Rumble from roll of the dice

Volcanology should be scientific not macho, says Clive Oppenheimer.

In 1988, a long-dormant volcano called Galeras began to rumble. This rang alarm bells in Colombia as it was only three years since an eruption of the Nevado del Ruiz volcano, some 500km away, had claimed the lives of 23,000 Colombians. Galeras rises above the regional capital, Pasto, a town of 300,000 inhabitants. The 1990s had been designated by the United Nations as the International Decade for Natural Disaster Reduction, and Galeras was one of 15 volcanoes worldwide accorded "decade volcano" status. The intention was that these volcanoes would become the focus of risk-mitigation efforts.

Concerned volcanologists arrived in Pasto after 1988 and worked alongside Colombian colleagues to map the volcano, interpret its past deposits, instal and develop a monitoring network and build relations with the authorities and the population. In January 1993, their joint efforts culminated in an international workshop at Pasto aiming to review the understanding of the volcano and promote further collaborations between Colombian and foreign scientists.

After several days of talk, on January 13, the participants split into groups to take part in field excursions on and around the volcano. The volcano monitoring experts set off for the summit crater of Galeras to try out geophysical and geochemical surveillance techniques that could probe the inner state of the volcano. Some five hours later, the volcano exploded. Six of the scientists and three hikers who had followed them out of curiosity were killed. Several other scientists were injured. Geologically speaking, the eruption was unexceptional, but its human consequences are still being mourned.

These two books focus on the events leading up to the disaster and debate its inevitability. They present very different sides of the science of volcanology and some of its practitioners. One is written by a volcanologist, Stanley Williams, the other by a geosciences graduate-turned-writer, Victoria Bruce. Although the books deal with the same set of facts about Galeras, they reach opposite conclusions concerning the extent to which the disaster might have been avoided. Both are based on careful research and extensive interviews with participants at the workshop, survivors of the explosion and relatives of those who died. Both are crisply and provocatively written, but Williams presents a far more personal account, for it was he who led the trip into the crater. He was seriously wounded and he discusses the consequences of his injuries, including the psychological effects, which he admits led him to go along with a media story, much touted in the aftermath of the tragedy, that he was the sole survivor.

The thrust of Bruce's book is to question why Williams led people up to the crater. Her argument hinges on interpretation of curious earthquake traces that were recorded by the Galeras volcano observatory seismographs before the eruption. These particular tremors are known as long-period events because of their low frequency. At Galeras, they are called tornillos because of their tapered, screw-like appearance in seismograms. Bruce reports the findings of a US Geological Survey seismologist, Bernard Chouet, who believes that tornillos presaged the 1993 eruption. They had preceded an explosion in July 1992, and they have been seen in association with a number of eruptions since January 1993. They are now thought to be related to escalating pressure within the volcanic plumbing system.

The question is, at Galeras in January 1993, could the volcanologists at the observatory and those leading the workshop and field trips have been expected to understand the significance of the tornillos ? Bruce says yes, and comes up with various lines of evidence, including a prominent abstract from the workshop proceedings, which likened the prevailing seismic and gas observations to those before the July 1992 eruption. Williams, not unnaturally, feels that accusations about a "smattering of tornillos " are easily made with the hindsight of several subsequent eruptions to confirm the trend. I expect many volcanologists would indeed forgive him for failing to decode the seismograms. But, interestingly, he rests his case by claiming that, at the time, "based on all available evidence, the consensus at the observatory was that Galeras was safe". Does this statement intentionally set out to deflect or diffuse any possible blame to an organisation (the observatory)?

Bruce's other line of attack is to expose the field party's lack of safety precautions. Williams states in his introduction that "one should always prepare for the worst on a volcano", but the truth is that little or no emergency planning was carried out before descending into the crater. She later quotes one of the other survivors as saying: "We were certainly not prepared. We were kind of dilettantesI We were all too cavalier." Bruce makes much of the lack of hard hats and heat-protective clothing among the participants and the limited radio contact they had with the observatory (the field team had only one radio between them). Also, she implies that Williams allowed the team to dawdle in the crater.

Williams counters by explaining that he felt it inappropriate to tell his colleagues what to wear in the field and that he had substantially trimmed the number of participants on the trip precisely to enhance the group's mobility at the summit. But he also describes a macho side to volcanology:

"Our profession is not for those who prize a secure, sedentary life," he claims. The implication is that the best volcanologists work on volcanoes; they are prepared to look an eruption vent in the eye and to inhale its noxious fumes without a mask. These views will dismay and be disowned by many of his colleagues.

This does not mean that volcanologists are always sober about the risks. The fact is that 29 volcanologists - out of a worldwide community numbering probably 1,000 - have been killed on volcanoes since 1975. Many others have been on volcanoes between eruptions that, with different rolls of the dice, could have caught them out. There are probably more near-misses than are acknowledged. A volcano is never a zero-risk environment, but the observations essential to understand how volcanoes work, and thereby help to protect populations, could not be obtained if nobody accepted the risks of installing seismometers and other sensors in hazardous areas. In this context, volcanology is not so different from some other risky scientific endeavours, such as polar exploration or working with pathogens. Williams says as much but avoids the more substantive point: that an awakening volcano that erupted six months previously in a way that could well have killed anyone near the crater has to be considered highly suspect. How much did the perceived scientific benefits of going to the crater outweigh the risks?

As a back-of-an-envelope calculation, consider spending five hours atop a volcano that erupts, on average, twice a year, with a 50 per cent chance of survival if it happens, and there is a greater than 1 in 2,000 chance of being killed. Put 20 people up there and the risk of someone being killed increases 20-fold. The kinds of eruptions that can hurt people within a few hundred metres of an explosive vent are so trivial as far as the volcano is concerned - mere throat clearings - that they can probably occur without detectable precursors. While attempts at real-time forecasting of such events using seismic and other data remain worthwhile, the first-order approach to mitigating risks for field scientists should be a careful assessment of the probability of an eruption.

Both books are gripping - though the Hammond Innes-style jacket and Boy's Own subtitle of the Williams book will not be to everyone's taste (perhaps a hint of the movie that is apparently already under discussion). Williams has the edge in having survived the eruption and known the victims. As an experienced volcanologist, he is also able to convey the science more accurately. On the other hand, Bruce stands outside the glass house and can afford to be franker in her criticism. She also relates in more detail the events surrounding the 1985 eruption of Nevado del Ruiz, which is a necessary part of the background to the Galeras tragedy. It is a pity, though, that neither author examines in more depth issues such as group-versus-individual responsibility, why volcanologists work on volcanoes, the calculation of risk and why scientists stop talking to one another. (Williams and Chouet - the seismologist who knew the meaning of the tornillos - were not on good terms.)

But in the final analysis, it is hard to deny that fate played its hand: Chouet was kept from attending the Pasto meeting because of a ban on travel to Colombia for US government employees; a scheduled power cut inclined the meeting organisers to delay the field excursions by one day; and Williams's group spent longer in the crater than expected. It is an agonising reminder of the thin line between a near-miss, which is laughed off afterwards, and disaster. Precisely because we tend to think that it will not happen to us, field volcanologists should take risk assessment as seriously as it is taken in the laboratory. If these two books remind us of the importance of vigilance over complacency, then something more than media froth could come of them. But they will do the science of volcanology a great disservice if they encourage its exponents to become regarded as foolish thrill seekers.

Clive Oppenheimer is lecturer in geography, University of Cambridge. He is monitoring a volcano in Nicaragua.