ÐÓ°ÉÔ­´´

The chinese finished building the Sanmenxia Dam on the Yellow River
in 1960. Four years later, the reservoir behind it had almost filled with
silt and was taken out of action. The dam, built to a Soviet design, had
promised everything. Plans published in the early 1950s boasted that it
would prevent floods, irrigate fields, generate hydroelectric power and
even halt damage from sheets of ice hurtling down the river in spring. The
authorities moved some 300 000 people off their land to make way for the
reservoir, which was 40 times bigger than Britain’s largest, Kielder Water.
But they had forgotten about the Yellow River’s silt.

The Yellow River is not called yellow for nothing. It carries a greater
concentration of silt than any other major river, collecting around 1.5
billion tonnes a year, mainly from the fine ‘loess’ soils of central northern
China. The 106-metre-high Sanmenxia Dam, which was built in the last gorge
before the river reaches the plains of lowland China, barricaded the river
at its muddiest, trapping much of its silt.

The Sanmenxia Dam is now holding back the water of the Yellow River
again, thanks to some extensive re-engineering. But the reservoir has only
one-third of its original storage capacity of 8 cubic kilometres. The eight
holes that once housed turbines have been modified to act as channels that
let silty water through.

The Chinese are unabashed. They plan to build a series of 15 dams in
a cascade down the Yellow River. As part of the programme they will build
a new dam, 170 metres high, downstream of Sanmenxia, to replace part of
its lost capacity. This time, however, they are planning for siltation.
The Chinese government’s engineers calculate that the storage capacity of
the new Xiaolangdi Dam will fall from an initial 12.6 cubic kilometres to
about 5 cubic kilometres after about a decade. The two dams will then protect
the lower plains of the Yellow River against the worst flood that can be
expected in a thousand years. The hope is that the dams will allow a reduction
in the massive job of constantly improving the dykes downstream. On its
flood plain, the river is contained between high dykes and, as silt has
accumulated on the channel bed, it has risen to 5 metres above the level
of the land around.

Chinese engineers have lived with silt for thousands of years. And,
except when swayed by foreign technical advisers, they usually take account
of it when planning their elaborate hydrological structures. But outside
the Middle Kingdom, silt often seems to come as a surprise.

India has built more large dams than any other nation. One of its first,
the Nizam Sagar Dam across the River Manjira in Andhra Pradesh, was finished
in 1931. By the early 1960s, two-thirds of the reservoir’s capacity had
been lost to silt. Since then, efforts to halt soil erosion in the river
catchment have helped keep the lost capacity at around this level. But less
than half the 2700 square kilometres of fields connected to the reservoir
now receive irrigation water.

For three decades, the rate of siltation behind the Nizam Sagar Dam
was 16 times that predicted by its state government designers. But far from
learning from this disaster, India has replicated it many times. The Sriram
Sagar Dam, also in Andhra Pradesh, was completed in 1970 and lost a third
of its storage capacity within two years because of silt settling in the
reservoir bottom. Since then the reservoir has reached a new equilibrium,
with as much silt flushed out as is deposited. But half its capacity is
occupied by silt. Further north, in the Himalayas, Indian engineers have,
on average, underestimated the rate of siltation of dams built in its steep
valleys by a factor of five, and there is little evidence that their skill
has improved.

This year, the Indian government hopes to begin work on a 260-metre-high
dam at Tehri in the Himalayan mountains of Uttar Pradesh (‘The dam that
should not be built’, New ÐÓ°ÉÔ­´´, 26 January). It is the first of a new
generation of very large hydroelectric dams intended for the Himalayas.
Yet it is being built on one of the most silt-laden and landslip-prone rivers
in the Himalayas, the Bhagirathi. The government’s engineers say that the
dam will function for 100 years. But there is continuing controversy about
the research on which this assessment is based.

A working group set up by the Indian government to review environmental
aspects of the Tehri Dam concluded that its designers had, either by accident
or design, published unreasonably low estimates of the sediment that would
accumulate in the dam. This had happened, said the working group, because
sediment monitoring stations along the Bhagirathi were placed where sediment
load was lowest, and because the stations monitored only silt suspended
in the water; they failed to measure the ‘bedload’, larger pieces of rock
and sediment that roll along the river bed. S. P. Nautiyal, former president
of the Wadi Institute of Himalayan Geology, which conducted its own detailed
study of the region, says that the life of the Tehri reservoir may turn
out to be only 30 or 40 years.

Directing nature’s power

Large dams are among the most awe-inspiring monuments to modern society.
They were pioneered by the American New Deal programme of public works in
the 1930s, which built the Hoover Dam on the Colorado river, and Stalin’s
great works on the rivers of Russia. Dams came to be seen as the springboard
for economic growth in the developing world, offering unending supplies
of cheap hydroelectricity for industry and a secure source of water to irrigate
fields. Since 1950, huge plugs of earth, rock and concrete have been placed
across some of the greatest river valleys on Earth, mostly in the tropics.
In the past 40 years, the amount of water trapped behind large dams has
increased 25-fold, and now amounts to around 5000 cubic kilometres. This
is a substantial interruption to the planet’s hydrological cycle; artificial
reservoirs now hold the equivalent of roughly 13 per cent of the total runoff
of rivers to the oceans.

Most engineers who have built dams in the tropics cut their teeth on
smaller dams in temperate regions. They have brought the gospel of large
dams to dozens of countries. But this has led to a fatal blindness. With
few exceptions, the threat of silt to reservoirs has been ignored or downplayed
in designs and economic appraisals. In temperate lands, weathering of rocks
takes place only slowly, whereas in the tropics high temperatures and frequently
greater rainfall make natural rates of weathering much higher. Chemical
weathering, in which water dissolves salts in rock, is typically 20 to 50
times faster in the tropics than in temperate lands. Chemical weathering
leaves behind a crust on rocks that is vulnerable to mechanical erosion
by wind or water.

The difference is clear from a comparison of siltation rates behind
dams on the Colorado, one of the US’s muddiest rivers, and on the Indus
in Pakistan. The Hoover Dam on the Colorado filled with silt at a rate of
only 0.3 per cent per year, an accumulation that has now almost ceased because
new dams upstream intercept the silt. Yet at the comparably sized Tarbela
Dam on the Indus, the siltation rate since its completion in 1974 approaches
2 per cent per annum. Hundreds of such dams in the tropics are likely to
become useless early in the 21st century.

Typically, engineers have assumed that the area of ‘dead storage’ in
reservoirs-the region at the bottom from which water cannot be extracted
because it is below the level of the reservoir outlets-will be sufficient
to hold any silt. But, in practice, due to the vagaries of hydrology, silt
often collects in the ‘live storage’ zones, along the sloping sides of reservoirs,
long before the dead zone is filled. After just six years of operation of
the Tarbela Dam, 44 per cent of silt deposits were in the ‘live’ storage
area, even though 78 per cent of the dead storage zone was still empty.

Following several decades of headlong construction, Western engineers
and their pupils in tropical countries are belatedly beginning to realise
the extent that they underestimated siltation. Khalid Mahmood, professor
of engineering in the department of international water resources at the
George Washington University in Washington DC, completed a study of the
problem in 1988 for the World Bank, which, in the past four decades, has
loaned more money than anybody else for dam construction in the tropics.

Mahmood estimates that every year about 1 per cent of the combined storage
capacity of the world’s reservoirs is being consumed by silt-a staggering
annual loss of 50 cubic kilometres, almost twice the capacity of the reservoir
behind the Hoover Dam. Even with highly conservative estimates of construction
costs, Mahmood puts the value of this loss at $6 billion each year, making
a total loss to date of $130 billion.

The loss is not merely a matter of wasted cash. As Mahmood points out
in his report: ‘In many basins, additional sites (for reservoirs) are hard
to find, and in general, remaining sites for storage reservoirs are more
difficult and, hence, more expensive to develop.’ The world is filling most
of the best sites for reservoirs with huge mounds of silt, most of it caked
so hard by the pressure of water above it that, when exposed, it is useless
for agriculture.

Hydroelectric power is frequently described as a ‘renewable’ resource
that we can tap into as long as the rivers flow to the sea. But from this
perspective, we are destroying this resource as surely as when we are digging
coal from a mine or pumping oil from the North Sea. In the process, we are
turning some of the most fertile river valleys on Earth into permanently
sterile tracts.

There is a huge variation in the amount of silt carried by the world’s
rivers. Among the largest rivers, China’s Yellow River stands out with average
sediment content of 22 000 parts per million. Other rivers with a reputation
for muddiness seem limpid by comparison. The Ganges-Brahmaputra basin, which
drains through India to Bangladesh, carries only 1700 parts per million
of sediment, roughly the same as the Indus.

In many places, dams almost eliminate sediment from rivers downstream,
causing potentially serious problems where farmers need the sediment to
fertilise their fields. The River Nile used to carry more than 100 million
tonnes of sediment downstream each year. The sediment, and the bacteria
that it contained, renewed the fertility of fields all down the Nile valley
and in its delta, the country’s main agricultural region, as well as sustaining
Mediterranean fisheries. Since the completion of the Aswan High Dam in 1964,
this flow has almost ceased. The coastline of the delta has retreated by
more than 2 kilometres in places and the eastern Mediterranean fisheries
have been wiped out.

Ghana completed the Akosombo Dam in 1966. Its hydroelectric power has
failed to transform the country’s economy, but behind the dam Lake Volta
has trapped so much of the River Volta’s silt that coastal erosion in neighbouring
Togo has washed away 10 000 homes in less than three decades. The Indus
in Pakistan used to carry about 440 million tonnes per year into the Arabian
Sea. Since the completion of the Mangla Dam and Tarbela Dam in 1965 and
1974 respectively, the flow has fallen to 100 million tonnes. Most of the
missing 340 million tonnes is clogging up the reservoirs.

On the Nile, the very large capacity of Lake Nasser behind the Aswan
Dam means that the 2 cubic kilometres of silt dropped there in the 25 years
since it was built have reduced the reservoir’s huge capacity by less than
2 per cent. But in the smaller reservoirs behind the dams on the Indus,
sedimentation is a serious problem. Behind the billion-dollar Tarbela Dam,
a delta of silt, deposited as the Indus enters the reservoir, is advancing
down the reservoir towards the dam itself.

The Tarbela Dam’s designers, the US consulting firm Tippetts Abbett
McCarthy Statton, knew that some delta would form, says Mahmood, but they
grossly underestimated how fast it would creep downstream towards the dam.
‘The actual delta crest after 9 years of operation (of the dam) was located
about 18 kilometres upstream of the dam instead of 45 kilometres.’ Unless
it is halted, the delta is likely to reach the dam next year. Engineers
are working to prevent the delta of silt from overwhelming the turbines.
Even if they succeed, the delta will remain a visible harbinger of the dam’s
ultimate fate. The Tarbela reservoir is likely to be all but useless within
two decades-its predicted 100-year life reduced to about 40 years.

Pakistan is hugely reliant on its two major dams on the Indus to provide
water for irrigation of its arid plains and for hydroelectricity to fuel
industrialisation. The government has already drawn up plans for a new dam
at Kalabagh, downstream of Tarbela, to make good the losses as its earlier
dams silt up. But the dam will occupy a much less favourable site for power
generation. A law of diminishing returns is already beginning to overtake
Pakistan’s efforts to harness the Indus.

In many parts of the world, one of the prime sources of silt is not
day-to-day weathering and erosion of rocks, but landslips, often triggered
by earthquakes. Such slips can make nonsense of estimates of the longevity
of reservoirs that are based on measurements of average sediment flows in
rivers. An earthquake in New Guinea in 1970 triggered landslips over an
area of 60 square kilometres that denuded slopes by an estimated 11 centimetres
in a region where the average rate of erosion was 2 centimetres in a 100
years. Mud flows that followed the eruption of Mount St Helens in 1980 quadrupled
sediment loads in the Columbia river, one of the major rivers of the American
northwest.

In the Himalayas, one of the most seismically active regions in the
world, landslips are large and frequent. Such slips could add substantially
to the silt load behind the proposed Tehri Dam on the Bhagirathi. The Bhagirathi
is a tributary of the River Ganges and its valley has, according to the
Geological Survey of India, been the scene of ‘a number of major landslides’
in the recent past. In 1978, a landslip at Kanaudiagad on the Bhagirathi
completely blocked the river, creating a natural dam of its own for several
days. A flood in the adjoining Alakananda valley in 1970 carried 3 million
cubic metres of sediment.

The Geological Survey points out that the operation of a dam on the
Bhagirathi will itself trigger new slides. As the level of the reservoir
rises and falls to meet the hour-by-hour needs for hydroelectric power in
India’s industrial cities, the crumbly shales of the bare reservoir sides
will be broken up by the constant wetting and drying. Whole hillsides, rising
many hundreds of metres from the valley bottom, could be destabilised. Some
geologists, such as Vinod Gaur, a former director of the National Geophysical
Research Institute in Hyderabad, have warned that landslides could cause
tidal waves that might overtop the dam, as well as rapidly filling the reservoir
itself.

An example of what could befall the Tehri, with potentially tragic consequences,
occurred in 1980, on a major tributary of the River Kosi, the third largest
river draining the Himalayas. The Kosi flows from Tibet through western
Nepal before joining the Ganges in India. A landslide first blocked the
tributary and then sent down the Kosi a flood of debris containing boulders
weighing up to 150 tonnes. In just one day some 60 million tonnes of sediment
poured down the river.

The problems of silt in reservoirs are most critical in Asia, where
young rocks are often still being raised up by the tectonic processes that
create mountain ranges. Erosion of these rocks provides around two-thirds
of the sediment entering the world’s oceans. The yield of sediment from
land here averages more than 380 tonnes per square kilometre per year, more
than twice the global average. In the Himalayas, rates are even higher.
The basins of the Yellow River, the Ganges and the Brahmaputra, yield 1000
tonnes per square kilometre.

A second region of the world producing a disproportionate amount of
sediment is Latin America. Here, too, there are many vulnerable dams. Particularly
in Central America, hydroelectric power is the dominant source of electricity,
and studies for the US government’s Agency for International Development
show that many reservoirs have serious siltation problems that threaten
the economic future of several countries. Guatemala is faced with spending
$100 million on remedial measures aimed at reducing the silt load that threatens
the central dam in its Pueblo Viejo Quixal project. El Salvador, Honduras
and Costa Rica have similar problems.

In Central America, the destruction of forests in the catchment areas
of dams is widely blamed for the siltation of reservoirs. There, as elsewhere,
remedial measures concentrate on forest planting, in the hope that trees
will bind soils and prevent them being washed off the land and into reservoirs.
But Mahmood warns that such initiatives frequently do not work, especially
in larger river basins. It is far from clear how important human activity
is in the creation of sediment in rivers, especially at long distances downstream
(‘Floods in Bangladesh: who is to blame?’, New ÐÓ°ÉÔ­´´, 13 April).

A study of 120 years of data on sediment in Coon Creek, Wisconsin, a
small river with a basin of just 360 square kilometres, found that while
forest clearance and agriculture appeared to erode thousands of tonnes of
soil each year, no more than 5 per cent of this excess soil flowed out through
the mouth of the creek. The rest was deposited on sandbanks and river beds
within the basin.

The implication is that efforts to reforest reservoir catchments, either
before building a new dam or to prolong the life of an existing dam, are
not likely to help the reservoir itself unless it has a small catchment.
The classic instance of such failure is the Mangla Dam on the Jehlam river,
a tributary of the Indus in Mahmood’s native Pakistan. A concerted effort
was made to reduce the sediment flow into rivers across almost 8000 square
kilometres of the reservoir’s catchment. The effort began in 1959, five
years before the dam was completed, and continued until 1989. Mahmood says
that the project was expected to reduce sediment load at Mangla by about
30 per cent. But published data from sediment monitoring stations in the
catchment show ‘no discernible difference in the sediment loads’, even in
the areas of most intense and earliest activity. It may be that, measured
over many decades, the efforts to keep soil on the land will be seen to
have influenced sediment load in the River Jehlam-but by then the Mangla
reservoir will probably be long abandoned, filled to its turbines with silt.

So what are engineers to do? First, they need to make more realistic
assessments of the likely sedimentation rates both before designing dams
and when estimating the likely working life of the structures. Secondly,
they need to assess what steps are needed to extend the life of the dams
in operation.

The world’s stock of large reservoirs-once gleaming and brand new-is
starting to age. As the incentive to give them a new lease of life grows,
techniques must be devised to clean them out. Those currently on offer include
flushing the reservoir with river water and dredging-or even building small
‘debris dams’ upstream to catch the sediment before it reaches the main
dam. But these options have not proved attractive in practice. Debris dams
can end up costing more than the main dam, says Mahmood.

Sediment flushing has been tried on Pakistan’s Warsak Dam on the River
Kabul, which lost 18 per cent of its capacity in the first year of operation.
But only about 6 per cent of the annual silt deposition was ever removed
by flushing, and the reservoir had to be almost emptied in order to do even
that. Dredging too is very expensive, costing about 20 times more than the
price of building a reservoir of a capacity equivalent to the volume of
sediment removed. Nonetheless, says Mahmood, dredging may be the only way
to save irreplaceable dam sites.

If Mahmood is right, then the task of making these ‘renewable’ resources
genuinely renewable will cost the world an annual sum of more than $100
billion. And that ignores the backlog of $2600 billion to clear out the
capacity already lost beneath mountains of silt growing at a rate sufficient
to fill three standard swimming pools every second.

Fred Pearce is an environmental journalist. He is currently writing
a book on the politics of water.

Further reading: Reservoir Sedimentation: Impact, extent and mitigation,
by Khalid Mahmood, World Bank Technical Paper No 71, published in 1988.
Available from the World Bank, Washington DC.