FIRE and MUD Contents

Rainfall, Acoustic Flow Monitor Records, and Observed Lahars of the Sacobia River in 1992

By Norman M. Tuñgol¹ and Ma. Theresa M. Regalado¹

¹Philippine Institute of Volcanology and Seismology.

ABSTRACT

Instrumental records and field observations of 1992 lahars show a strong correlation between acoustic flux (a measure of ground shaking near the lahar channel) and discharge rates as observed at the Mactan watchpoint. Integration of this correlation over time suggests that the cumulative volume of lahars during the 1992 rainy season was about 1.0x10⁸ cubic meters, a figure in reasonable congruence with the mapped 7x10⁷ cubic meters of 1992 lahar deposits.

Correlation of acoustic flow data and rainfall suggests that increase in lahar magnitude is roughly proportional to the amount of triggering and sustaining rainfall. Furthermore, lahars along the Sacobia River seem to be triggered by rainfall with intensities greater than I= 46D^-1.5, where I is the rainfall intensity in millimeters per minute and D is the duration of rainfall expressed in minutes. This remarkably low trigger threshold is due at least in part to low permeability and infiltration capacity of the fine ash-rich Pinatubo pyroclastic-flow deposits.

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

Virtually all of the lahars that flowed down the Sacobia-Bamban River were derived from the Sacobia pyroclastic fan, which is drained by the Sacobia, Abacan and Pasig-Potrero Rivers (fig. 1). Pyroclastic deposits in the Sacobia and the Abacan watersheds have an estimated volume of approximately 7x10⁸ m³, of which about 40 percent, or about 3x10⁸ m³, was expected to be eroded within a decade (Pierson and others, 1992). During 1991 alone, at least 1.6x10⁸ m³ of lahar deposit filled channels and buried nearby fields (Martinez and others, this volume). After the April 4, 1992, secondary pyroclastic flow, which cut off the Abacan River from the Sacobia pyroclastic fan, the Sacobia River became the sole channel draining the northeast lobe of that fan and carried about 0.7x10⁸ m³ of additional debris in 1992.

Figure 1. The 1992 lahar monitoring network along the Sacobia-Bamban River, northeast of Mount Pinatubo. The flow sensor shown in this figure is FM 1, Lower Sacobia, and the rain gauge is RG F, Sacobia, of Marcial and others (this volume).

Lahars along the Sacobia River are generally induced and triggered by rainfall. Large-scale lahar damming and lake breakouts similar to those of the Mapanuepe Lake, southwest of Mount Pinatubo (Umbal and Rodolfo, this volume), are not a significant lahar-triggering mechanism in the Sacobia watershed. Rain-induced debris flows have been the subject of many scientific papers, but only a few scientists have studied the detailed relationship of rainfall to debris flows in a tropical volcanic setting. Caine (1980) synthesized published records of 73 rainfall events associated with shallow landsliding and debris-flow activity from all over the world and came up with a limiting threshold for the initiation of debris flows, expressed in terms of the intensity and duration of the triggering rainfall. This data base included undifferentiated volcanic and nonvolcanic events. A similar study of the volcanic debris flows of Mayon volcano, Philippines, by Rodolfo and Arguden (1991) identified a higher threshold than that of Caine. For the lahars of Pinatubo, Umbal and Rodolfo (this volume) found that an average rainfall intensity of 0.3 to 0.4 mm/min over a period of 30 to 40 min was sufficient to trigger lahars on the Marella watershed in August 1991.

OBJECTIVES AND SCOPE OF THIS PAPER

This study adopted the approach of Regalado and Tan (1992), who correlated rainfall with lahar generation along the Marella River in 1991 by using rain-gauge and acoustic flow monitor (AFM) data. Here, we use instrumental and field data acquired along the Sacobia River during 1992 to (1) correlate the Sacobia AFM records with estimated flow discharge, (2) estimate the cumulative volume of 1992 lahars along the Sacobia River by using calibrated flow monitor data, (3) estimate the minimum intensity and duration of rainfall that triggered lahars along the Sacobia River, and (4) correlate the total amount of rainfall with the magnitudes of the associated lahars (rainfall-runoff relation).

SACOBIA LAHAR MONITORING IN 1992

For the 1992 rainy season, the system of lahar monitoring in the Sacobia watershed, as shown in figure 1, consisted of:

1. a rain gauge installed at midslope on the volcano to measure the amount of rain falling on the pyroclastic flow deposits;
2. an acoustic flow monitor (Hadley and LaHusen, 1991) installed near the lower portion of the pyroclastic fan, just below the lahar source area but well upstream from population centers, and
3. manned watchpoints for visual observations, measurements, and sampling of active lahars.

THE SACOBIA RAIN GAUGE

The Sacobia rain gauge is located about 10 km east of the crater, within the Sacobia pyroclastic fan, which provides the source materials for the lahars along the Sacobia River. The instrument consists of a standard tipping-bucket recorder that has been modified to accommodate ash fall (Hadley and LaHusen, 1991; Marcial and others, this volume). The data are transmitted to a computer at the Pinatubo Volcano Observatory at Clark Air Base (PVO-CAB) and recorded as the cumulative number of tips through time (1 tip = 1 mm of rainfall; Marcial and others, this volume). The data may be viewed in graph form on the computer screen. The Sacobia rain gauge, together with weather forecasts from the Philippine Atmospheric, Geophysical and Astronomical Services Administration (PAGASA) and the weather station of the U.S. Naval Air Station at Cubi Point, served as a lahar early-warning system during the 1992 lahar season.

THE SACOBIA ACOUSTIC FLOW MONITOR

The Sacobia acoustic flow monitor is located about 4 km downstream from the rain gauge, near the lower end of the Sacobia pyroclastic fan (fig. 1) (Marcial and others, this volume). The flow monitor records vertical ground velocities (in cm/sec x 10 ^-6) generated by the passage of a flowing mass along the river channel (Hadley and LaHusen, 1991, 1995; R. LaHusen, written commun., 1993). The flow monitor data, like those of the rain gauge, are radiotelemetered to PVO-CAB. Normally, the instrument sends data every 30 min. When it detects a sustained high-amplitude flow, however, an alarm is set off, and it starts transmitting every minute until the flow vibrations drop below the alarm threshold. Like those of the rain gauge, realtime data are available in graphical as well as numerical formats.

The flow monitor has been programmed to record acoustic amplitudes in different frequency bands: the low-frequency range (10-100 Hz), the high-frequency range (100-300 Hz), and the broad-range band (10-300 Hz). A lahar, being much more debris laden than normal streamflow or surface runoff, generates a low rumbling noise that correlates best in the acoustic records with the signals in the low-frequency range (fig. 2; Suwa and Okuda, 1985; Zhang, 1990; Hadley and LaHusen, 1991).

Figure 2. Acoustic-flow-monitor record of the August 28, 1992, flow events. The three events between 1000 and 1800 were debris flows dominated by low-frequency vibration, so the low-frequency wave band (10-100 Hz) recorded these lahars better than did the higher frequency band (100-300 Hz).

LAHAR WATCHPOINTS

During the 1992 lahar season, the Philippine Institute of Volcanology and Seismology (PHIVOLCS) maintained two observation posts along the Sacobia River: one near the Mactan Gate of Clark Air Base, about 15 km from the crater, and another at Sitio Maskup, Barangay Dolores, Mabalacat, Pampanga, about 8 km farther downstream. During rainy weather and lahars, flow discharges were estimated from active channel widths, estimated flow depths, and surface velocities (measured by timing floating objects as they traveled a known distance). Whenever possible, samples of the flows were collected and their temperatures measured, but this was not practicable during big lahar events.

Observations at the Mactan watchpoint were used in correlating field and instrumental data, as this watchpoint is closest to the instruments. Some change of flow magnitude might occur along the 2- to 2.5-km reach from the flow monitor site to the Mactan watchpoint, but proportionality between flow magnitude past the AFM site and flow magnitude at Mactan watchpoint was probably consistent enough to make correlation of AFM records and observations at Mactan meaningful.

CALIBRATION OF AFM RECORDS AND CORRELATION WITH RAINFALL

CALIBRATING THE ACOUSTIC FLOW MONITOR

To correlate flow monitor records and observed discharge and thereby "calibrate" the Sacobia AFM, we had first to determine flow traveltimes between the AFM and our watchpoint at Mactan. Figure 3 shows a striking correspondence between the graph of the low-frequency flow monitor record and the hydrograph at the Mactan watchpoint for lahars of August 29, 1992. Distinct flow pulses can be seen and correlated, from which we see that flows took between 5 and 15 min to travel from the flow monitor site to the watchpoint. However, not all flow signatures were as easily distinguishable as those of the August 29 events. Small-amplitude flow signals, particularly those below the alarm threshold amplitude and thus recorded at 30-min intervals, were correlated with lahars observed at Mactan within 40 min.

Figure 3. Low-frequency flow sensor record and the corresponding hydrograph at the Mactan watchpoint for the lahars on August 29, 1992.

For quantitative comparisons, we integrated the data over defined flow events or segments of flow events, rather than use instantaneous seismic amplitudes and discharge measurements. We define "acoustic flux" as the summed product of each amplitude reading and its duration and estimated this flux for 31 flow segments from 16 flow events in July and August (table 1 and fig. 4). Discharge and integrated volume measurements of the same flows were made at Mactan.

Figure 4. A, Observed discharge (D) at Mactan versus low-frequency flow sensor amplitude (A) for observed flows in July and August 1992. B, The measured flow volume (V) versus the "acoustic flux" (F) for observed flows in July and August 1992. All values are averaged over the duration of each event. Acoustic flux is a product of amplitude and duration.

The slope of the best fit between discharge and acoustic flux (fig. 4A) will vary greatly with distance of the sensor from the river channel, channel roughness, and type of flow. Therefore, the "calibration" shown here cannot be transferred to other sites even at Pinatubo, much less at other volcanoes. However, provided that the Sacobia AFM instrument is left unchanged, future AFM data can be a reasonable proxy for discharge, especially when darkness, severe weather, steam, and hazard prevent direct measurement of that discharge.

Table 1.Flow measurements made at the Mactan watchpoint correlated with the low-frequency (10-100 Hz) flow sensor data for observed events from July 1 to August 29, 1992.

[Values are for the specified time periods of actual observations at Mactan watchpoint. Amplitude and discharge are averages for those periods; acoustic flux is a product of amplitude and duration. Volumes were calculated by integrating instantaneous discharge through the hydrograph and thus differ slightly from the product of average discharge and duration. HF, hyperconcentrated streamflow; DF, debris flow; NSF, normal streamflow]

CORRELATING FLOW MONITOR AND RAIN GAUGE DATA

Any marked deviation from the background amplitude in the low-frequency flow monitor record may correspond to a lahar along the Sacobia. In 1992, however, some flows with AFM amplitudes of less than 100 units were muddy streamflows (nonlahars) rather than hyperconcentrated streamflows or debris flows. To avoid consideration of muddy streamflow, we considered only events with peak amplitude of at least of 100 units.

Rainfall events that were recorded by the rain gauge from June 1 to September 3, 1992, were classified into lahar-related and nonlahar events. Slightly modifying the definitions of Rodolfo and Arguden (1991), we define lahar-triggering rainfall as a rainfall event that includes no pauses longer than 30 min and results in a flow that registers an amplitude of at least 100 units in the low-frequency flow monitor record; sustaining rainfall is the additional rain that falls during the flow. The total rainfall is the sum of triggering and sustaining rainfall. Other rainfall events are considered nonlahar rain.

DISCUSSION

ESTIMATING LAHAR MAGNITUDE FROM AFM DATA

Figure 4A shows a general increase in average discharge with the average amplitude recorded in the low-frequency band by the flow monitor for 31 defined flow segments, representing 17 flow events, in July and August 1992. The relationship may be approximated by the equation

Q = 0.24A (correlation coefficient (r) =0.76),

where Q is the discharge at Mactan in cubic meters per second, and A is the average amplitude for each event. An exponential correlation actually yields a higher r value of 0.81, but the best-fit equation,

Q = 10.7e^0.029A,

indicated improbably high discharges for amplitudes of 2,000 and above.

Figure 4B shows that the flow volume at Mactan (discharge integrated over time) increased in rough proportion to the acoustic flux (AFM amplitude integrated over the same period). The correlation is best represented by a linear fit:

V = 14F (r=0.61),

where V is the estimated volume of each flow event in cubic meters, and F is the instrumental acoustic flux. This equation gives a total volume of about 1.0x10⁸ m³ for all 58 lahars detected by the flow sensor from June to early September (table 2). An exponential best-fit curve,

V = 6891e^2.46x10F,

has a lower r (.56) and gives unrealistically high flow volumes for modest acoustic flux values.

The 1992 lahar deposits were estimated to be about 7x10⁷ m³ on the basis of aerial photographs and rough thickness estimates in the field. We judge this to be in remarkably good congruence with the 1.0x10⁸ m³ estimated from the linear equation, especially in light of the fact that the volume of a lahar shrinks as water drains from it.

Table 2A. Lahars detected by the Sacobia flow monitor from June 4 to September 3, 1993, and the calculated flow volume for each event.

[This table includes many events that occurred during times when we were not present at the Mactan watchpoint. Accordingly, the acoustic fluxes and volumes in table 2A are generally higher than those in table 1]

Table 2A. Lahars detected by the Sacobia flow monitor from June 4 to September 3, 1993, and the calculated flow volume for each event--Continued.

Table 2B. Rainfall events that did not result in lahars.

[These events define the triggering threshold for lahars (fig. 5)]

TRIGGERING RAINFALL

Figure 5 presents data from 58 lahar-triggering and 94 nonlahar rainfall events as recorded by the Sacobia rain gauge from June 1 to September 3, 1992. A power curve,

Figure 5. Intensity (I) versus duration (D) of 58 lahar-triggering (filled squares) and 97 nonlahar (empty squares) rainfall events, derived from the flow monitor and rain-gauge data, from June 1 to September 3, 1992. The minimum threshold curve, described in the text, is the curve of rainfall above which lahars will nearly always result.

I = 46D^-1.5

(where I is the intensity in millimeters per minute, and D is the duration in minutes of the rain event), plots above all but 5 of 94 nonlahar rainfall events. This suggests that rain falling at a rate of at least 0.3 mm/min for 30 min or 0.1 mm/min for 1 hour is enough to trigger lahars along the Sacobia River.

Because we had only one rain gauge in the entire 50 km² Sacobia watershed, we need to ask whether the recorded rainfall is representative of the actual rains that triggered and sustained the lahars. The average size of convective storm cells at Pinatubo was about 10 km in diameter (Pierson and others, this volume), and thus roughly the size of the Sacobia watershed, but some rain was much more localized and was either heavier or lighter than that for the watershed as a whole (Hadley and LaHusen, 1991; Pierson and others, this volume). This is apparent in figure 5, which shows that many lahars occurred even when the recorded rainfall was comparable to that which at other times did not cause lahars; in fact, a few lahars have occurred after practically no recorded rain at all! In some and probably most of these instances, localized lahar-triggering rain occurred elsewhere in the watershed but not at the site of the rain gauge. Given this problem, we have estimated the threshold for lahar triggering (fig. 5) to be the maximum rainfall that does not trigger lahars, rather than the minimum (apparent) rainfall that does trigger lahars. With a large enough data set, the effects of localized rainfall become apparent, and we are reasonably confident that rainfall greater than the threshold curve in figure 5 usually will trigger a lahar.

The threshold defined in figure 5 for Mount Pinatubo is much lower than that of Rodolfo and Arguden (1991) for debris flows at Mayon:

I= 2.16D^-0.38

(converted from their equation I = 27.3D^-0.38, where their I was in millimeters per hour and their D was in hours). The threshold for Mayon was determined exclusively for debris flows, while the present study includes both debris flows and hyperconcentrated streamflows, and the threshold for hyperconcentrated streamflow is surely lower than that for debris flows. It is also likely, however, that the pyroclastic materials of Pinatubo are also mobilized more easily because they contain more fine ash, which causes lower permeability and infiltration capacity. Low permeability and low infiltration rates enhance runoff that eventually bulks up through headward and lateral erosion to form lahars.

RAINFALL AND LAHAR VOLUME

Figure 6 shows that the volume of each lahar event, as estimated from flow monitor data, is directly proportional to the total amount of rainfall that triggered and sustained the flow (r=0.77). Regalado and Tan (1992) defined a similar linear relationship between rainfall and the magnitude of lahars, as observed and as recorded by the flow monitor, for the Marella watershed in 1991. Their graph is reproduced here (fig. 7) for comparison. For 1992 events in the Sacobia watershed, instrumentally recorded rainfall also correlated well with the frequency of lahar occurrences (Martinez and others, this volume). Close correlation between rainfall and lahars was already known from field observations; the apparently linear relation between rainfall and (inferred) lahar volume is a new and potentially useful finding. Although the volume of lahars generated by a specified amount of rainfall will certainly decrease over time, as the watershed "heals," the relationship presented here might help to predict the volume of lahars that would result from various assumptions about future rainfall.

Figure 6. Total rainfall and the volume of the resulting lahars, the latter being calculated from the relationship between acoustic flux and observed discharge (fig. 4A), from June 1 to September 3, 1992.

Figure 7. Rainfall versus acoustic flux for August-September 1991 events in the Marella watershed (from Regalado and Tan, 1992). For these events, acoustic flux could not be converted to volume of lahars because considerable deposition and erosion of sediment, and addition of tributary flow, took place between the flow monitor and the nearest observation point, Dalanaoan, San Marcelino.

CONCLUSIONS

1. Lahars along the Sacobia River are triggered by rainfall exceeding the threshold curve

I = 46D^-1.5,

or above about 0.3 mm/min for 30 min.

2. The total volume of a flow increases linearly with total rainfall (that is, the sum of triggering and sustaining rainfall).

3. Sacobia flow monitor records and field estimates of discharge during 1992 were sufficiently well correlated that we could use flow monitor records to estimate discharge during periods for which we had no direct observations.

4. With this continuous estimator of discharge, we estimated the cumulative flow volume of lahars along the Sacobia channel at Mactan to be 1x10⁸ m³, a figure that correlates well with 7x10⁷ m³ of observed lahar deposits.

5. Future use of rain gauges to forecast lahar discharge should increase the density of rain gauges in each watershed. Future use of acoustic flow sensors to estimate lahar discharge should colocate the flow sensor at the observation post [editor's note: this was done at Mactan in 1993], make corrections for flow densities, and establish new calibrations at each flow sensor site.

ACKNOWLEDGMENTS

Rick LaHusen developed the AFM and modified the rain gauge used for instrumental data collection, and Kevin Hadley, Jeff Marso, and the PHIVOLCS Instrumentation Group (Joey Marcial, Arnold Melosantos, Ronald Pigtain, Arnold Chu, and Ramses Valerio) installed and repaired the instruments under difficult and hazardous conditions. Field observations of lahars were by members of the 1992 Pinatubo Lahar Monitoring Group: Onie Arboleda, Art Daag, Edwin de la Cruz, Peejay Delos Reyes, Atoy Garduque, Rey Macaspac, Mylene Martinez, Jed Paladio-Melosantos, Alex Pataray, Weng Quiambao, Bella Tubianosa, Mike Dolan, Rick Dinicola, Yvonne Miller, and the Newhall family. Staff of the Pinatubo Volcano Observatory--Jimmy Sincioco, Mike Eto, Manny Isada and Alex Ramos--watched the instruments for lahar alerts and watched out for our team of field observers. Onie Arboleda, Mylene Martinez, Ishmael Narag, Weng Quiambao, and Obet Tan offered critical help in data interpretation, as did Rick LaHusen, Chris Newhall, Tom Pierson, R.S. Punongbayan, Kelvin Rodolfo, and Rene Solidum in reviews.

REFERENCES CITED

Caine, N., 1980, The rainfall intensity-duration control of shallow landslides and debris flows: Geografiska Annaler, v. 62, p. 23-27.

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, p. 67.

Hadley, 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.

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.

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

Pierson, T.C., Janda, R.J., Umbal, J.V., and Daag, A.S., 1992, Immediate and long-term hazards from lahars and excess sedimentation in rivers draining Mount Pinatubo, Philippines: U.S. Geological Survey Water Resources Investigation Report 92-4039, 37 p.

Regalado, M.T., and Tan, R., 1992, Rain-lahar generation at Marella River, Mount Pinatubo: unpublished report, Quezon City, National Institute of Geological Sciences, University of the Philippines, 17 p.

Rodolfo, K.S., and Arguden, A.T., 1991, Rain-lahar generation and sediment-delivery systems at Mayon Volcano, Philippines: Sedimentation in Volcanic Settings, SEPM Special Publication no. 45, p. 71-87.

Suwa, H., and Okuda, S., 1985, Measurement of debris flows in Japan: Proceedings of the IVth International Conference and Field Workshop on Landslides, Tokyo, p. 391-400.

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

Zhang, S., 1990, Geosound characteristics and measurement of debris flow at Jiangjia Gully, in Wu, J., Kang, Z., Tian, L., and Zhang, S., eds., Debris flow observations and research in Jiangjia Gully, Yunnan: Beijing, Science Press, p. 141-164.

FIRE and MUD Contents

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

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