It began 2 April 1991, 35 years ago this month.

Hidden targets

There are many volcanoes in the world. At any one time, perhaps 20 volcanoes may be erupting. Over a full year, on average around 50 volcanoes erupt. Some of these may erupt several times – as Kilauea has done episodically in spectacular fashion, while Great Sitkin has been quietly and continuously erupting for years. Around 500 volcanoes have had known eruptions in the past 500 years, and another 1000 volcanoes are known as potentially active, in the sense that there are indications of eruptions in the past 10,000 years, e.g., a caldera! Around 30 of these are in Iceland and according to the USGS, there are 170 in the US. Who knew! Volcanoes that haven’t erupted for more than 10,000 years are generally considered extinct – unless it is Yellowstone in which case even 50,000 years is seemingly not enough for a downgrade.

Detailed monitoring is limited to volcanoes which are deemed to pose significant hazards. Extinct volcanoes are generally left alone. Potentially active volcanoes also tend to be given limited attention. Even the 500-odd volcanoes with historical eruptions may be left wanting. Most attention is focussed on a minority of most frequently active volcanoes.

In some cases we may still be too relaxed about well-monitored, active volcanoes. The recent past is not always a good indicator of what is coming: not all eruptions are the same, even when from the same volcano. For instance, Etna is amongst the most frequently erupting volcanoes on Earth, with eruptions that are notable but are not too significant. However, in 1669 Etna destroyed villages with massive lava flows which even reached the walls of Catania. Going back further, before 15,000 years ago Etna had massive explosions with ash reaching as far as the future site of Rome. And it had suffered massive land slides, luckily in the distant past. When monitoring Etna, we do so against the backdrop of  those events, not just the recent photogenic eruptions. Krakatau is another example. Anak Krakatau was regularly seen erupting, and photos of its eruptions were used as eye candy on many websites. But the sudden collapse in 2018 came as a surprise and killed 400 people. In hindsight the signs of the instability had been there for many months before. The best observed volcanoes are not necessarily the best monitored.

Volcano monitoring is now an international endeavour. Satellite imaging transcends borders. Expertise in analysing seismographs exists in many countries, but the manpower (or funding) may be lacking: it is something where countries can help each other. As has been proven in the story of Pinatubo.

The beginning

Perhaps it started 16 July 1990 with a large earthquake. The magnitude 7.8 earthquake occurred on a fault zone running through north central Luzon, the Philippines’ largest island. Steam clouds were seen from one of those unmonitored, unrecognized volcanoes: Pinatubo, and small earthquakes were felt. Catholics nuns living close to Pinatubo went to Manilla a few weeks later, 3 August,  to report these events to PHIVOLCS.  PHIVOLCS investigated and found evidence of a large landslide, but that the steam had come from the known thermal area. PHIVOLCS consulted with USGS but  there was not enough to call it volcanic activity,  The investigators concluded that the landslide had uncovered more of the thermal area but otherwise they did not suspect any changes.We still don’t know whether this was the beginning.

History

Pinatubo had attracted little attention. Luzon has over 30 known volcanoes, of which three are frequently active: Taal, Mayon, and Bulusan. Of the 30, 12 are considered extinct. Pinatubo is located in a mountainous region on the west of Luzon. There are two other volcanoes in this mountain range: Natib and Mariveles. If you have never heard of them, before the 1990’s you would not have heard about Pinatubo either. Mariveles has a 4-km caldera, but has not erupted for 4000 years. Natib has a large caldera, 6 km across, but has not erupted for perhaps 50,000 years. The region was not high on the list of volcanic hazards! It was clear that the rare eruptions here were caldera-size, but there are many places in the world where that is the case and those eruptions were long ago.

Pinatubo consisted of a set of separate domes located across an old caldera, within a densely forested region. The tallest dome was around 1700 meters high, but stood only 600 meters above the surroundings and 200 meters above the next tallest neighbouring mountain. Even the USGS described it as ‘inconspicuous’.

Those surrounding mountains were all part of Pinatubo. They were relics of a stratovolcano which had self-destructed, after which modern Pinatubo grew up on the ashes. Volcanic tuffs are widespread in the region. Although it was clear that they had come from this general direction, Pinatubo had not been identified as the source, and one researcher even described Pinatubo as ‘non-volcanic’. Only in the 1970’s was Pinatubo itself studied in more detail, as part of a site study for a potential nuclear plant. (The plant was built in the 1970’s ,about 50 km south of Pinatubo, but it was never commissioned. That is an interesting story.)

Warren Smith explored the region over a century ago and wrote in ‘The geology of Luzon’ (1913) ‘The rocks are volcanic extrusives [..] there are no active volcanoes along this line, and the old volcanic stocks are pretty well eroded.’ He added a footnote ‘It was reported by Mr. Snyder of the Bureau of Lands that smoke was seen issuing from the top of one of these peaks.’ Sadly, he does not say what he meant with ‘these peaks’!

In 1983, Wolfe & Self (Geophysical Monograph Series. The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands: Part 2) summarised what was known about the mountain:

Mount Pinatubo. The highest volcano on the Bataan Lineament is a composite structure of calc-alkaline affinity rising to 1745 m. It is located at the point where the Iba fault zone, a graben, intersects the Bataan Lineament. There have been no historic eruptions of Mount Pinatubo, but there are reportedly two roaring vents on the summit and sulfur is actively depositing. Dates of 635, 2330, and 8050 years B.P. have been obtained by the I4C method [Wolfe, 1981] from material within mudflow (lahar) deposits. Since there has been no thorough study of this volcano, there may have been younger eruptions. One of the most recent eruptions resulted in a voluminous plagioclase-rich crystal tuff (ignimbrite), which filled stream valleys southwest toward San Marcelino, northwest into the Iba graben and north-easterly toward Tarlac and Angeles.

The Bataan lineament is the series of volcanoes on the west side of Luzon, stretching from Taal northwestwards. Pinatubo is 100 km beyond the end of this lineament but the authors state that it is related. The ‘roaring vents’, presumably the geothermal area, seem a bit of a concern. There is no source given for this information, but it would be unexpected from a volcano that had not erupted for centuries! Something similar in, say, Auckland would set alarm bells ringing. It is worth noting that the local Aeta people did mention small explosions having occurred during previous generations.

Pinatubo was known since the 1960’s to have a geothermal field, on the northwest flank within a 150-m wide depression. The field had weak steaming with some sulfur depositions and hot springs. Plans had even been made for a geothermal energy facility. A few exploratory wells were drilled in 1989 and 1990. Results were presented in 1990, by D. Michels who stated ‘A fumarole near the summit emits gases with magmatic characteristics’ (https://www.osti.gov/biblio/888499). In view of later developments, this may have been telling! The high acidity of the fluids made them unsuitable for exploitation, being too corrosive. They provide a geological cross section of the mountain:

The wells revealed a temperature of 330C at 1 km depth, seemingly higher than expected. They also give the cryptic statement of ‘increased boiling due to rock heat that increased during later stages of testing’. (The testing was done between 18 Dec 1989 and 26 Jan 1990.) They also found excess CO2, indicative of magmatic fluids. This was all interpreted in terms of the eruption of 500 years, with high temperatures attributed to dry rock which had cooled only slowly. In hindsight, was this instead an early indication of a re-awakening of Pinatubo?

To the east of Pinatubo are parts of an old, 4-km wide caldera wall. The remnants of ancient Pinatubo are outside of this wall. This old Pinatubo existed 1 million years ago, and includes several of the nearby peaks. After a long phase of quiescence, self-destruction occurred more than 35,000 years ago, in a VEI-7 eruption which formed the old caldera. It is known as the Inararo eruption. After that, a new Pinatubo grew up in the same place. It had a number of ‘eruptive episodes’ (4 or 5?) separated by quiet phases. The last activity before the 20^th century was around 500-600 years ago. it left deep ash beds in various places. It should be noted that it can be difficult to distinguish old pyroclastics flow deposits from those of lahars, so it is not always clear what is caused by what, and this can leave the size of prehistoric eruptions unclear.

Awakening

This much was known before the initial warnings in 1990.

Nothing happened for the next 7 months. Starting on 15 March 1991, local residents began to feel earthquakes. The steaming from the thermal area had not changed and no rockfalls had been noted. It is not clear whether this information made it to PHIVOLCS. The local area was inhabited by Aetas, from the indigenous population, who perhaps counldn’t easily pick up the phone to Manilla! This information later came through members of the Negrito People’s Alliance of Zambales (LAKAS).

The earthquakes increased in the morning of April 2, and large phreatic explosions started at 4pm that day, lasting into the evening. A strong sulfur smell was noticed.  Again the nuns went to Manilla to  report to PHIVOLCS,  two days later. The PHIVOLCS Director immediately sent an airplane to carry out an aerial survey. It found nine thermal vents along a line close to the original thermal area, and a line of deep craters which had formed in the explosions.

The investigators still did not think this was unusual. It was just steam activity from a known geothermal area. The PHIVOLCS Director, Ray Punongbayan, disagreed and felt it should be investigated further. PHIVOLCS installed the first seismograph already in the afternoon of the next day. It was located downstream along the Maraunot river and radio-linked to their headquarters. The result immediately moved Pinatubo way up their list of priorities. 223 earthquakes were detected in the first 24 hours. Four further stations were installed during the following week, but now with data written to a floppy disk (remember those?) to be retrieved manually. Daily volcanic updates were issued. The seismic activity decreased somewhat during April but did not cease. Two of the vents remained strong and even carried some ash.

Assistance

This was when the USGS became involved. It had a program to help developing countries with exactly this. The program was funded through USAID (recently abolished), and was called the Volcano Disaster Assistance Program – VDAP in short. It kept a supply of equipment ready for just such a crisis. The Clark Air Base was located close to Pinatubo, and they asked VDAP to come with the next available flight saying “you have $20,000, and do not feel compelled to spend it all.” On arrival, the first thing they saw was an advert for the TV series ‘Last days of Pompei’!

By the end of April, a ‘Pinatubo Volcano Observatory’ (PVO) had been set up in empty bed rooms at Clark Air Base, making use of the on-site infrastructure. The air base had good reason to assist: built on deposits from a previous Pinatubo eruption, it could consider itself in a potential line of fire!

A variety of instrumentation had to be installed, in a mountainous, sometimes forested and inaccessible area. The US army knew the area well as they used it for training exercises. They provided helicopters for transporting seismographs, tiltmeters and one gas spectrometer, linked to the PVO office by radio. In these ancient times, the most important piece of equipment was possibly the printer. Some of the lessons learned from the St Helens eruption came in handy, such as,  don’t station your observatory people within reach of the eruption! This became an issue for both teams: PHIVOLCS observers stationed at Poontabo, 20 km northwest of Pinatubo, and USGS at Clark Airforce Base was a similar distance east of the mountain. In the end, both of these observatories had to be abandoned.

The PVO was a collaboration between PHIVOLCS and USGS/VDAP, helped by the fact that the two lead scientists knew each other well. That was important: there was a danger that warnings given by a US institution would be perceived as aimed solely on the Clark Air Base. The local population might receive them with delay or might ignore them: PHIVOLCS would know best how to reach them. Vice versa, The US army would be more likely to heed warnings by a US institution! (In the end, they almost didn’t.)

Clark Air Base was enormous, extending over some 40km² and located as close as 15 km to Pinatubo. The main housing of the site was in the hills east of Pinatubo, even closer to the mountain. The base was used during the second world war (mainly by the Japanese), the Korean war and the Vietnam war. But Pinatubo was a war it lost: after the eruption the US lease was not renewed and the base remained empty for years.

Warnings

The first task of the combined group was to investigate the volcanic hazard map. That was easy – there wasn’t one. Instead, the group drew out the largest extent where deposits from the previous eruptions were found (using a set of new radiocarbon age determinations) and designated that as the warning area. This turned out (in hindsight) to be bot remarkably accurate and a major underrestimate. Evacuation plans were made for the quarter of a million people within the hazard area, of whom a third later actually evacuated. The volcanologists found that the flat area of the air base (ideal for landing strips) had been created by Pinatubo pyroclastics, but themselves (unwisely) in the hazard area.

The seismographs showed that the mountain was rumbling but no clear pattern emerged during the next weeks. There was however a tenfold increase in SO2 emissions, a warning sign of the presence of fresh magma. The earthquakes were centred about 5 km NNW of the mountain, in a region with young-looking faults, and were at a depth of around 5km. This was not near the steaming vents: that area also showed earthquake activity but shallow and much less intense. The large extent of the earthquake region indicated this was more than a minor plumbing issue.

An alert system was created involving the local authorities. The system of numeric warnings worked very well: it was easy to comprehend for the public, who were often unfamiliar with volcanic threats and technical language such as ‘tephra’ ‘magma’, etc. Nowadays, different countries use different system for warning levels. PHIVOLCS still use this same system. New Zealand uses 6 levels. The US, perhaps more visually oriented, uses colour coding. But the basic purpose remains the same: to reduce the information to a single word. The alert level was raised to 2 (straight from zero) on 13 May. This was the first official forecast of a chance of an eruption.

A video made by Maurice Krafft on volcanic hazards became very helpful in communicating the dangers.Nowadays we know well how effective videos are in convincing people, but this was still experimental in the 1980’s. (In a sad coincidence, Maurice and Catherine Krafft died on 3 June that year, in a pyroclastic flow on Mount Uzen.)

Run-up

The steaming intensified during this period, and contained increasing amounts of ash, turning the white steam brown. On 28 May, the ash cloud reached 5 km high. The steam acidified the Maraunot river which reached a pH of 2.45 in early April, having been pH neutral before. Plants withered and aquatic life died.

After this, the situation began to change, with more signs of intent. The earthquake activity shifted to the region of the steaming vents, and grew more shallow. Harmonic tremor was seen. At the same time, the SO2 emission declined tenfold, a sign that a plumbing conduit had become inactive or blocked. On June 3, an evacuation was ordered for 20,000 people within 10 km of the summit. The alert level was raised to 3 on 5 June.

On 6 June, after an initial quiet period, a shallow earthquake swarm in the venting area, combined with inflationary tilt on the east flank, heralded that an eruption was becoming likely. At this time PHILVOCS moved their seismographs to Poonbato, further from Pinatubo, as they were becoming dangerous to reach. The PVO office was moved to the far side of the Clark Air Base almost 10 km further from Pinatubo – in hindsight, still too close.

A lava dome began to form on 7 June, just north of the venting area. It grew quickly, leading to an alert level increase to 4 on that day. The evacuation radius was now set at 18 km, covering 120,000 people (later increased to 25 km). The evacuation area also included the air base, but the army was not yet convinced and stayed put. On 9 June, ash clouds were seen rolling down the Maraunot river. This was interpreted as the onset of an eruption, and the alert level was raised to 5 – perhaps a little premature. Only on 10 June did the Clark Air Base evacuate. That was just in time: the first real eruption started June 12. (The 20-km high ash cloud was called the ‘appetizer” by the volcanologists,)

On June 15, 2am the eruption went for the cataclysmic main course. Pyroclastic flows nearly reached the Air Base, stones fell like rain (which also fell) and the volcanologists in their more distant office had to shield deep inside the structure not entirely sure of their own survival. Ash blanketed a region far outside the perceived danger zone. In one mountain valley, the ash was later measured at 200 meters thick.

In the end, 58,000 people evacuated, many ending up in refugee centres. The PHIVOLCS/USGS warnings had been spectacularly effective. Without them, the death toll would have been in the tens of thousands. Now, ‘only’ 300 people died in the eruption itself, and 100 more died later in lahars. (The oft-quoted death toll of 847 includes deaths that occurred in the evacuation centres, from dirty water and from measles to which the Aeta had little resistance.)

The forecasts of the areas at risk of pyroclastic flows and lahars had been accurate. But a major factor had been missed: the area of ashfall extended much further out, and combined with the heavy rain from a typhoon that hit during the major eruption, caused roofs to collapse in cities and places well beyond the danger zone. Even the US navy base to which the Clark Air Base evacuated was badly affected – the evacuees should have gone further.

There are many vivid descriptions of this eruption, still the largest eruption since Katma in 1912.

It seems amazing, in view of the damage and area covered, that this was ‘only’ a low VEI-6! A VEI-7 would have obliterated the land out to a three times larger distance, affecting millions.

Afterthoughts

The common thread in discussions at the time involved uncertainty. An eruption is forecast, not scheduled, and the forecast comes with a probability. But the affected people can’t handle probabilities. They won’t go until there is certainty. (Even the air base was not fully evacuated until they saw the volcanologists panic.) And if an eruption does not happen – there is always a chance – the people won’t listen a second time. It is a seemingly unsolvable conundrum.

How about the VDAP program? The outcome was a great success. But the role of PHIVOLCS in the reports seemingly diminished over time as the volcanologists focussed more and more on the impending eruption. There were separate activities: most detailed observations of eruption plumes came from PHIVOLCS, while seismic modeling and gas monitoring were done by USGS. Alert levels appear to have been decided by USGS. It feels like the role of PHIVOLCS at leater times the development was a bit underplayed. But it may have been different at the time: this impression comes from reading the reports, each written for a different audience.

I do not know whether VDAP will continue, at a time when the USAID program seems largely terminated. But I hope it will. There are 1500 volcanoes in the world that could potentially erupt. The major eruptions seem to come from volcanoes that had been quiescent for a long time: Pinatubo, El Chicon, Hudson, Katmai, Krakatau, Tambora. The next one may well have been missed from the monitoring programs. An organization that can step in at short notice to provide help where needed is sorely needed. At Pinatubo, it helped prevent a volcanic disaster. It can do so again.

As one of the scientists later noted: “What you think is the worst case, can actually happen“.

Albert, April 2026

Sources used for text and images:

https://archive.org/details/InthePathofaKillerVolcano/NOVA.S20E05.In.the.Path.of.a.Killer.Volcano.1993.VHSRip.AAC2.0.x264-rattera.mp4

https://www.usgs.gov/news/featured-story/remembering-mount-pinatubo-25-years-ago-mitigating-a-crisis

https://pubs.usgs.gov/pinatubo/sabit/

Other reading:

https://earthjournalism.net/stories/eruption-lahar-and-resilience-the-aftermath-of-mt-pinatubo-eruption-in-the-philippines

https://www.volcanocafe.org/when-pinatubo-turned-the-tide/

And of course our very own VDAP April-1 spoof of 2017:

https://www.volcanocafe.org/the-usgs-volcano-observatories-can-we-save-vdap/