FIRE and MUD Contents

People's Response to Eruption Warning: The Pinatubo Experience, 1991-92

1 Philippine Institute of Volcanology and Seismology.

By Jean Tayag,1 Sheila Insauriga,1 Anne Ringor,1 and Mel Belo1


Two posteruption surveys, one in 1991 and another in 1992, assessed whether eruption warnings were received, understood, and used by citizens to take protective action. The 1991 survey showed that 71 percent of the total number of respondents (234) were forewarned; the remaining 29 percent learned of the hazard on June 12 by seeing the first large explosive events, a fact that indicates some weakness in the warning transmission.

Evacuation orders were issued by concerned Disaster Coordinating Councils or local government officials soon after danger zones were declared on April 7, June 7, and June 14-15, 1991. Eighty-six percent of the respondents received an evacuation order, but 30 percent of these people received it 2 or more days after it was issued.

Of those forewarned, 82 percent took protective action, including 46 percent who evacuated. However, within the group that evacuated, some waited two or more days after receipt of an evacuation order before moving, and some merely evacuated their women, children, and elderly or evacuated but returned. Some who did not evacuate as advised thought the eruption would not be strong enough to affect their places; others were reluctant to leave behind their houses and household effects, livestock, and crops, especially at harvest time; still others had no ready means of transport and could not walk long distances, or they believed that their God, Apo Namalyari, would not let them come to harm. Eventually, all but 5 of 234 respondents evacuated, either before or during the eruptions.

Communities in which LAKAS, an organization of the indigenous Aetas, was active showed the most exemplary operation of the system:transmission was total and response was consistently appropriate. These communities were reached by an information drive that featured the Maurice Krafft videotape on volcanic hazards, which he made for the International Association of Volcanology and Chemistry of the Earth's Interior (IAVCEI).

When Pinatubo threatened again to erupt in 1992, more than 90 percent of the respondents were forewarned and responded appropriately, indicating a marked improvement in the system. However, some overreaction was observed as an evacuation order was received by respondents who lived outside the danger zone. The errant evacuation order was traced to two sources: (1) some local government officials, who interpreted the Alert Level 5 released by the Philippine Institute of Volcanology and Seismology on July 14, 1992, to mean an eruption similar to that of June 12, 1991, and, hence, evacuation of the 20-kilometer-radius danger zone and (2) a popular radio announcer who broadcast that an eruption was imminent within 72 hours. The discrepancy between the warning message released by the source and that which was actually received appears to be a simple transmission problem. But other factors, including some features of the alert levels, may have inspired overexpectations and overreactions.

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.


If timely warning can be given of an impending disaster-causing event, the severity of the resulting disaster or adverse consequences can be reduced. The Mount Pinatubo 1991 eruption provides an excellent example of how accurate forecasting and timely warning saved lives from the destructive agents unleashed by a violent eruption. The number of casualties at the height of the June 1991 eruptions was small (only 200 to 300) despite the violence of the explosions and the vastness of the area affected. Early, perceptible signs from the volcano and prompt warning and mobilization of disaster-response officials minimized the human losses. It is precisely on account of its success that the Pinatubo warning system makes an interesting object of review. Its strengths, as well as its imperfections, provide insights on how other volcano-eruption warning systems could be developed or improved.

The degree to which the severity of the disaster can be reduced by warning depends on the interplay of the major components of a warning system, namely (1) the source and timing of the warning, (2) the warning message, (3) the warning transmission, and (4) the recipients' response (modified from UNDRO, 1986).

A team from the Philippine Institute of Volcanology and Seismology (PHIVOLCS) assessed all four aspects of the Pinatubo warning system to identify areas of success and those which needed improvement. The review involved two sampling surveys among the affected households: the first conducted within a month after the June 1991 major eruptions and the second during the month following the declaration of Alert Level 5 in July 1992.


The 1991 survey was by stratified random sampling of respondents who had lived in the danger zones or zones recommended for evacuation. Respondents were selected from barangays that lay within 10 km, 10 to 20 km, and 20 to 40 km of the volcano's preeruption summit, radii that formally defined danger zones (fig. 1). 1990 census figures indicate that the barangays within the 10-km and 10- to 20-km danger zones had 7,653 households, or 41,100 residents; the 20- to 40-km danger zone, which included 106 barangays in 17 towns, had 58,696 households and more than 331,000 inhabitants (National Statistics Office, 1990).

Figure 1. Map of the Mount Pinatubo area showing recommended evacuation zones ("danger zones") of various radii, and barangays cited in the text.

The only recommendation for evacuation in 1992 was for the danger zone within <10 km of the summit. Very few people were affected, because most former residents of this zone had remained in evacuation camps or resettlement areas since 1991. Our survey was conducted among the next nearest population, from the 10- to 20-km danger zone. On the eastern side of the volcano, most barangays within the 10- to 20-km danger zone that were sampled in 1992 had only about half of their original pre-1991 eruption populations. The other residents had either relocated or were still in evacuation centers. On the western side, most of the former residents of the sample barangays in this zone were living (officially) in the relocation sites, but many were also spending days, weeks, or even months on their preeruption land planting and gathering food, whenever they felt it safe enough to do so. Some looked on the relocation site as a kind of "bakasyunan" or vacation home.

In both surveys, sampling size was determined by using a normal variable (z) value of 1.96 (see appendix 1 for the formula and computation). The survey covered only the survivors and is biased in favor of those who took precautions. However, those who died constituted a very small percentage of the population at risk, so the resulting bias is deemed insignificant. The respondents were of two types: households (with the household head or an adult household member as respondent) and key informants from among barangay and municipal officials. Household respondents were randomly selected from lists of household heads provided by barangay leaders, with substitutions when the original respondents were either unavailable or unwilling. Interviews were conducted by PHIVOLCS staff and volunteers (local school teachers) with the aid of interview schedules (appendix 2) and, as needed, interpreters.



Normally, the source of eruption warning should be the entity tasked to study and monitor active volcanoes. In the case of the Philippines, this entity is PHIVOLCS. But when Mount Pinatubo started showing signs of restiveness in April 1991, PHIVOLCS had no monitoring at the volcano and, hence, no warning system for the area.

Consequently, it was not PHIVOLCS that recognized the first signs of volcanic unrest but, rather, indigenous Aetas who lived on the slopes of the volcano. Some of these Aetas, members of Lubos na Alyansa ng mga Katutubong Ayta sa Sambales (LAKAS) (Negrito People's Alliance of Zambales), reported their observations to PHIVOLCS through Sister Emma, a sister of the Franciscan Missionaries of Mary (FMM) who was doing missionary work among the Aetas.

Upon receipt of the LAKAS report, PHIVOLCS immediately began to monitor Pinatubo and, thenceforth, became the principal source of warnings. Details of the monitoring activities and chronology of preeruption events are given by Sabit and others (this volume) and Wolfe and Hoblitt (this volume); details of preeruption warnings are given by Punongbayan and others (this volume). Those warnings provided enough lead time for the beleaguered inhabitants to pack up and run away from the volcano.

The Warning Message

The warning message consisted of hazard zonation maps, alert levels, and "danger zones," which were zones of recommended evacuation, simplified from the hazard maps. Preliminary hazard zonation maps were disseminated by PHIVOLCS on and after May 23, 1991. These maps delineated the areas likely to be affected by the destructive agents, namely, pyroclastic flows, ash fall, and lahars. These maps illustrated the probable extent of the most probable hazards and served as guides for evacuation of endangered communities. Since the major eruption of June 15, 1991, the lahar hazard part of these maps has been updated several times.

Alert levels were designed to describe various levels of eruptive activity and danger. These provided information on the condition of the volcano, including whether its activities would likely culminate in an eruption. The alert levels were based on instrumentally derived data and daily visual observations. The original scheme of alert levels that was released on May 13, 1991, is shown in table 1.

Table 1. Alert levels for Mount Pinatubo, May 13, 1991.

Table 2. Alert levels and danger zones issued on Mount Pinatubo, 199192 (PHIVOLCS, variously dated).

Daily volcano bulletins and special advisories on the volcano's condition always indicated the alert level and an associated danger zone that should be avoided and evacuated. PHIVOLCS' recommendations for evacuation were translated and transmitted by the concerned Disaster Coordinating Councils (DCCs) or local government officials as evacuation orders. A chronology of alert levels and danger zones declared in connection with Pinatubo's activity in 1991-92 is presented in table 2.

When the temporary seismic station installed near the volcano recorded high seismic activity on the first 3 days of operation, April 5-7, PHIVOLCS declared a danger zone of 10 km radius that was centered on the volcano's summit and advised evacuation of the residents from the area. Initially, volcanologists considered employing an alert level terminology used at other Philippine volcanoes but opted to design a new one for Pinatubo (table 1). The 10-km danger zone was reiterated when the alert level scheme shown in table 1 was officially adopted on May 13, 1991, and Alert Level 2 was raised. It was maintained even when Alert Level 3 was raised on June 5.

When Alert Level 4 was declared on June 7, the danger zone's radius was increased to 20 km. On June 14, this was further expanded to the 30-km radius. During the June 15 explosions, the danger zone was expanded to a 40-km radius to allow for the possibility of devastating, large-scale pyroclastic flows of a caldera-forming eruption. The danger zone was shrunk back to a 20-km radius on June 18, though Alert Level 5 remained. On September 4, the alert level was lowered to 3, and the danger zone was shrunk back to a 10-km radius. The alert level was further lowered to 2 on December 4. Alert Level 2 remained in effect until the volcano started manifesting a resurgence of activity in July 1992.

In 1992, renewed seismicity prompted PHIVOLCS to raise the Alert Level to 3 on July 9 and then to 5 on July 14, when viscous lava reached the surface and began to form a dome. The 10-km danger zone, in effect since September 1991, was maintained throughout the 1992 unrest.

The volcano's 1992 activities were entirely different from its 1991 eruptions. These were characterized by quiet effusion of lava and dome building punctuated by minor explosions and hence were not as explosive and hazardous as the 1991 events. Realizing the need to reflect these differences in the alert level scheme, PHIVOLCS revised the definitions of alert levels toward the end of the year. The first three alert levels were retained with only a slight revision of Alert Level 3 interpretation, but Alert Levels 4 and 5 were substantially modified (table 3).

Table 3. Revised alert levels for Mount Pinatubo (revised December 1992).

Alert Level 4 will be used only for impending hazardous explosive eruptions or for ongoing eruptive activity that involves only small explosions or lava dome extrusions. Alert Level 5 will be used only for large explosive eruptions in progress. Definitive time windows for the occurrence of an eruption, such as "eruption possible within 2 weeks" for Alert Level 3 and "eruption possible within 24 hours" for Alert Level 4, were modified to "within days to weeks" and "within hours to days," respectively.

Transmission of Warning

At the five most active volcanoes being monitored by PHIVOLCS--Mayon, Bulusan, Taal, Hibok-Hibok, and Canlaon--eruption warnings are usually passed through the appropriate DCC. When one of these volcanoes manifests abnormal behavior, PHIVOLCS interprets its changing behavior and decides whether or not to send warnings and, if so, when. As soon as PHIVOLCS decides to issue a warning, it notifies the Office of the President and the national and local DCCs, through Volcano Bulletins and advisories that explain the condition of the volcano and recommended actions. The DCCs, in turn, transmit the warning to those at risk and respond in various other ways. Although transmission of warnings is officially the responsibility of the DCCs, PHIVOLCS observatory personnel help deliver warnings to nearby inhabitants. Later, PHIVOLCS' main office might release information to the media to clarify and explain the volcano's condition.

This warning procedure was modified in the case of Pinatubo. Warning messages were formulated at PHIVOLCS' main office and transmitted simultaneously through the DCC hierarchy, major national and local newspapers, radio and television stations, nongovernmental organizations (NGOs), and directly to the endangered inhabitants.

Multipath warning transmission has been found to create confusion, duplication, and administrative problems in some situations. This is why, at other monitored Philippine volcanoes, warnings and evacuation advice are passed, as much as possible, through the concerned DCCs. The effectiveness of the modified transmission procedure adopted at Pinatubo was assessed by use of two indicators: (1) consistency between the warning message released by the source (PHIVOLCS) and the message received by the recipients and (2) the time gap between issuance from the source and receipt by the target public.

Respondents were asked if they received any eruption warning and (or) evacuation order, and, if so, when. The 1991 survey showed that 71 percent of the 234 respondents knew of the impending eruption before June 9, 1991, the date on which Alert Level 5 was issued, either through their own observation (9 percent) or through their own observation and forewarning from PHIVOLCS, media, local officials, or other people (62 percent). Before June 12, the date of the first large explosive events, 82 percent of the respondents knew of the danger. Subtracting the 9 percent (roughly) who knew only through their own observation, it appears that about one-fourth of the respondents were not reached by warnings before explosive eruptions began. By June 14, 99 percent of the respondents knew of impending danger, from continued warnings and, especially, from observing the preparoxysmal eruptions (Wolfe and Hoblitt, this volume).

Danger zones that were delineated by PHIVOLCS served as basis for the DCC's issuance of evacuation orders. Evacuation orders were transmitted soon after the danger zones were declared by PHIVOLCS, on April 7, June 7, and June 14-15. Those within the 10-km danger zone should have received their order as early as April 7; those within the 10- to 20-km danger zone, on June 7; and those within the 20- to 40-km danger zone, on June 14-15.

Two hundred and two (86 percent) of the 234 respondents received an evacuation order while 14 percent did not. Of the 32 respondents who did not receive an evacuation order, all except two (one from the <10 km danger zone and one from the 10- to 20-km danger zone) were from the more distant, 20- to 40-km radius danger zone (table 4). Most received their evacuation orders on the day or 1 day after the danger zone was declared by PHIVOLCS; others received the evacuation order 2 or more days later, reflecting some delay in the transmission. Delays in transmission were reported in all the danger zones.

Table 4. Receipt of eruption warning and evacuation order by date.
[Household survey, 1991; number of respondents was 234; EW, eruption warning; EO, evacuation order; cum%, cumulative percentage of the 234 respondents]

Table 5. Respondents who received preeruption warning and (or) false evacuation order.
[Household survey, 1992; number of respondents was 130; all respondents for the 1992 survey were from the 10- to 20-km danger zone]

In 1992, 94 percent of the respondents learned of the impending eruption on or before July 14, the day PHIVOLCS issued Alert Level 5 (table 5). Almost all the respondents received the warning from multiple sources, with PHIVOLCS, the media, and military officials as the most common transmitters.

Throughout the 1992 activity, PHIVOLCS merely reiterated the continued enforcement of the 10-km danger zone. Therefore, evacuation was recommended only for those who had returned to the <10-km danger zone despite advice against reoccupation of the area. Because the 1992 survey was confined to the 10- to 20-km danger zone, no respondent was expected to have received an evacuation order. Surprisingly, 8 percent of the respondents (all from Pampanga) reported that they received an evacuation order either from municipal or barangay officials or through the media.

One Municipal Disaster Coordinating Council (MDCC) official admitted that the council decided to order evacuation of barangays beyond the 10-km but within the 20-km radius (including one community that was already living in a relocation center) on the night of July 15. The council even provided vehicles to bring the evacuees to the evacuation centers. According to the town official interviewed, they wanted to play it safe. If the anticipated eruption would be similar to the June 12, 1991 eruption, then the 10- to 20-km danger zone would be affected, so why should they wait?

Respondents from the villages Sapangbato and Margot of Angeles City reported that sometime before July 14, 1992, a popular radio announcer, citing PHIVOLCS as his source, broadcast that Mount Pinatubo would erupt within 72 h. It is interesting to note that the PHIVOLCS alert levels do not include one that indicates that the volcano may erupt within 72 h. The signal with the closest time reference is Alert Level 4, which means that eruption is possible within 24 h. But Alert Level 4 was not used in 1992, as the Alert Level jumped from 3 to 5.

The People's Response to Warnings Issued

The final test of a warning system's effectiveness is the receipt of and appropriate response to the warning by the target recipients. The warning, no matter how timely, accurate, and precise, will not be of any value unless the recipient of the warning takes appropriate defensive action.

Response to 1991 Warnings and Evacuation Orders

One hundred sixty-seven respondents, representing 71 percent of the total number of respondents (234), had forewarning of the eruption and were asked what they did when they learned that the volcano was going to erupt. About 81 percent of those who received forewarning took appropriate action, by evacuating immediately (the response that was called for in the case of those living within the 10-km radius as early as April and those living within the 10- to 20-km radius starting June 7) or by taking some other defensive action (such as preparing for evacuation, convening a meeting, disseminating the warning, seeking further information or confirmation, or observing for further signs, responses that were appropriate at radii of 10 to 20 km from April to June 7 and at radii of 20 to 40 km until June 15). Eight respondents from the <10-km danger zone who should have evacuated immediately merely took other precautionary actions like preparing for evacuation, seeking additional information, or watching out for further developments. Another 13 respondents (from the 20- to 40-km danger zone) overreacted by evacuating before they were ordered to do so.

In contrast, 13 percent of those who were forewarned either waited for the eruption or ignored the warning; and 5 percent ran without definite destination, prayed, or cried without taking any defensive action (table 6).

Table 6. Response to preeruption warning in each of the danger zones.
[Household survey, 1991; number of respondents was 167]

Table 7. Response to evacuation order, 1991.
[Number of respondents was 234]

Table 8. Reasons for not evacuating immediately as when advised.
[Household survey, 1991; number of respondents was 69]

The responses to evacuation orders (a step beyond warnings) indicate that all the respondents except five (2 percent) eventually evacuated (table 7). Fifty-eight percent of all respondents evacuated when and as advised, and an additional 11 percent evacuated even without or before receiving evacuation order. But 23 percent delayed evacuation and 6 percent evacuated selectively. Households that evacuated selectively either (1) sent their women, children, sick and elderly to safety while the able-bodied adult males stayed behind or (2) evacuated all together but then allowed some member(s) of the household to return home (usually during daytime).

Table 8 lists some of the reasons given by those who dallied or evacuated selectively. Some thought that the eruption would not be strong enough to affect their place; some were reluctant to leave behind their property and livelihood, especially as it was harvest time; some had no ready means of transport and could not walk long distances; and some believed that their God, Apo Namalyari, would not let them come to harm. One respondent did not want to leave immediately because it was fiesta time.

Responses to 1992 Warnings

In 1992, PHIVOLCS advised the inhabitants of the 10- to 20-km danger zone to avoid the 10-km danger zone (where some residents would otherwise hunt or gather food or tend farm plots), be alert to possible deterioration in the volcano's condition, and prepare for this possibility. About 8 percent of the respondents received an evacuation order (table 5), some from their local officials, others through radio. Two percent of the respondents evacuated their entire households and another 3 percent evacuated selectively (table 9). The latter were from one community that was ordered by the municipal DCC to evacuate but, instead of complying fully, sent only its women, children, elderly, and sickly to the evacuation centers, where they stayed for about 3 months. Another community, about 15 km from the volcano, was ordered to evacuate and was even provided with vehicles for evacuation, but it refused to comply because residents believed that Apo Namalyari would protect them.

Table 9. Response to preeruption warning and false evacuation order.
[Household survey, 1992; number of respondents was 130; all from the 10- to 20-km radius danger zone]

Had there been a real need for evacuation, the noncompliance of the recipients would have exposed them to danger. But, because the evacuation order was an overreaction on the part of the concerned officials and misinformation on the part of the radio announcer who broadcast "warning" of an imminent eruption within 72 h, their noncompliance led to no harm. Nevertheless, the more appropriate response on the part of the recipients of the evacuation order should have been to seek further information or verification of the order, instead of not to comply.

Among those who took other defensive action (table 9), some overreacted by suspending their normal activities, such as going to school or going to work, for 1 to several days. This overreaction occurred principally while waiting for the "imminent eruption within 72 hours" that was broadcast by the irresponsible radio announcer sometime before July 14, 1992.


The fact that most respondents took appropriate defensive actions and evacuated as advised indicates that the warning system worked well enough in 1991. It performed even better in 1992. Nevertheless, some aspects could still be improved. For improvement, the following findings are particularly important:

1. The failure, in 1991, of 18 percent of those who were forewarned to take any defensive action and the delayed or selective evacuation of 34 percent of those who received evacuation orders indicate some failure to stimulate protective action.

Inhabitants who received warnings and evacuation orders but did not take defensive action obviously lacked appreciation of the magnitude of the dangers posed by the volcano. Even those who delayed evacuation or evacuated selectively showed lack of understanding of the gravity of the threat. Either the information drive launched by PHIVOLCS and other disaster response organizations before the eruption did not reach these respondents or the information campaign failed to drive home to them the magnitude of the threat and the urgency as well as the possibility of avoiding the volcano's fury.

It is worth pointing out that all of the respondents contacted by the LAKAS organization showed the exemplary appropriate response. All (except one old man who chose to die rather than leave his home) prepared and evacuated promptly. These respondents recounted that, before the eruptions, the eruption threat and the hazards posed by the volcano had been explained to them by PHIVOLCS and other officials. They had also been shown the videotape on volcanic hazards produced by the late Maurice Krafft for IAVCEI (Punongbayan and others, this volume).

Some other Aetas did not fare as well. According to one informant (a Protestant pastor), a group of Aetas was about to evacuate (on the 15th of June) along with the others who were fetched by chartered buses. But these people changed their minds when they could not read sign boards on the buses that indicated which should be boarded by Villar residents, by Moraza residents, and so on. After boarding at random and being twice informed that they were in the wrong bus, they were so embarrassed that they decided to return to the mountain and seek refuge in the so-called caves, saying that Apo Namalyari would protect them.

Two other informants said half of the residents of sitio Lomboy were very reluctant to evacuate. Most of them did not want to leave their belongings, crops, and livestock and believed that Apo Namalyari would not let them come to harm. Many of them did not believe that the eruption would be strong enough to affect their places. Some feared lowlanders would burn their crops and homes. A Korean pastor was finally able to convince them to leave, but they put off their departure until the next morning and spent the night in some kind of natural shelter that they called caves. That night, pyroclastic flows buried the caves and killed those inside.

The findings of the survey corroborated news reports about the reluctance or refusal of some endangered inhabitants to leave the danger zones. In April, Aeta tribesmen who refused to move out reportedly said "they were afraid to leave their 'precious belongings&'" (Alcayde, 1991) or reasoned that they could not leave because their camote crops were due for harvesting. The latter group promised to leave as soon as the harvest was done (Gob, 1991). One Aeta leader who stayed behind was quoted to have said "Mahina lang siguro ang pagsabog dahil hindi naman ito narinig dito sa Belbel" ("The eruption was probably weak because it wasn't heard here at Belbel"), referring to the April 2 explosion (Empeno, 1991).

By June 9, Mayor Richard Gordon of Olongapo City was reported to have dispatched trucks to "clear" barangays within the 20-km danger zone where "there were still some Negritoes who chose to stay where they were, because of their livestock and other properties" (Villanueva and Dizon, 1991). One of the holdouts compromised by sending his family not to an evacuation center but to a place a bit farther away from the volcano, saying "Hindi naman daw kami aabutan ng pagputok ng bundok" ("We heard that the eruption will not reach us") (Cortes, 1991).

Hours after intensified ash emission on June 9, evacuees who earlier refused to leave were reported to have finally climbed into trucks brought by rescuers. Hundreds of Aetas with their belongings and work animals lined the roads, waiting for trucks to bring them down to evacuation centers. But there were still others who refused to be evacuated (Velarde and Bartolome, 1991). The mayor of San Marcelino reported that during rescue operations on June 9, 10 Aeta families opted to stay, believing that the eruption was nothing serious--"para lang daw malakas na bagyo'yan" ("it is just like a strong typhoon") (De Villa, 1991).

As late as June 11, Zambales Governor Deloso reported that some 200 tribesmen still refused to leave their settlements in Barangays Moraza, Nacolcol and Maguisguis. He added that the men may have wanted to stay to harvest their palay and camote crops so they could repay their loans to the Land Bank of the Philippines (anonymous, 1991a). An Aeta in Moraza who defied the evacuation order and stayed on to keep an eye on his home, farm, and carabao (water buffalo) was quoted to have said "We fear the volcano, but if we left our carabaos, we'll die" (Morella, 1991). In another barangay, Belbel, the barangay captain reported that some 252 tribesmen also refused to leave their homes (Anonymous, 1991b).

The anecdotes from the survey informants, and these news reports, highlight some of the communication and cultural problems with which the warning system had to contend. These problems indicate a need for hazard-awareness promotion that is more intensive and broader in outreach than was possible during the 2-month period from the time the volcano started showing signs of restiveness up to the time of the major explosions.

2. In 1991, the failure of 18 percent of the respondents to receive preeruption warning before June 12, the failure of 14 percent of the respondents to receive an evacuation order at any time (even after the June 15 eruption), and delay in the receipt of evacuation order by 26 percent of the respondents all indicate some deficiencies in the transmission system. In the <10- and 10- to 20-km danger zones, it is possible that the transmission network did not reach the most remote areas or the communities that were on the move.

Again, it is worth noting that in the case of the communities with a grassroots organization like LAKAS, warning transmission was total despite difficulties of transportation and terrain. The organization ensured that everyone received the warning and evacuation order.

The official warning system was unable to reach all residents of the large, 20- to 40-km danger zone during the short, hectic time that that zone was in effect (June 15-18). Communities in this zone are easily accessible but too numerous to reach in such a short time. Broadcast radio served as the principal channel for warning communities in this area. The warning communication system was improved in 1992 by the distribution of two-way radios to barangay leaders.

3. In 1992, the receipt of a false evacuation order by 8 percent of the respondents is a clear case of discrepancy between the warning message released by the source and the message transmitted to the concerned inhabitants. Issuance of this evacuation order for communities outside the official danger zone may have been a simple case of caution or overreaction on the part of the local officials. However, incentives for evacuation such as the availability of relief and emergency resources and the usual outpouring of sympathy might also have inspired the move.

The overreaction may also be traced, at least in part, to the warning messages released by PHIVOLCS. The United Nations Disaster Relief Office (UNDRO, 1986, 1987) advised, among other things, that warnings should be consistent in content and as specific as practicable in their information concerning the magnitude of the event, the place at which it is expected, and the time when it will occur. The 1991 alerts were indeed specific--in terms of expected magnitude, areas likely to be affected, and time of occurrence--but they were specific to the 1991 eruptive activities. The application of these same alerts to the less explosive and less hazardous 1992 events may have given rise to undue concern and inspired exaggerated media reporting.

After the 1992 experience, revision of the alert levels was in order. The revision removed the implication that eruptions could be predicted to the nearest hour or day, especially at volcanic systems in which the vent was already open. The original alert levels focused mainly on the imminence or occurrence of a large explosive eruption. The revised alert levels allow for differentiation of large and small eruptions. Missing still are the recommended actions for each of the alert levels and each of the danger zones. How to incorporate these without making the scheme of alert levels inflexible and too specific remains to be studied.

One specific aspect of the PHIVOLCS warning messages that went against UNDRO advice was the inconsistency in the danger zones associated with the various alert levels. In 1991, Alert Level 4 was associated with a 20-km danger zone, and Alert Level 5 was associated with both a 20-km and a 40-km danger zone. A 40-km danger zone was declared because there was concern about pyroclastic flows from a big eruption and the possibility that a caldera might form. The alert level-danger zone association, though not intentionally established, lingered, so that when Alert Levels 4 and 5 were released in 1992, an understandable reaction was to react as in 1991 and evacuate the 10- to -20-km danger zone.

There are at least two options for rectifying this source of potential misunderstanding. One is to explore the possibility of striking some correspondence between alert levels and danger zones. The assumption is that it is possible to determine the areas likely to be endangered by each type and magnitude of activity referred to in each alert level. However, eruptions vary in style and intensity, so such a correspondence may not be feasible. The alternative is to consciously dissociate the alert levels from danger zones, define a permanent danger zone, and keep other danger zones open-ended and adjustable.

Another possible improvement in the alert level scheme would be to reword the "Interpretations" and specifically the phrase "eruption is possible within 2 weeks [or 24 hours]." That phrase was variously interpreted to mean "eruption will occur 2 weeks [or 24 h] hence" or that "an eruption would occur within 2 weeks [or 24 h]." Ironically, the wording was actually chosen to avoid making specific predictions. Rather, it was meant to define a window in which an eruption was possible and to indicate disappearing margins of safety. Thus, at Alert Level 2, an eruption within the next 2 weeks was judged unlikely, but at Alert Level 3, this was no longer true. At Alert Level 3, an eruption was unlikely within less than 24 h, but at Alert Level 4 all reassurances of safety was gone--an eruption could occur at anytime (C.G. Newhall, written commun., 1994). The Pilipino translation of the phrase "eruption is possible within. . ."--"maaaring mangyari ang pagputok sa loob ng" would have conveyed the message that the authors meant to convey. A Pilipino version of the alert level scheme could be pilot tested the next time one of our volcanoes becomes restive.

The broadcast of a warning that an eruption was imminent within 72 h, falsely attributed to PHIVOLCS, triggered discussions on the wisdom of the modified warning transmission procedure adopted at Pinatubo. Critics of the multipath transmission procedure claimed that had PHIVOLCS stuck to the DCC channel instead of directly dealing with the media, reporters would not have been able to cite it as their source for their false or sensationalized reports. The traditional DCC channel would certainly minimize the warning source's need to deal with the media and make it easier to pinpoint responsibility for erroneous reporting. But it would also limit the area that could be reached, given a short lead time for warning dissemination. The institution of an emergency broadcast system might provide a mechanism for effectively involving media in warning transmission. The concept is not new, and there have been attempts to establish such a system. The idea is worth reviving. An emergency broadcast network could be identified, with media representatives officially identified and properly trained to handle warning and emergency response operations. This move might not eliminate exaggerated or fabricated reports but could minimize the effect of such reports if people listen to and believe only the official transmitters.

4. The fact that 94 percent of the respondents in 1992 knew of the impending eruption before Alert Level 5 was released and that 92 percent responded appropriately indicates improvement in warning transmission as well as in inducing optimal response. However, because no evacuation was required, the improvement in the percentage of appropriate response may be more apparent than real. The question remains, would this percentage be as large should there be a call for evacuation of areas beyond the 10-km radius?

Mount Pinatubo's continuing activity provides an excellent opportunity to continue the development of the eruption warning system. We hope the experience in evolving a suitable warning system for Pinatubo will yield some valuable lessons for issuing warnings at other active volcanoes.


The authors are deeply indebted to Dr. Dennis Mileti and Dr. C. Dan Miller, whose comments and suggestions did not only enrich the final output but provided a rich source of learning for the authors as well, and to Dr. Chris Newhall for his patience, relentless prodding, and meticulous attention to detail.


Alcayde, Jerry, 1991, 1876 families move out of volcano area: The Philippine Star, April 22, 1991.

Anonymous, 1991a, 200 Aetas won't leave endangered settlements: Philippine Daily Inquirer, June 11, 1991.

Anonymous, 1991b, `Big bang' looms, Yanks flee Clark: Philippine Daily Inquirer, June 11, 1991.

Cortes, Joseph, 1991, A roof or a wall they carried along, to remind them of home: The Manila Times, June 9, 1991.

De Villa, Arturo, 1991, Aetas may go hungry: Daily Globe [Manila], June 10, 1991.

Empeno, Henry, 1991, Mt Pinatubo already spewing lava, (say) Aetas: The Manila Times, April 23, 1991.

Gob, Fely, 1991, Rains bring death to Pinatubo evacuees: Daily Globe[Manila], April 22, 1991.

Morella, Cecil, 1991, Volano eruption displaces Aetas: Manila Bulletin, June 11, 1991.

National Statistics Office, 1990, 1990 census of population and housing, Report No. 2-74C (Pampanga); Report No. 2-95C (Tarlac); and Report No. 2-99C (Zambales), Population by City, Municipality and Barangay: NSO, Manila.

Philippine Institute of Volcanology and Seismology (PHIVOLCS), variously dated, Daily Volcano Bulletins, Pinatubo Volcano, April 7, 1991-December 1992: Quezon City, PHIVOLCS.

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.

Sabit, J.P., Pigtain, R.C., and de la Cruz, E.G., this volume, The west-side story: Observations of the 1991 Mount Pinatubo eruptions from the west.

United Nations Disaster Relief Organization (UNDRO), 1986, Social and sociological aspects, in Disaster prevention and mitigation, v. 12: New York, United Nations.

------1987, Public information aspects, in Disaster prevention and mitigation, v. 10: New York, United Nations.

Velarde, Cherry, and Bartolome, Noel, 1991, Pinatubo erupts: Malaya [Manila], June 10, 1991.

Villanueva, Marichu A., and Dizon, Romy, 1991, Mass evacuation starts in Pinatubo: Manila Standard, June 9, 1991.

Wolfe, E.W. and Hoblitt, R.P., this volume, Overview of the eruptions.


Appendix 1. Computation of sampling size, household survey, 199192.

Appendix 2. Distribution of respondents, 1991 and 1992.

Appendix 3. Interview schedule.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey
This page is <>
Contact: Chris Newhall
Last updated 06.11.99