Excerpt from Silenced Rivers: The Ecology and Politics of Large Dams,
by Patrick McCully, Zed Books, London, 1996
Every dam site has unique geological characteristics. Gaining a thorough understanding of these characteristics is expensive and time–consuming: millions of dollars may have been spent on a geological survey before it finds that a site is unsuitable for a dam. It is therefore normal for dams to be designed with only a partial knowledge of local site conditions – the builders just have to hope that they will not find any unstable formations which will fail to support their foundations or cause the roofs of their tunnels to come crashing down. More than three–quarters of 49 projects assessed in a 1990 World Bank study of hydropower construction costs were found to have experienced unexpected geological problems of some kind. The study concluded that for hydrodams "the absence of geological problems should be treated as the exception rather than the norm."
It is well established (although little known by the general public) that large dams can trigger earthquakes. The first observation of possible reservoir–induced seismicity (RIS) was noted for Algeria’s Quedd Fodda Dam in 1932; the first extensive study of the correlation between increased earthquake activity and variations in reservoir depth was made in the 1940s for Hoover Dam. Today there is evidence linking earth tremors and reservoir operation for more than 70 dams. Reservoirs are believed to have induced five out of the nine earthquakes on the Indian peninsula in the 1980s which were strong enough to cause damage.
As with most aspects of seismology, the actual mechanisms of RIS are not well understood, and it is impossible to predict accurately which dams will induce earthquakes or how strong the tremors are likely to be. Most of the strongest cases of RIS have been observed for dams over 100 metres high – but dams just half this height are also believed to have induced quakes. Reservoirs can both increase the frequency of earthquakes in areas of already high seismic activity and cause earthquakes to happen in areas previously thought to be seismically inactive. The latter effect is the most dangerous as structures in areas thought to be quiescent are not built to withstand even minor earthquakes. Complicating the picture further are five reservoirs, including Tarbela in Pakistan, where a reduction in local seismic activity was noted after impoundment.
The most widely accepted explanation of how dams cause earthquakes is related to the extra water pressure created in the microcracks and fissures in the ground under and near a reservoir. When the pressure of the water in the rocks increases, it acts to lubricate faults which are already under tectonic strain, but are prevented from slipping by the friction of the rock surfaces.
For most well–studied cases of RIS, the intensity of seismic activity increased within around 25 kilometres of the reservoir as it was filled. The strongest shocks normally occured relatively soon – often within days but sometimes within several years – after the reservoir reached its greatest depth. After the initial filling of the reservoir, RIS events normally continued as the water level rose and fell but usually with less frequency and strength than before. The pattern of RIS is, however, unique for every reservoir.
The most powerful earthquake thought to have been induced by a reservoir is a magnitude 6.3 tremor which flattened the village of Koynanagar in Maharashtra, western India, on 11 December, 1967, killing around 180 people, injuring 1,500 and rendering thousands homeless. The dam was seriously damaged and power cut off to Bombay, causing panic among its populace, who were able to feel the quake 230 kilometres from its epicentre. The epicentre of the tremor and numerous fore– and aftershocks were all either near the Koyna Dam or under its reservoir.
RIS is suspected to have contributed to one of the world’s most deadly dam disasters, the overtopping of Vaiont Dam in the Italian Alps in 1963. The 261–metre Vaiont – the world’s fourth highest dam – was completed in 1960 in a limestone gorge at the base of Mount Toc. As soon as the reservoir started to fill, seismic shocks were recorded and a mass of unstable rock debris on the side of the mountain started to slide toward the reservoir. After reaching a maximum depth of 130 metres in late 1960, the reservoir was partially drained, and the seismic activity and slope movement almost stopped. The reservoir was then filled again, provoking a new increase in tremors. Despite the tremors, engineers and geologists, according to a later engineering report, decided "that the mass . . . would keep moving so slowly that no problems would occur."
The experts were wrong. Heavy late summer rains in 1963 swelled the reservoir. In the first half of September, 60 shocks were registered and the movement on Mount Toc started to accelerate. On the night of 9 October, 350 million cubic metres of rock broke off Mount Toc and plunged into the reservoir. The gargantuan wave resulting from the impact overtopped the dam by 110 metres – the height of a 28–storey building. About two minutes later the downstream town of Longarone was levelled and almost all its inhabitants killed. Altogether 2,600 people died.
The actual relationship between the seismic activity and the landslide is not certain, but it is likely that the numerous shocks at the very least hastened the collapse of the mountainside.
Leonardo Seeber, a seismologist at the Lamont–Doherty Earth Observatory at Columbia University, New York, believes that official maps which show the areas most at risk of earthquakes should also indicate the increased risk near many reservoirs. If this were to happen, communities near reservoirs could presumably demand compensation to "earthquake–proof" buildings, greatly increasing the cost of dams. The dam industry would probably strongly oppose any such measures which would raise awareness of RIS. Seismologist Harsh Gupta, Vice–Chancellor of Cochin University in India and a professor at the University of Texas, notes a "general reluctance in parts of the engineering community, worldwide, to accept the significance or even the existence of the phenomenon of reservoir–induced seismicity." Action in the courts could force the dam industry to accept the importance of RIS: a 1994 article in the Journal of Environmental Law and Litigation concluded that people who suffer from induced quakes would have grounds under US law to sue the operators of a reservoir.