TOWNS and villages downstream of Montserrat’s Soufrière Hills volcano have been abandoned for almost a decade now. Of around 12,000 people who once lived on this Caribbean island, two-thirds left. The others live crammed into the island’s northern half, while the south, including the former capital, Plymouth, remains an exclusion zone where nobody may stay overnight.
By day, however, the exclusion zone is far from deserted. On many of the hills and valleys alongside the volcano, cows and donkeys still graze. Farmers have permission to visit their property, and many do, some in the hope of being able to resettle there one day.
But in July 2003, even this tenuous existence was thrown into disarray, when the volcano staged its largest eruption since it became active in 1995. This led to the unexpected deaths of more than 40 animals, and, as it turns out, an interesting geological discovery.
Advertisement
Seismometers began to react on the morning of 12 July. Soon the small Montserrat Volcano Observatory was crowded with scientists anxiously scanning the instrument panels and monitors. Suddenly one of the seismometers went off-line – the sensor had taken a direct hit from a pyroclastic flow made up of ash, blocks of rock and superheated steam.
The drama continued for 18 hours. At one point, the volcano’s collapsing dome was losing material so fast that GPS receivers recorded the ground rising spontaneously as it was relieved of the pressure from around 50 million cubic metres of rock in 40 minutes. The flows reached the ocean, and port authorities in neighbouring Guadeloupe radioed to say that a small tsunami had smashed 15 fishing boats.
As soon as it was safe volcanologists Marie Edmonds and Richard Herd drove out to survey the damage. It was the morning of 14 July. The dome at the top of the volcano had entirely collapsed, dumping around 200 million cubic metres of pyroclastic material into just one spot, the Tar river valley. But they were deeply puzzled.
“A splurge of rock, gas and steam spread from the shoreline, fanning out to high slopes that had escaped the original flow”
Near the coast, the pyroclastic flows seemed not to have come from the volcano, but from the river delta, where the Tar river runs into the sea. The closer they got to this spot, the thicker the pyroclastic deposits became. What is more, trees facing the ocean had rocks embedded in them only on their coastal side – the side away from the volcano. And these flows travelled up to 3.5 kilometres, reaching valleys farmers had been told were safe for animals. As it was, 40 to 50 animals were dead.
At first the pair blamed the damage on turbulent pyroclastic flow, which sometimes shoots off sideways or even back up the hill when it hits an object. In 1902, when Mount Pelée erupted on the nearby island of Martinique, one stream of pyroclastic flow broke off and tore sideways along the coast. But at Soufrière Hills, there was no obstacle to produce the turbulence, and no way it could produce such devastation. “We needed a mechanism by which a surge could flow back on land at up to 120 degrees away from its original pyroclastic flow direction, and more importantly, have enough energy to reach 300 metres above sea level,” Edmonds says.
Perhaps the flat river delta had caused the pyroclastic flow to back up, like the splash-back in a shallow pool at the base of a waterfall. But then they found half-metre chunks of rapidly cooled volcanic rocks as far as 2 to 3 kilometres away from the parent pyroclastic flow. Much too far away for mere splash-back.
At this point, Edmonds and Herd began to consider an interesting possibility: could pyroclastic material hitting the seawater have caused a secondary eruption? It is well known that pyroclastic material can interact with seawater to produce what are known as hydrovolcanic eruptions. “Krakatoa, Tambora, and Santorini all had hydrovolcanic components to their explosive activity when water flooded their magma chambers. It is a huge source of energy,” Edmonds says. Fine ash has a large surface area and can rapidly dump its heat in the water, vaporising it explosively.
It is likely such an explosion did indeed occur at the mouth of the Tar river. Unlike shield volcanoes, such as Hualalai on the island of Hawaii, where lava rolls downhill little faster than walking pace, the lava on Montserrat is so viscous it cannot escape easily. Instead it builds up into a sticky dome. Over time, the dome becomes unstable. When it finally blows, the plug of lava shoots out of the top, releasing pyroclastic material that flows at more than 200 kilometres an hour and reaches temperatures of over 300 °C.
By July 2003, the dome on Soufrière Hills was a kilometre high and contained around 100 million cubic metres of hard lava. When it blew, the resulting pyroclastic flow was funnelled into the Tar river valley. No one saw what happened next. But it must have been something like this. Where hot rock met the ocean at the mouth of the river there must have been lots of steam and small explosions. The steam probably smashed the rock apart exposing the burning-hot surface to the water, creating yet more steam. By midnight on 13 July, the volcano was blowing fine ash too, which explodes even more violently on contact with water. What followed was almost certainly a huge mushroom-shaped explosion.
At its base there would have been a surge of pyroclastic material, gas and steam spreading out from the shoreline, fanning out to high slopes on parts of the coastline that had escaped the original pyroclastic flow. Indeed the legs of the dead animals were marked by this surge, although they probably died from inhaling the hot gas.
Last year Edmonds and Herd told the Western Pacific Geophysics meeting in Hawaii about their discoveries. Both are now based there with the US Geological Survey and their report will appear in Geology next month (vol 33, p 245). “This may be the first time such an eruption is being described as coming back on non-impacted areas,” says John Ewert of the USGS in Vancouver, Washington state.
The first time a base surge was ever linked to a hydrovolcanic eruption was in 1965, when lake water spilt into the caldera of the Taal volcano in the Philippines. But the idea that this can happen in a river delta is new. “Usually you get only small-scale interaction when flows hit the sea – local boiling of seawater, black jets of tephra shooting up into the air,” Edmonds says. But in this case there was a large explosion, which could only have happened if a large pyroclastic flow hit the sea, generating an explosion column and a base surge that thundered back onto the land. “This has never been documented previously.”
This time on Montserrat, farm animals were the only casualties. But the research has implications for the people near the exclusion zones of volcanoes in heavily populated areas. So far, authorities are not taking this phenomenon into account. “The hazard implication is that the pyroclastic surges do not necessarily have to come from the volcano,” warns Edmonds.
The problem is that no one knows the probability of a dome collapse causing a base surge. On Montserrat, the base surge was the result of an enormously large dome collapse coupled with pyroclastic flow being funnelled into one river valley. But similar geography exists at many dome collapse volcanoes near coastlines. One example is Vesuvius, which, being active and close to inhabited areas, is probably the most dangerous volcano in the world. More than 550,000 people live in the most dangerous “red zone” between the volcano and the sea. This area can already be evacuated in around seven days, but 15,000 people are expected to take a €30,000 deal to leave the danger zone in the next 15 years. Local authorities are also considering converting residences in the zone to bed and breakfast accommodation: tourists can rebook their holidays far more easily than residents can move out of their homes.
“More than 1 million people would be at risk if Vesuvius went the way of the Soufrière Hills”
Volcanologist Franco Barberi of the University of Rome says the risk analyses for Vesuvius do not even consider the possibility of a base surge from the ocean. But more than 1 million people live just beyond the red zone, and they would all be at risk if Vesuvius went the way of Soufrière Hills.
All over the world, populations near coastal volcanoes are growing, Ewert points out. “We create risk as we put ourselves and our things in what have always been hazardous areas,” he says. As well as the Caribbean islands, Indonesia, the Philippines and parts of New Zealand are all areas in which there are substantial numbers of people living near coastal volcanoes.
Even without knowing the exact probability of a base surge, Ewert says authorities should already be drawing larger buffer zones round volcanoes where valleys reach the sea. It is a clear message: wake up to the risks of pyroclastic base surges, or go the way of the cows.
