ANY day now, a gargantuan wave could sweep westwards across the Atlantic towards the coast of North America. A mighty wall of water 50 metres high would hit the Caribbean islands, Florida and the rest of the eastern seaboard, surging up to 20 kilometres inland and engulfing everything in its path. If you thought the tsunamis that periodically terrorise the Pacific Ocean were big, consider this: the Atlantic wave will be five times bigger. It will start its journey 6000 kilometres away, when half an island crashes into the sea.
Simon Day of the Benfield Greig Hazard Research Centre at University College London has discovered that a huge chunk of La Palma, the most volcanically active island in the Canaries, is now unstable. 鈥淚f the flank of the volcano slides into the ocean, the mass of moving rock will push the water in front of it, creating a tsunami wave far larger than any seen in history,鈥� says Day. 鈥淭he wave would then spread out across the Atlantic at the speed of a jet airliner until it strikes coastal areas all around the North Atlantic.鈥�
The idea that collapsing islands can cause giant tsunamis dates back to the 1960s when Jim Moore, a geologist with the US Geological Survey in Menlo Park, California, was studying early bathymetric maps of the sea floor around the Hawaiian islands. He spotted what he thought were huge chunks of volcanic rock strewn across the seabed up to 140 kilometres from the nearest islands. Moore believed these were the debris from titanic landslides.
Advertisement
It was not until the early 1990s that Moore鈥檚 suggestion began to be taken seriously. By then, higher-resolution maps of the sea floor showed evidence for dozens of landslides in the Hawaiian islands. And sea-floor surveys near R茅union in the Indian Ocean, the Marquesas in the western Pacific, Tristan da Cunha in the South Atlantic and El Hierro in the North Atlantic showed that collapses of oceanic volcanoes into the sea occur worldwide.
Moore also suggested that the Hawaiian landslides generated giant waves. He attributed marine deposits in the Hawaiian islands that lie up to 375 metres above sea level to the action of tsunamis, though this is still highly controversial (快猫短视频 supplement, 7 August 1999, p 4). But most scientists now agree that island collapses around the world would inevitably have caused gigantic tsunamis. For instance, giant wave deposits found in the Bahamas coincide with a past collapse in the Canaries (see 鈥淚sland catastrophe鈥�).
Day began his work on the Canary Islands in 1994, at the invitation of Spanish geologist Juan Carlos Carracedo from the Volcanological Station of the Canary Islands in Tenerife. Carracedo had already found evidence for at least one collapse on the neighbouring island of La Palma, which had apparently come from the extinct Taburiente volcano in the north of the island. He wanted to know if there was any risk of a future collapse of the Cumbre Vieja, the active volcano that forms the southern half of La Palma.
The best way to predict the volcano鈥檚 future, Day believed, was by studying its past. Over the next two years, he carefully surveyed the summit area of the Cumbre Vieja, identifying dozens of volcanic vents that had been formed by successive eruptions over the past hundred thousand years. From several of these vents, Day collected samples of lava and charcoal from trees burnt by the molten rock. His colleague, Herve Guillou of the Laboratory of Climatic and Environmental Sciences in Gif-sur-Yvette in France, then used potassium-argon or carbon dating to find out when the vents formed.
By 1997, Day and Carracedo had enough data to assemble a detailed geological map of the Cumbre Vieja. They found that most of the vents were organised in three rift zones, laid out in a three-pointed 鈥淢ercedes star鈥� configuration, facing south, north-east and north-west. The north-east and north-west rift zones have become inactive in the past few thousand years, but the remaining southern rift zone has extended northwards, gradually bisecting the volcano.
In a paper published late last year, Day and his colleagues suggested that this structure could lead to a mighty landslide. Day believes that the western flank of the volcano is becoming gradually detached from the eastern half. What鈥檚 more, he thinks the western flank is subtly altering in shape, making it easier for magma to break through to the surface. During the three most recent summit eruptions in 1585, 1712 and 1949, vents opened up on the western flank of the volcano, while none appeared on the eastern flank.
Day has other evidence to support his theory. During the 1949 eruption, the island was racked by two large earthquakes. A day later, Spanish seismologist Juan Bonelli Rubio discovered a crack on the summit of the Cumbre Vieja. It could not be a new volcanic vent, as no lava or steam was coming from the fissure. Bonelli Rubio suggested that it was the fault responsible for the earthquakes.
During 1995, while mapping the summit region, Day noticed something unexpected about the 1949 fault: rather than opening up horizontally the way vent fissures do when magma rises to the surface, the land surface on the western side had slipped down 4 metres relative to the eastern side. 鈥淚t is the first time we鈥檝e seen a fault like this on the volcano,鈥� says Day. For him, the appearance of the fault in 1949 proved that the whole of the western flank of the Cumbre Vieja is poised to collapse into the Atlantic Ocean.
It鈥檚 hard to imagine what would happen if half a trillion tonnes of rock slid into the sea. But Hermann Fritz, a PhD student at the Swiss Federal Institute of Technology in Zurich, has spent several years modelling how landslides generate waves when they fall into water. Earlier this year, he constructed a lab model of the western flank of the Cumbre Vieja in a wave tank. The model is an elongated wedge-shaped block resting on a 10-degree slope with the tip of the block lying just under the water. When the block is released, it slides down the slope generating a wave, which is recorded by a high-speed camera.
Fritz found that the sliding block generated a long, shallow, fast-moving wave 鈥� the classic profile of a tsunami. Scaling up 10,000 times, the model predicts that in real life the crest of the wave generated by the collapse of the western flank of the Cumbre Vieja would initially be a staggering 650 metres above normal sea level, more than enough to submerge the tallest building in the world. Fritz admits that there is a big size difference between his model and the real tsunami, but he has no doubt that the dimensions of the wave are in the right ballpark.
Where would the wave go? Because the unstable flank of the Cumbre Vieja faces west, if it collapsed the resulting tsunami would race across the Atlantic towards North America. It wouldn鈥檛 be half a kilometre high when it arrived, though. Waves that radiate out from a single point, like ripples on a pond, decrease in height as they travel. But it would still be a monster. If the La Palma collapse produced a single tsunami, it would be 40 to 50 metres high when it reached the American coast. 鈥淭hat mass of water moving at that speed would remove buildings, trees, roads, everything,鈥� says Day.
Given the evidence that the Cumbre Vieja is poised to collapse and the catastrophic consequences of such an event, Day had to find out when the volcano might let go. First, he wanted to know if the fault that appeared in 1949 is still moving. Over the past few years, Jane Moss, a student at the College of Higher Education in Cheltenham, Gloucestershire, has used the satellite-based Global Positioning System (GPS) to monitor the positions of 20 markers around the volcano on both sides of the fault. Late last year, she concluded that the fault has completely stopped moving 鈥� at least for now.
If nothing is pushing the fault today, why did it form in the first place? Day believes that water trapped in the volcano is the key. On the face of it, water is not an obvious factor: the volcanic rock of La Palma is so permeable that when rain falls it quickly soaks through, leaving the surface as dry as a bone. But inside the volcano, impermeable dykes 鈥� columns of magma that fed previous eruptions 鈥� cut through the permeable volcanic rubble. These dykes act like a series of dams, trapping rainwater (see Diagram).FIG-mg22594402.JPG
Day realised that when magma rises towards the surface through wet rock, the water in the pore spaces within the rock is heated and expands. Because liquid water is incompressible, small increases in water temperature can significantly increase the water pressure in a confined space. Working with Derek Elsworth of Pennsylvania State University, Day performed some calculations to see if the pressure generated by heating water would be enough to make the rock fracture. Sure enough, relatively modest warming gave huge increases in water pressure. At 1 kilometre below the surface, for instance, a temperature increase of as little as 15 掳C would increase water pressure from around 160 atmospheres to 400 atmospheres 鈥� more than enough to split the rock and cause a collapse.
Gary McMurtry at the University of Hawaii in Honolulu agrees that water is vital, but disagrees with Day and Elsworth on exactly how water causes collapses. McMurtry thinks collapses in the Hawaiian islands are triggered when rainwater penetrates shallow magma chambers and causes so-called phreatomagmatic eruptions 鈥� large, violent explosions that occur when hot magma comes into contact with water and turns into a pressurised mixture of molten rock and steam.
鈥淓verywhere we look on the sea floor around the Hawaiian islands, we see volcanic ash. That鈥檚 telling us that there has been explosive volcanism,鈥� says McMurtry. He believes the force of these explosions, rather than the expansion of liquid water, causes volcanic island collapses. However, Day argues that this model does not apply to the Canaries, where magma reservoirs lurk deep in the Earth鈥檚 crust, out of the reach of rainwater. He points out that on La Palma in 1949, no explosive activity was recorded on the volcano when the fault line appeared.
Whether the expansion of liquid water or steam explosions are the driving force, Day and McMurtry agree that the combination of water and magma triggers collapse. Day says this leads to an important prediction: if the Cumbre Vieja collapses, it is likely to happen during a future eruption. 鈥淚t is reassuring that the Cumbre Vieja isn鈥檛 going to collapse spontaneously,鈥� he says. But it鈥檚 also worrying. After all, the volcano erupts once every few decades.
Still, not every eruption happens in the right place to cause a collapse 鈥� or even to make the fault more unstable. The last one was in 1971, but that took place on the southern tip of the island, well away from the unstable summit ridge. According to Day, how dangerous an eruption is depends on how long it lasts, where on the volcano it occurs and how much magma is involved. 鈥淚t鈥檚 unlikely that the volcano will collapse during the next eruption,鈥� he says, and it may take many more before the fault finally gives. But when it does give, watch out Florida.