they might ultimately save a few communities right at the edge of where sediment would have naturally spread. But they have been built at great cost--roughly half of the monies available for coping with this disaster, plus loss of opportunities to use the same monies for nonstructural or even better structural measures, and losses attributable to a false sense of security until lahars were at hand.

Among the questions raised anew were:

What engineering solutions can be technically viable? In order to be viable, what design specifications would they need?

If an engineering solution saves a town for a year, or so, is it worth the associated costs? The direct cost of the dike is simple to calculate; indirect costs, and the value of saving a town for a year, are more subjective.

Are some engineering solutions actually aggravating lahar problems, for example, by initially helping lahars to reach far downstream, and now by causing avulsions to move upstream and thus damage towns on the edges of alluvial fans?

Politically, economically, and scientifically, what are the most practical options for catch basins, that is, land that can be used for sediment storage until the lahar crisis has passed?

We are neither engineers nor policymakers so we will not presume to answer these latter questions. Rather, we mention them to illustrate the continuing dilemma faced by all who seek to mitigate the lahar hazards of Mount Pinatubo: how to help people stay in their communities wherever possible and yet not waste resources where communities cannot be defended against lahar without spending an inordinate amount of money and (or) running a high risk that lahar defenses will ultimately fail. Scale and topography are major parts of the dilemma: solutions that might have worked where lahars are smaller, governments are wealthier, and valleys are deeper, are not readily transferable to Pinatubo. Denial of the hazard and difficulties in reaching political consensus are also important factors. Rather than asking that sediment be kept from all populated and cultivated areas, citizens and their leaders would do better to decide which areas must be protected and which can be temporarily sacrificed and to ensure that both hazard and compensation are fairly distributed.

SUGGESTIONS FOR FUTURE WARNINGS

Can lahar warnings at Pinatubo be improved? Yes, without question. Some improvements relate to incomplete data and uncertainties about process, and others relate to when and how warnings are conveyed. For technical improvements of long-range forecasts, we need to keep tracking:

• Sediment budgets, including how much sediment that predates 1991 is being incorporated into current lahars, how much sediment is flushing straight through the river systems, and, especially, how rapidly the annual sediment yields are declining.
• Details of topographic change in lowlands, with a contour interval of 5 m or less. Excellent maps have been made for limited areas, for a few flight dates, but to our knowledge there are no plans to make repeated, high-precision photogrammetric surveys of topographic change around the entire volcano.
• Reaches of lahar deposition, because those reaches will be the sites of the next channel avulsions and overbank lahar damage. This effort requires frequent remeasurements of cross sections of each major river draining Pinatubo.

Increasingly, GIS technology can be used to analyze, display, and disseminate current information about watershed response and risk. All long-range planning for lahar mitigation must anticipate and be adaptable to continued, rapid, major changes in the hydrologic system, such as stream captures. For example, captures of Abacan drainage by the Sacobia, and part of the Sacobia drainage by the Pasig-Potrero, are unlikely to be the last such events, and GIS-based hazard maps can be changed accordingly.

For technical improvements in short-range forecasts, we still need:

• Timely transmission of typhoon tracks and detailed evaluations of southwest monsoon air masses, from PAGASA to scientists and RDCC-III personnel at Pinatubo.
• Continued operation of rain gauges and flow sensors on the upper and middle slopes of Pinatubo, with commensurate technical support for those instruments and calibration of flow sensors to actual discharges.
• Better realtime communication with WPC/RDCC, perhaps aided by a duplicate data-receiving station at WPC to serve as a backup and to counteract skepticism that we encountered in 1991-1992. A more ambitious extension of this would be to link all lahar watchers--police, military, scientists, and others--by a robust radio communications network, supplemented where possible by voice telephone and facsimile machines.

Equally or even more important than technical refinements in forecasts, we need to strengthen the channels by which our scientific results are actually used in lahar mitigation measures. Four specific improvements in our presentation would be:

• Additional, highly visual presentations of hazard, such as videotapes and working models that are prepared in cooperation with those who are experienced in shaping public opinion and public policy.
• Warnings that speak clearly and firmly despite inescapable uncertainties in the outcome of such complex events as lahars. A graphic illustration of this need came during raging lahars in early September 1992, when several villages were being overrun by lahars and when the chief of RDCC Region III (Gen. Pantaleon Dumlao) said to one of us, "Yes, you warned us, but you didn't shout loudly enough." Scientists are trained to be conservative in their warnings, but there are times when we must make our best guesses loudly and clearly, even without a full set of data.
• Better use of scientific consensus in public statements. Differences of opinion among scientists are inevitable and have occurred. Some have concerned purely scientific matters; others have concerned recommendations for mitigation. Unfortunately, these differences have been played out in the newspapers rather than in private discussions among scientists, and we have therefore lost some credibility, as a group, in the eyes of decisionmakers. Development of written scientific consensus before public statements would help all parties.
• Equal attention to decreasing lahar hazard as to increasing hazard, partly to avoid undue alarm, partly to retain our credibility for those times when urgent warnings are needed, and partly to help people return to their land where such return is reasonably safe.

FUTURE LAHAR HAZARD

The nature of Pinatubo lahar threat was discussed in earlier sections. Here, we call special attention to three anticipated changes in hazard: extreme rainfall; diminution of annual sediment yield; and continued, even increasing, distal sedimentation.

Extreme Rainfall

Rainfall in the Pinatubo area is highly variable from one year and one month to the next, depending on typhoons and on sustained "siyam-siyam" southwest monsoon rainfall. Rainfall during 1991 and 1992 was close to the long-term mean (table 2). In contrast, that at Cubi Point in 1993 (7,165 mm, with 3,102 mm in August alone) was far above the mean and set new records for that station. Rainfall in excess of 2,500 mm/yr in Manila and Clark Air Base occurs, on average, about once every 10 yr, most recently in 1972 (>3,000 mm in the July-August period) and 1986 (>2,000 mm in August alone); none of our east-side stations approached these high levels yet, but, clearly, such levels can be reached.

Daily rainfall can also be heavier than has occurred since the eruption. We do not know the maximum daily rainfall during 1991 to 1993 at Clark Air Base or that at Cubi Point during 1993; during 1991 and 1992, the 24-h maximum at Cubi Point was 280 mm (J.S. Oswalt, oral commun., 1991). We have recorded higher amounts at many of our stations on Pinatubo itself (highest=744 mm in 24 h, from 0600 July 25 to 0600 July 26, 1994 at QAD), but we do not have a long-term average against which that could be compared.

Thus, in terms of the heavy rainfall that can generate severe lahars, rainfall in 1991 to 1992 was well below historical maxima, and within 1 standard deviation of average values. 1993 rainfall at Cubi Point was an historical maximum, while that at other west-side stations was probably normal. We expect that the next decade will include at least one additional year with heavy rainfall, >2,500 mm/yr on the east side and >5,000 to 6,000 mm/yr on the west side.

Diminution of Annual Sediment Yield

Lahars and rapid sediment accumulation will diminish and eventually stop. In terms of sediment volume, we are at the time of this writing (July 1994) about two-thirds through the crisis (fig. 3; table 3); we are farther along in some east-side watersheds and not as far along on the west. Both the geographic scope and the timescale of further mitigation options, including resettlement and engineering measures, should recognize that sediment yields are diminishing rapidly and will soon become more manageable, perhaps without need to undertake massive measures.

Continued Distal Sedimentation

As vegetation recovers and as other aspects of the 1991 watershed disturbance heal, runoff, like sediment yield, will diminish. However, high rates of runoff may linger a few years longer than high rates of sediment yield. If so, two phenomena will result. First, bank erosion in downstream channels will be high, so maintenance of remaining dikes will remain a major task. Second, some of that erosion will be of lahar deposits themselves, and the remobilized sediment will be washed farther downstream. One can think of Pinatubo sediment as moving into distal lowlands in two steps--first as lahars that are deposited on alluvial fans around Pinatubo and second as gradual but persistent movement of that sediment into more distant channels, for years or even decades after lahars have ended. Low-gradient river channels, even far from Pinatubo, will experience gradual siltation and diminution of channel capacity. As that capacity diminishes, normal rains will produce overbank flooding. Dredging will be required as a long-term flood-control measure in those areas, and the need for distal dredging may soon outstrip the need for dikes and other structures nearer to Mount Pinatubo.

CONCLUSIONS

Erosion and lahars in posteruption Pinatubo watersheds have filled lowland channels at rates that are unprecedented in the history of volcano hazards mitigation or sediment control. Almost all of the sediment transport occurs as lahars, which are episodic events dependent on heavy rain. Sediment yields at Pinatubo, on the order of 10⁶ m³/km²/yr during 1991-93, were an order of magnitude larger than at Sakurajima and at Mount St. Helens immediately after eruptions.

Rates of change in the lahar hazard are as remarkable as the initial magnitude of that hazard. Sediment yields are declining, on average by 20-30 percent per year and in some drainages by half or more each year. Sediment yields have decreased most sharply in east-side watersheds, where volumes of 1991 pyroclastic-flow deposits are small to moderate, slopes are relatively steep, and vegetation recovery is relatively rapid. In the Sacobia-Pasig watershed, three-fourths of the expected sediment has already arrived in the lowlands. Even on the west side, more than half of the expected sediment has arrived in the lowlands. Rates of sediment yield in 1993, normalized for watershed area and rainfall, ranged from one-third of 1991 values (in the Sacobia-Pasig system) to 100 percent (or more?) of 1991 values (in the Balin Baquero-Bucao) (table 4).

Despite long-range decline of sediment yield, days or weeks of extreme rainfall can still generate bank-full or overbank flow of several thousand cubic meters per second of dense lahar slurry. As channels are filled to capacity, each new storm and rainy season, and especially extreme rainfall, will cause new avulsions and shift flows wildly from one side of an alluvial fan to the other. One day or one season a community might be relatively safe; the next, it might be buried in lahar deposits. Then, as quickly as it was buried, that area might be isolated from future hazard.

Scenarios of lahar hazard offered by PHIVOLCS and PLHT/ZLSMG--long-range and immediate--have been dire yet generally accurate. Undersecretary of Public Works and Highways T. Encarnacion said to us, in 1992, "You were right . . .  unfortunately." Some of our warnings have been based on hard data, others on general experience and intuition. We have walked a fine line between issuing timely warnings (some were backed only by sparse data) and reliable warnings (backed by ample data). Most citizens and most public officials have responded constructively, being willing to make precautionary evacuations, even after a few false alarms. Mutual understanding, courage to risk being wrong, and tolerance in case of false alarms are essential.

Hazard maps, booklets, and videos have correctly identified the most hazardous areas; immediate warnings from rain gauges, flow sensors, and manned watchpoints alerted people who remained in these communities to move to high, safe ground at particularly critical times. Of the three sources of immediate warnings, rain gauges and manned watchpoints were immediately useful, while flow sensors are gradually gaining acceptance. Civil defense has tried with reasonable success to fine-tune evacuations so that they are neither too large nor too small. After serious problems with false alarms in 1991, warnings are now reliable enough that most people are choosing to evacuate when warned, albeit at the last minute.

Of the more than 50,000 people who lost their homes to lahars, an unknown but substantial percent would have been killed had there been no warnings. Actual deaths from lahars, in contrast, are approximately 100--many times fewer than might have been killed without warnings. Jeepneys (local public minivans) and dump trucks full of property--furniture, household goods, farm animals, and harvested crops--accompany the exodus of evacuees, so we judge that much portable property has also been saved, though poignant scenes of people digging through fresh mudflow deposits to salvage belongings indicate that much has also been lost.

Hazard maps and estimates of long-term sediment yield have been important input to some decisions on long-term lahar mitigation. Major relocation sites have been sited (or, in the case of Rabanes, canceled) on the basis of such maps. Year-to-year emergency plans have been made on the basis of such maps. But only some decisions about engineering control of lahars have utilized available scientific information. Many small dams and small and large dikes have been built in the face of scientific judgment that they are too small or in places where any structure will be overrun. Information about expectable volumes of sediment is only one and sometimes a relatively minor factor in mitigation decisions, in competition with differences in perception of both risk and solutions (table 5), peoples' reluctance to leave and "sacrifice" their land, a lack of attractive alternatives for those at risk, political competition, and cost of mitigation measures.

Table 5. Differences in perception and approach between various parties in Pinatubo lahar mitigation.

• approach natural hazards as forces to be understood and, when necessary, avoided, rather than as forces to be controlled.
• think over a wide range of spatial and time scales, from cubic centimeters to cubic kilometers and from minutes to millennia.
• think of natural hazards as complex, ongoing, ever-changing processes that need to be studied continuously. Frequent changes in facts and interpretations seem perfectly natural, indeed, good.
• must often be the bearers of bad news and must therefore be prepared for hostile receptions.

Engineers:

• are trained to control or otherwise design against natural hazards (floods, landslides, and others). To not do so is professionally awkward, just as it is for a doctor to "give up" on a patient.
• think principally on the scale of their structures and the design life of those structures.
• are trained to design with available or quickly obtainable data and to add a factor of safety for future change. Opportunities to change design after construction are limited.

Residents:

• view natural hazards as God's will, largely beyond their control but possibly preventable by prayer and engineering works.
• must deal with harsh reality of finding food and shelter, which before a lahar is certainly easiest on their own land.
• have emotional as well as economic attachment to the places in which they were born and raised.

Politicians and other policymakers:

• view natural hazards as political and policy problems.
• need to make decisions on technical issues even when technical advice is conflicting or uncertain, and even when the natural situation is sure to change.
• are caught in a quandary between being sympathetic to public pressure for quick fixes, yet knowing that resources for coping with the lahar hazard are limited and must be used to everyone's advantage.
• are sensitive to issues of reelection and duration of their own responsibility.

ACKNOWLEDGMENTS

We are pleased to acknowledge lively discussions with those to whom we provide warnings. They have constructively critiqued and used our warnings, sometimes even before we have decided exactly what the warning should be. We also thank those who have supported our work, financially and logistically, including His Excellency Fidel V. Ramos (as former Chairman of the National Disaster Coordinating Council and currently as President of the Republic of the Philippines), Secretary Renato de Villa (Chairman, National Disaster Coordinating Committee), Secretary Corazon Alma de Leon (Chairperson of the Mount Pinatubo Commission), Secretary Jose P. de Jesus and Secretary Gregorio Vigilar (Department of Public Works and Highways), Secretary Ricardo T. Gloria (Department of Science and Technology), Congressmen Rolando Andaya and Emigdio Lingad (Chairman and Vice-Chairman, respectively, of the House Committee on Appropriations), Gen. Pantaleon Dumlao (Chairman of the Regional Disaster Coordinating Committee, RDCC III), Mayor Richard J. Gordon (now Chairman of the Subic Bay Metropolitan Authority) and Congresswoman Katherine H. Gordon (the First District of Zambales), Mayor Sally Deloso of Botolan, Director Joel Muyco of the Mines and Geosciences Bureau, Dr. Eduardo Sacris of Dizon Mines (Benguet Corporation), faculty of the University of the Philippines' National Institute of Geological Sciences, other friends of the PLHT/ZLSMG, USAID/Philippines, USAID/Office of Foreign Disaster Assistance, the Philippine Air Force, and U.S. military forces in the Philippines. Yvonne Miller of the University of Geneva and Mike Dolan of Michigan Technological University joined our efforts in 1992 and 1993 as student volunteers; Sheila Agosa of the University of Washington joined as a student volunteer in 1994. The work of PLHT in 1991 and ZLSMG in 1992-93 was supported by grants from the Earth Sciences and International Programs Divisions of the National Science Foundation. Numerous private individuals in the United States contributed to an emergency fund to support PLHT in 1991.

We acknowledge the seminal contributions of our late colleague, Dick Janda, and honor his memory with senior authorship. While most of us were still preoccupied with day to day events, Dick was already thinking in terms of watershed processes and how an understanding of those processes would lead to better forecasts.

REFERENCES CITED

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

Bautista, C.B., this volume, The Mount Pinatubo disaster and the people of Central Luzon.

Bureau of Soils and Water Management, 1991, Mudflow and siltation risk map: Bureau of Soils and Water Management, Quezon City, 1 sheet, scale 1:250,000.

Cola, R.M., this volume, Responses of Pampanga households to lahar warnings: Lessons from two villages in the Pasig-Potrero River watershed.

Hadey, K.C., and LaHusen, R.G., 1995. Technical manual for an experimental acoustic flow monitor: U.S. Geological Survey Open-File Report 95-114, 24 p.

Iwakiri, T., 1992, Photo album on lahar warning systems around Mount Pinatubo: unpublished report, Japan International Cooperation Agency, 27 p.

Janda, R.J., Meyer, D.F., and Childers, D., 1984, Sedimentation and geomorphic changes during and following the 1980-1983 eruptions of Mount St. Helens, Washington (parts 1 and 2): Shin Sabo, v. 37, no. 2, p. 10-21 and v. 37, no. 3, p. 5-19.

LaHusen, R.G., 1994, Real-time monitoring of lahar using ground vibrations [abs.]: Abstracts with Programs, Geological Society of America, Annual Meeting, Seattle, p. A-377.)

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.

Martinez, M.L., Arboleda, R.A., Delos Reyes, P.J., Gabinete, E., and Dolan, M.T., this volume, Observations of 1992 lahars along the Sacobia-Bamban River system.

McPhee, John, 1989, The control of nature: New York, Farrar Straus Giroux, 272 p.

Mercado, R.A., Lacsamana, J.B.T., and Pineda, G.L., this volume, Socioeconomic impacts of the Mount Pinatubo. eruption.

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

Philippine Institute of Volcanology and Seismology, 1992a, Lahar hazards of Mount Pinatubo: unpublished map, 1 sheet, 1:50,000.

------1992b, Ang Lahar: PHIVOLCS, Quezon City, 19 p. + maps.

------1994, Lahar studies: Quezon City, PHIVOLCS Press, 80 p. [Report of studies sponsored by UNESCO.]

Philippine Institute of Volcanology and Seismology and the National Economic Development Authority, 1992a, Pinatubo lahar hazards map: National Economic Development Authority, Manila, 1 sheet at scale 1:200,000; 4 sheets, Quadrants 1-4, at scale 1:100,000.

------1992b, Pinatubo Volcano Lahar Hazards Map: National Economic Development Authority, Manila, 4 sheets at scale 1:200,000.

Pierson, T.C., and Costa, J.E., 1994, Trends in sediment yield from volcanoes following explosive eruptions [abs.]: Abstracts with Programs, Geological Society of America, Annual Meeting, Seattle, p. A-377.

Pierson, T.C., Daag, A.S., Delos Reyes, P.J., Regalado, M.T.M., Solidum, R.U., and Tubianosa, B.S., this volume, Flow and deposition of posteruption hot lahars on the east side of Mount Pinatubo, July-October 1991.

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

Pinatubo Lahar Hazards Taskforce (PLHT) (Umbal, J.V., and Rodolfo, K.S., compilers), 1991a, Lahar hazard map of the Santo Tomas river system, Pinatubo Volcano, Luzon, Philippines (as of August 23, 1991): unpublished map, 1 sheet, scale 1:36,400.

Pinatubo Lahar Hazards Taskforce (PLHT) (Paladio, M.L., Umbal, J.V., and Rodolfo, K.S., compilers), 1991b, Lahar hazard map of the Bucao river system, Pinatubo Volcano, Luzon, Philippines (as of September 5, 1991): unpublished map, 1 sheet, scale 1:36,400.

Punongbayan, R.S., Besana, G.M., Daligdig, J.A., Torres, R.C., Daag, A.S., and Rimando, R.E., 1991a, Mudflow hazard map: PHIVOLCS-NAMRIA, 1 sheet, scale 1:100,000 (released July 31, 1991).

------1991b, Mudflow hazard map: PHIVOLCS-NAMRIA, 1 sheet, scale 1:100,000 (released October 15, 1991).

Punongbayan, R.S., Newhall, C.G., Bautista, M.L.P., Garcia, D., Harlow, D.H., Hoblitt, R.P., Sabit, J.P., and Solidum, R.U., this volume, Eruption hazard assessments and warnings.

Punongbayan, R.S., Tungol, N.M., Arboleda, R.A., Delos Reyes, P.J., Isada, M., Martinez, M., Melosantos, M.L.P., Puertollano, J., Regalado, T.M., Solidum, R.U., Tubianosa, B.S., Umbal, J.V., Alonso, R.A., and Remotigue, C., 1994, Impacts of the 1993 lahars, long-term lahars hazards and risks around Pinatubo Volcano, in Philippine Institute of Volcanology and Seismology, Lahar Studies: Quezon City, PHIVOLCS Press, p. 1-40.

Punongbayan, R.S., Umbal, J., Torres, R., Daag, A.S., Solidum, R., Delos Reyes, P., Rodolfo, K.S., and Newhall, C.G., 1992a, A technical primer on Pinatubo lahars: PHIVOLCS, Quezon City, 21 p.

------1992b, Three scenarios for 1992 lahars of Pinatubo Volcano: unpublished report, PHIVOLCS, 15 p.

Regalado, M.T., Delos Reyes, P.J., Solidum, R.U., and Tuñgol, N.M., 1994, Triggering rainfall for lahars along the Sacobia-Bamban River in 1993, in Philippine Institute of Volcanology and Seismology, Lahar Studies: Quezon City, PHIVOLCS Press, p. 41-45.

Regalado, M.T., and Tan, R., 1992, Rain-lahar generation at Marella River, Mount Pinatubo: unpublished report, University of the Philippines, 6 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 eruptions [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 62.

------1992, The continuing Pinatubo lahar threat: unpublished position paper, October 28, 1992, 7 p. Excerpts published in "At Large," Philippine Daily Inquirer, November 11-12, 1992.

Rodolfo, K.S., and Umbal, J.V., 1992, Catastrophic lahars on the western flanks of Mount Pinatubo, Philippines: Proceedings, Workshop on Effects of Global Climate Change on Hydrology and Water Resources at the Catchment Scale, February 3-6, 1992, Tsukuba, Japan, p. 493-510.

Rodolfo, K.S., Umbal, J.V., and Alonso, R.A., 1993, A rational analysis of the Santo Tomas-Pamatawan Plain and lahar system of southwestern Mount Pinatubo, Zambales Province, Philippines: unpublished report, Zambales Lahar Scientific Monitoring Group, 16 p.

Rodolfo, K.S., Umbal, J.V., Alonso, R.A., Remotigue, C.T., Paladio-Melosantos, 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., Janda, R.J., de la Cruz, E., Gabinete, E., Eto, I., Isada, M., Sexon, M., and Hadley, K.C., 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., 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.

Tuñgol, N.M., and Regalado, M.T.M., this volume, Rainfall, acoustic flow monitor records, and observed lahars of the Sacobia River in 1992.

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

Umbal, J.V., Rodolfo, K.S., Alonso, R.A., Paladio, M.L., Tamayo, R., Angeles, M.B., Tan, R., and Jalique, V.,1991, Lahars remobilized by breaching of a lahar-dammed non-volcanic tributary, Mount Pinatubo, Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 63.

U.S. Army Corps of Engineers, 1994, Mount Pinatubo Recovery Action Plan: Long Term Report: Portland, Oregon, 162 p. + 5 appendices.

U.S. Geological Survey-Philippine Institute of Volcanology and Seismology (USGS-PHIVOLCS), 1994, Serious but rapidly declining hazards at Mount Pinatubo: unpublished progress report of PHIVOLCS-USGS collaboration, 1992-mid-1994, Quezon City, 59 p + 14 attachments.

Voight, B., 1988, Countdown to catastrophe: Earth and Mineral Sciences (Pennsylvania State University), v. 57, p. 17-30.

------1990, The 1985 Nevado del Ruiz volcano catastrophe: Anatomy and retrospection: Journal of Volcanology and Geothermal Research, v. 44, p. 349-386.

Zambales Lahar Scientific Monitoring Group (ZLSMG) (Alonso, R.A., Rodolfo, K.S., and Remotigue, C.T., compilers), 1993, Lahar hazard map of the Santo Tomas-Pamatawan plain, southern Zambales province, as of September 1, 1993: unpublished map, 1 sheet, scale 1:36,400.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey
This page is <https://pubs.usgs.gov/pinatubo/janda/>
Contact: Chris Newhall
Last updated 06.11.99