'Earthquake risk' from dams

African dams may need more careful planning

By Alex Kirby BBC News Online environment correspondent, in Nairobi

Large dams in mountainous regions could threaten people living near them by stressing the Earth's crust to danger levels, a scientist says.

The researcher says there have been recorded cases in several countries of dam construction causing earthquakes. Large-scale mining, he believes, can sometimes produce the same result.

He says parts of Africa are especially vulnerable because of the tectonic forces that are shaping the continent.

The scientist, Chris Hartnady, is a former associate professor in the department of geological sciences at the University of Cape Town, South Africa.

He was attending a conference here, the Africa Mountains High Summit, hosted by the United Nations Environment Programme (Unep).

Pulled apart

Professor Hartnady says in his presentation: "Large areas of the African continent are in an unstable, tectonically active state, and especially in the mountain regions substantial danger is posed to growing populations.

"The economic cost of seismic and volcanic disasters is likely to escalate dramatically during this century.

If you dig a big enough reservoir, you're going to get earthquakes

Bill McGuire
Benfield Greig Hazard Research Centre

"Mountain areas appear very attractive places in which to site reservoirs or hydro-electric schemes. However, in east and southern Africa, these high-lying areas are usually associated with tectonically active belts near faults and rifts in the Earth's crust."

Bill McGuire, director of the Benfield Greig Hazard Research Centre in London, UK, says rising seismic and volcanic damage is a worldwide problem.

"There's no question that if you dig a big enough reservoir, you're going to get earthquakes.

"The Three Gorges Dam in China is going to be a big problem," he told BBC News Online.

"There's also the worry that if you build a dam in mountainous terrain that you will get landslides as it fills."

A huge landslide behind the Vaiont dam in northern Italy in 1963 took the lives of over 2,500 people when a wave of water and debris spilled over the dam and swept away a small town, he explained.

Professor Hartnady says the African continental crust is stressed to the "fracture criticality" limit.

Build dams if you must, but engineer them much more sensitively than we do now

Chris Hartnady

He told BBC News Online: "Partly this is because of the African superswell, a mass of warm volcanic rock which is rising under much of south-eastern Africa, producing a buoyancy effect which is helping to pull the crust apart.

"And partly it's the forces at play in the boundary zone between the Nubian and Somalian tectonic plates. Mining triggers earthquakes in South Africa's high veldt."

Real and present danger

In late October 1995, the reservoir behind the Katse dam in Lesotho began to fill. Days later people started feeling earth tremors, and one measuring 3.1 on the Richter scale was recorded on 3 January 1996.

"I am positive that was cause and effect. In 1964, a dam was built at Koyna in India's Western Ghats. There was a big earthquake in the region in 1967 - cause and effect again.

"So build dams if you must, but engineer them much more sensitively than we do now."

Professor Hartnady believes geohazards, including earthquakes, volcanic activity and shifting soils, are an underestimated problem.

He says: "I wonder whether for Africa they may be a more real and present danger than climate change. We badly need more research, on the sort of scale of the effort going into climate change.

"Geohazards are a problem in developed countries too, in places like California and Japan. The US Geological Survey has a wealth of expertise. But some of their knowledge just isn't applicable in places like Africa with much slower rates of motion.

"In the San Andreas Fault in California, the rate is something like 30-50 mm a year, compared with 5-10 mm in the Rift Valley in East Africa. But while 5 mm a year may not be sexy, it could be serious."

Signal in the noise

Compared with the centuries-old records and monitoring from developed countries, Africa's detailed seismological monitoring goes back only about 40 years, Professor Hartnady says.

But he is confident that science means it will soon be possible to predict the risks much more accurately.

"We need to leapfrog to a new kind of technology, and it does exist," he says. "There's everything space geodesy can offer, including the global positioning system, very long baseline interferometry, and satellite laser-ranging. These will measure current rates of plate motion.

"I can't tell people now when the risks will become acute. There could be a catastrophe tomorrow, or we could get through this century without one.

"Come back in five to 10 years, though, when we can start to distinguish the signal from the noise, and I'll give you the numbers. The sooner we begin sophisticated monitoring, the sooner we'll have the answers."