FIRE and MUD Contents

Instrumental Lahar Monitoring at Mount Pinatubo

By Sergio Marcial,¹ Arnaldo A. Melosantos,¹ Kevin C. Hadley,² Richard G. LaHusen,² and Jeffrey N. Marso²

¹Philippine Institute of Volcanology and Seismology.

²U.S. Geological Survey.

ABSTRACT

Rain gauges and experimental acoustic flow monitors (AFM's) were installed near drainages of Mount Pinatubo after the eruption of June 15, 1991. The AFM's use exploration-class geophones to detect ground vibration caused by passing flows, principally debris flows and hyperconcentrated streamflows. Data from the rain gauges and AFM's are telemetered to the Pinatubo Volcano Observatory, where they are used by the Philippine Institute of Volcanology and Seismology for realtime detection and warning of lahars. Although some lahars were missed during the period of initial installation in 1991 and during periods of sensor or telemetry malfunction, every lahar that passed an operating sensor was detected, and civil defense was warned. Civil defense, in turn, used these alerts to query manned watchpoints and to raise public lahar-alert levels.

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

INTRODUCTION

The explosive eruption of Mount Pinatubo on June 15, 1991, deposited 5-6 km³ of loose, unconsolidated, pyroclastic-flow and tephra-fall debris on the slopes of the volcano (W.E. Scott and others, this volume). This debris, when eroded by rainwater, flows downslope as lahars, which erode river banks on the upper slopes and bury residential and agricultural lands in the adjoining lowlands. By the end of the 1992 rainy season, lahar deposits had covered an area of at least 360 km² (Mercado and others, this volume) and affected densely populated municipalities around Mount Pinatubo.

Because many people are reluctant to leave threatened, lowland communities until lahars are imminent, the Philippine Institute of Volcanology and Seismology (PHIVOLCS) and others set up warning systems to provide as much as several hours of time for them to move to safer ground. Those warning systems have included rain gauges, conventional 1-Hz seismometers that are also used for monitoring volcanic and tectonic seismicity, conventional 10-to 300-Hz exploration seismometers and associated acoustic flow monitor (AFM) circuitry, trip wires, and manned watchpoints.

This paper describes a network of high-frequency AFM's and automated rain gauges that were brought to Pinatubo by the Volcano Crisis Assistance Team (VCAT) of the U.S. Geological Survey and that are now operated and maintained by PHIVOLCS. Data are received at PHIVOLCS' Pinatubo Volcano Observatory, which passes lahar warnings to the Watch Point Center (WPC) of the Regional Disaster Coordinating Council (RDCC), at Camp Olivas in San Fernando, Pampanga.

PREVIOUS INSTRUMENTAL MONITORING OF LAHARS

In the Philippines, previous instrumental monitoring of lahars relied on conventional 1-Hz seismometers, also used for volcano monitoring (Bautista and others, 1986). Relatively high-frequency tremor at Mayon Volcano was correlated visually with lahars in progress. Signals were strongest when lahars passed along gullies near the seismometers, and peak signals corresponded to peak discharge. Successful warnings were given to concerned authorities and to the public.

However, because 1-Hz seismometers record signals between 1 and 10 Hz that can travel through a whole volcano, it was difficult to pinpoint the exact channel along which a lahar was flowing. It was also difficult, at times, to distinguish lahar signals from other sources of tremor (for example, pyroclastic flows and jetting of steam from summit fumaroles).

Trip wires, though not previously used in the Philippines, have been used elsewhere with mixed success. In pairs or series, they can detect the approximate magnitude of a flow (according to the level of the wire that has been broken) and track flows past upstream and then downstream sets of wires. Such trip wires are also subject to vandalism or accidental breakage, cannot provide information about the flow after they have been broken, and must be restrung each time they are broken.

1991 INSTRUMENTAL MONITORING NETWORK

Three weeks after the paroxysmal June 15, 1991, eruption, the Volcano Crisis Assistance Team (VCAT) of the U.S. Geological Survey (USGS) brought a newly developed set of AFM's capable of recording acoustic vibrations in nearby rocks and soil as lahars flow along channels (Hadley and LaHusen, 1991; in press). The flow sensors were supplemented by automated rain gauges, so early warnings could be made of possible flows and rainfall and lahars could be correlated (Bautista and others, 1991; Tuñgol and Regalado, this volume). All components were designed to be relatively inexpensive, quick to install, and durable. Each field station is powered by a 12-V battery charged by a solar panel. Data from both the AFM's and rain gauges are telemetered to the Pinatubo Volcano Observatory (PVO, on Clark Air Base, and also in PHIVOLCS' Main Office, Quezon City), where they are received by a computer and made available in realtime in both graphical and numerical formats.

The combined network of flow sensors and rain gauges serves three roles: to provide immediate warnings of lahar hazards, to collect data for studies of the hydrologic aftermath of the eruption, and to test technical aspects of the system itself.

The rain gauges were installed high in the lahar source areas, free of nearby obstructions. In 1991, six rain gauges and seven flow monitors were deployed at strategic places around the volcano (fig. 1; table 1). Sites for rain gauges were selected on relatively high ground in the lahar source regions, with helicopter access and line-of-sight telemetry to PVO or to a repeater site.

Figure 1. Location of rain gauges and flow monitors in 1991-92 (solid symbols) and 1993 (open symbols). For 1991-92, RG (rain gauge) A, Mount Cuadrado; RG B, Marella, at the BUGZ seismic station; RG C, at PI2 seismic station; RG D, Mount Culianan; RG E, Gumain; RG F, Sacobia. FM (flow monitor) 1, Lower Sacobia; FM 2, Marella-Santo Tomas; FM 3, Abacan Gap; FM 4, Upper Sacobia; FM 5, Gumain; FM 6, Pasig-Potrero; FM 7, Balin Baquero. New stations for 1993 were RG D1, Kamanggi; RG E1, Summit rim, east-southeast of CRAZ seismic station; RG G1, O'Donnell-Upper Bucao; FM 31, Lopoton Creek; FM 41, Upper Bucao; FM 51, O'Donnell; and FM 71, Maraunot.

Table 1. Rain gauge and acoustic flow sensor stations at Mount Pinatubo.

¹Each acoustic flow sensor (1-7, middle digit here) can be queried on three channels, for example, 010, 011, and 012, representing the full band, low band, and high band of data.

Flow monitor sites were selected along the middle reaches of major lahar-bearing rivers, below most of the lahar source region yet far enough above population sites as to maximize lead time in warnings. Sites are on canyon walls or hilltops, safely above flows but close enough to record high-frequency ground vibration (typically, <100 m and, sometimes, <20 m from the stream). Telemetry is generally repeated through rain gauge sites. Flow sensors were installed along the Marella River, on the Balin Baquero River just downstream from its confluence with the Maraunot River, in the upper and lower reaches of the Sacobia River, along the Abacan River just downstream from its divergence from the Sacobia River, and along the Pasig-Potrero and Gumain Rivers (fig. 1). In 1992, sensors were in the same places but, owing to difficulties of access and maintenance, some malfunctioned, especially late in the rainy season. Details are given under "Installation and Maintenance of Field Sites." In 1993, several sensors were moved to higher and safer locations (table 1).

MONITORING INSTRUMENTS

RAIN GAUGES

The rain gauge counts tips of a magnetic-switch tipping bucket (fig. 2A); field calibration showed that each tip of gauges that were in the field in 1991-92 represented rainfall of 1±0.1 mm. Because ash from mild eruptions and secondary explosions affects operation of the buckets and leads to erroneous readings, one of us (R.G. LaHusen) designed an Ash-Resistant Rain Gauge (ARRG) in which rain falls through a 12-cm-diameter orifice and collects in a vertical PVC pipe, closed at the bottom. A spillover outlet about halfway up the column (fig. 2A) leads to the tipping bucket mechanism in a separate, ash-protected enclosure. The PVC pipe is filled with water to spilling level, and subsequent rain falls into the water column to cause an equivalent overflow to the tipping bucket. Any ash entering the ARRG settles to the bottom of the PVC tube and safely away from the tipping mechanism. Because the volume of ash is negligible in comparison to the volume of rainwater, it does not seriously inflate the rain measurements.

Figure 2. Setup of (A) a rain gauge station and (B) an acoustic flow monitor station.

The rain gauges use a microprocessor-controlled digital telemetry platform. Each time the bucket tips, the magnetic switch senses the tip and sends a signal to a digital counter. Data are transmitted every 30 min or every 10 tips, whichever comes first. In practice, data are transmitted every 30 min during periods of light or no rain and every 10 tips during heavy rainfall. Occasionally, when heavy rains persist for several days, frequent transmission, clouds, and ash on the solar panel prevent recharging and drain a battery completely.

Rainfall is plotted as a cumulative value, in inches (fig. 3A), and is displayed for any period of time within a given year. Typically, rainfall over the previous few minutes, hours, and days has been of the most utility for forecasting lahars.

Figure 3. Example of rainfall and flow data, July 14, 1992. A, Cumulative rainfall. B, Full-frequency band (10-300 Hz). C, Low-frequency band (10-100 Hz). D, High-frequency band (100-300 Hz). Data were from RG F (Sacobia) and FM 1 (Lower Sacobia); flow observed 10 min later, about 2 km downstream from FM 1, was transitional between a debris flow and a hyperconcentrated streamflow.

ACOUSTIC FLOW MONITOR

The AFM (fig. 2B) uses inexpensive, ruggedized, L10AR digital exploration geophones to monitor ground vibration caused by lahars, streamflow, and other sources of similar frequency (Hadley and LaHusen, in press). The geophones have a -3-dB flat response between 10 Hz and 300 Hz and a maximum dynamic range of about 3.8x10^-3 cm/s vertical ground velocity (Fig. 3B). Ground vibration is monitored continuously in three frequency bands: the full band (10-300 Hz), a low band (10-100 Hz), and a high band (100-300 Hz).

A microprocessor-controlled digital telemetry platform measures the amplitude of the signal every second and then sends the data to the base station, at the Pinatubo Volcano Observatory on Clark Air Base, once every 30 min in normal mode. If the signal in a preselected frequency band exceeds a preprogrammed threshold value (adjusted for each site) for longer than 30 s (in 1993, >45 s), the AFM sends immediate reports that are flagged as "alerts." It continues to send flagged data at 1-min intervals for as long as the signal level in that frequency band remains above the threshold level. When the signal drops below the threshold, the AFM resumes normal operation, transmitting every 30 min. The triggering duration was set to 30 (later, 45) s to avoid recording volcanic and tectonic earthquake spikes, and the geophone was buried at least 1 m deep to avoid unnecessary noise from wind and rain.

When plotted against time, AFM data define a pseudohydrograph (fig. 3B) in which signal strength is a proxy for relative discharge. Potentially, with calibration to actual measured discharge at each AFM site, absolute discharge can be estimated from AFM data (Tuñgol and Regalado, this volume).

The acoustic flow monitor overcomes several difficulties that are inherent in the use of trip wires. AFM's provide enough data before and after passage of a lahar that the signal requires no corroboration, and information about the lahar is provided even after a trip wire in the same place would have broken. Furthermore, there is no need to visit each site after each lahar.

ROUTINE DATA RETRIEVAL

At the base station, a laptop computer displays raw data, time of transmission, and a channel identity that identifies the field station from which the data came and the type of data (for example, the low-frequency band of vibration, or rain gauge bucket tip count). Data are plotted onscreen and on an attached printer; they are written from the hard disk to a floppy and transferred to another computer for graphing or analysis with another software package, BOB (Murray, 1990). The BOB program is used primarily to display data from various rain gauges and flow sensors on single screens and pages so that the times and magnitudes of events can be compared from one part of the volcano to another.

REMOTE CONTROL OF FIELD SITE PARAMETERS

The base station computer is also used to issue commands to the field stations, to reset event-detection parameters or transmit intervals, or to interrogate a unit for up-to-the minute data. Communication between the rain gauges, flow monitors, and the base computer is through two-way UHF radio. The transmitter for one site also serves as the repeater for other sites. Each rain gauge is identified by a unique letter of the alphabet, and each acoustic flow monitor by a unique number (table 1).

Instrument control menus are available for setting and checking of operating parameters such as clock time of field instrument, repeater paths, routine transmission interval, next transmission time, duration for event detection, and threshold levels. When threshold levels of activity are exceeded, the instrument begins frequent "event mode" transmissions. The threshold for rainfall is fixed at 10 tips within less than the routine 30 min, and the thresholds for AFM amplitude are set between 150 and 1,000 AFM units (cm/s x10^-6) for 30 (later, 45) s or longer. AFM parameters had to be varied from site to site and empirically determined. The threshold settings during 1991 and 1992 were for Upper Sacobia, 300; Lower Sacobia, 700 in 1991 and 900 in 1992; Abacan, 1,000; Marella, 500; Pasig-Potrero, 300; Gumain, 200; and Maraunot, 150 AFM units.

There are also special commands for checking the condition of field instruments. Only one station and one authorized operator control the network; one additional, receive-only station receives and graphs data but cannot change the parameters of the field instruments.

INSTALLATION AND MAINTENANCE OF FIELD SITES

The lahar monitoring system was designed to operate under conditions of heavy rainfall, lightning, wind, and volcanic ash. Care was taken in site selection, installation, and maintenance to keep the system operating despite the hostile environment. Areas prone to erosion or lightning strikes were avoided where possible.

Precautions notwithstanding, we have had trouble with some sites. The biggest problems have been lightning, theft, and ash fall from secondary explosions. Even though each station uses one or two 100 A-h, 12-V batteries, connected in parallel and charged by an 18-W solar panel, stations close to the pyroclastic deposits often need to have ash cleaned from their solar panels. Solar panels are inclined 15 degrees to the south (Philippine standard) for optimum charging and also for protection against strong winds; steeper inclinations have been required in areas of unusually heavy ash fall from secondary explosions.

A partial list of problems encountered to date includes, in 1992, the following: the Culianan rain gauge and the Pasig-Potrero flow sensor were struck by lightning, batteries drained on the middle Sacobia rain gauge when ash covered the solar panel, and the lower Sacobia flow sensor was stolen. Wet ash also got into the BUGZ rain gauge and shorted out the battery. During the 1993 rainy season, the magnetic switch on the BUGZ rain gauge malfunctioned, the Summit rain gauge was damaged by lightning or buildup of static electricity, and the battery of the lower Sacobia flow sensor was stolen. Before a new battery could be brought to the site, the rest of the unit was stolen. Midway through the 1994 lahar season, telemetry line-of-sight problems are blocking data from (mostly new) sites on the west side, and ash from secondary explosions clogged the mechanism of the middle Sacobia rain gauge and covered the solar panels of the Pasig-Potrero flow sensor and (briefly) the upper Sacobia rain gauge. An alternate flow sensor that was installed at a police-manned lahar watchpoint was offline briefly when it was disconnected from our solar panel to charge a police battery.

PRELIMINARY RESULTS

In monitoring lahars of Mount Pinatubo from 1991 through 1993, the acoustic flow monitor system proved to be effective in lahar detection and warning. Some lahars were missed before instruments were installed and during instances of battery or other electronic failure; however, to our knowledge, no lahar passed by an operating flow sensor without detection. Empirically, flow sensors begin reporting increased vibration several minutes (as much as 1 km) before a lahar reaches the AFM station and continue above background until the tail of a lahar has passed. Data from the lower Sacobia and Pasig-Potrero flow sensors during the 1992 rainy season are given in Martinez and others (this volume) and Arboleda and Martinez (this volume), respectively; data from the 1993 lahar season are given in PHIVOLCS (1994).

Personnel who monitor the data at PVO can detect lahars 0.5 to 1 h earlier than watchers at manned, civil defense watchpoints, which are (necessarily) several kilometers downstream of our unattended instruments. PVO provides early warning to watchpoints manned by police and army personnel and receives from them, in return, a modest amount of information about the flow height, speed, and other flow characteristics. Utilization of PHIVOLCS' warnings by civil defense leaders is generally good but not perfect (Janda and others, this volume). Visual observations by PHIVOLCS staff were used to calibrate the Lower Sacobia flow sensor in 1992-93 (Martinez and others, this volume; Tuñgol and Regalado, this volume).

Data from several lahars during 1991 and 1992 suggest that energy from debris flows and hyperconcentrated streamflows is concentrated in the low band of the AFM's (10-100 Hz), while that from normal streamflow is concentrated in the high band (100-300 Hz). For example, on July 14, 1992, a flow past the Lower Sacobia flow monitor from 1800 to 2230 was seen, just downstream at the Mactan watchpoint, to be transitional between debris flow and hyperconcentrated flow. Its energy was strongly concentrated in the low band (fig. 3C,D).

Because the number of rain gauges per watershed (1 or 2) is small and rainfall is often localized, some lahars occur and are detected by AFM's without a corresponding rainfall signal. In addition, some lahars are caused by breakouts of small lakes, with or without concurrent rain. These "lahars without recorded rainfall" occur often enough that PVO reports them to the Office of Civil Defense without waiting for a rainfall signal. For example, at 1500 on October 12, 1991, a lahar was recorded by the Pasig-Potrero flow sensor without recorded or observed rains and was reported promptly to RDCC. No damage or casualties were reported.

Conversely, it is rare that heavy recorded rainfall does not result in a lahar, especially if the ground surface has been saturated by rain in preceding days. We now use such records to indicate that a lahar is likely. In one example, at 1800 on September 7, 1991, a general warning of lahars was issued on the basis of heavy rain at the Sacobia rain gauge. Less than an hour after the warning was issued, a lahar from the Pasig-Potrero River damaged San Antonio, Bacolor, and killed several people. Unfortunately, at this time the Pasig-Potrero flow sensor had not yet been installed. Field investigation conducted after the event revealed that the lahar event was brought about by the breaching of a newly formed lake in the upper reaches of the Pasig-Potrero River, presumably by overtopping and (or) erosion related to the heavy rainfall. Because the network of rain gauges is sparse, rainfall data cannot always tell us the magnitude of lahars or even the river channel in which they are flowing, but the AFM's provide this missing information.

Potential improvements in the network include for some AFM's to be placed at manned watchpoints and at variable distances back from the river channel. Such placement is needed to calibrate AFM data to actual discharge. Further correlation of rainfall and AFM signals and a denser network of rain gauges are needed for prediction of lahar discharge based upon rainfall. Finally, all stations need annual maintenance and special protection against moisture and ash in the electronics and against ash on the solar panels.

CONCLUSIONS

Judging from results from 1991 through 1993, a network of telemetered rain gauges and acoustic flow monitors is yielding important data for understanding and warning of Pinatubo lahars. Key factors in its success are (1) the small size and low power requirements of units, allowing them to be installed high on the volcano close to where lahars form, and (2) the choice of acoustic sensors rather than trip wires for flow monitoring.

At its best, the system provides civil defense and lowland residents with up to several hours of warning of impending and actual lahars. Typically, the system provides 0.5 to 1 h of extra lead time before flows are observed by manned watchpoints. At its worst, near the end of the 1992 lahar season, many stations were inoperative because of battery drain, site erosion, lightning strikes, and theft. New and repaired stations were installed in time for the 1993 and 1994 rainy seasons.

ACKNOWLEDGMENTS

The authors thank the Pinatubo Volcano Observatory staff for their valuable help and the U.S. Navy, U.S. Marines and the Philippine Air Force for their helicopter support for installations, repair, and maintenance of rain gauges and flow monitors. We also thank USAID Philippines and USAID Office of Foreign Disaster Assistance for financial support. Lastly, we thank M.L.P. Bautista, M.L.P. Melosantos, B.S. Tubianosa, R.U. Solidum, C.G. Newhall, R.S. Punongbayan, A.B. Lockhart, and M. Reid for constructive comments on our paper.

REFERENCES CITED

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

Bautista, B.C., Bautista, M.L.P., and Garcia, D.C., 1986, Seismic monitoring: A useful tool for mudflow detection at Mayon volcano, Albay, Philippines: Philippine Journal of Volcanology, v. 3, no. 2, p. 90-108.

Bautista, B.C., Bautista, M.L.P., Marcial, S.S., Melosantos, A.A., and Hadley, K.C., 1991, Instrumental monitoring of Mount Pinatubo lahars, Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 63.

Hadley, K.C., and LaHusen, R.G., 1991, Deployment of an acoustic flow-monitor system and examples of its application at Mount Pinatubo, Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 72, no. 44, p. 67.

------in press, Technical manual for an experimental acoustic flow monitor: U.S. Geological Survey Open-File Report.

Janda, R.J., Daag, A.S., Delos Reyes, P.J., Newhall, C.G., Pierson, T.C., Punongbayan, R.S., Rodolfo, K.S., Solidum, R.U., and Umbal, J.V., this volume, Assessment and response to lahar hazard around 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.

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

Murray, T.L., 1990, A user's guide to the PC-based time-series data-management and plotting program BOB: U.S. Geological Survey Open-File Report 90-56, 53 p.

PHIVOLCS, 1994, Lahar studies: Quezon City, PHIVOLCS, 80 p. Special report of UNESCO-sponsored studies.

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.

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

FIRE and MUD Contents

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

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