The point is that geo-hydrological systems are hugely complex, just like the weather system. Chaos theory suggests that a minor perturbation in some corner has some large result elsewhere: as in 'the beating of a butterfly's wings may cause a storm'.

There are several well-known examples of the unintended, and grave, consequences, of large-scale terra-forming experiments gone awry. One is in the Florida [Images] Everglades in the US; another is the absolute cataclysm in the Aral Sea in ex-Soviet Central Asia; and a third is the construction of the Aswan High Dam in Egypt.

There are smaller examples, such as in the recurring inundation in California's former flood plains in the Central Valley and the dead moonscape of Mono Lake; in the filling-in of most of Bangalore's small water bodies; in the mis-handling of the marshlands in Louisiana that magnified Hurricane Katrina's impact, etc. These are less spectacular, but they, too show us how little we know about the engineering of water bodies.

Consider what happened in the Florida Everglades wetlands. The government, using highly questionable economic models, asked the US Army Corps of Engineers to drain the marshland, build levees and canals, and straighten out meanders in the rivers. But the result has been disastrous. After decades of floods, loss of species, and general mayhem, Florida is now thinking of undoing much of this, at a cost of many billions of dollars!

The Aral Sea was, as recently as 1960, the fourth largest fresh-water body in the world. Soviet authorities decided to divert the Amu Darya river that fed the sea. They wanted to grow cotton, a notoriously thirsty crop, in a semi-arid area. The result: the Aral Sea has lost more than two-thirds of its surface area, there is pesticide contamination from runoff. What remains is an increasingly briny, dying sea, where the fish have been wiped out. And the cotton? Well, there is no cotton crop either.

The negative outcomes of building massive dams are many, and the positives often small and relatively short-lived. Take the giant Aswan Dam on the Nile, a showcase Soviet project. It is true that the Aswan does generate electricity and irrigation water, but the very life-giving silt from the famed annual floods that has sustained Nile delta agriculture for millennia has now become a problem. The catchment area has silted up faster than expected (so the useful lifetime of the dam is shortened). Evaporation losses from the reservoir are enormous (3.5 cubic miles, close to the total amount of water used in a whole year by a mid-sized rich country!). There have been outbreaks of disease from the stagnant water. Basically, the Nile has been turned from a blessing into a curse.

I imagine there will be terrible, but as yet unknown, consequences to the huge Three Gorges Dam in China, too. The consequence of their plan to divert the Brahmaputra northwards in Tibet, of course, are known: north India will become a desert. Three Gorges may only have local impact, but the Brahmaputra certainly will affect India. There are quite possibly, other, unknown and unpredictable outcomes as well.

We can consider a couple of other terra-forming experiments. One, of course, is the release of hydrocarbons from human activity in general and the burning of fossil fuels for electricity and for transportation in particular. A vivid illustration of the law of unintended consequences: surely the melting of glaciers and the rise in sea-levels were not foreseen when Henry Ford unveiled the Model T.

The second is the paving over of forest and agricultural land -- and river basins -- for cities and for factories. Forests and agricultural land tend to be sponges for rainwater and help sustain the hydrological cycle. Cities, in addition to being generators of CO₂ and heat, force water to run off. Industrialisation and urbanisation disrupt global cycles of heat, water and CO_2.

Therefore, based on prior experience, there are major risks in embarking on giant geo-hydrological projects. The problem is that we simply do not know enough. Computer models and simulations are not powerful enough, and we do not know all the independent variables involved. This is the reason that I have been an opponent of the proposed river-linking project in India, even though good friends of mine argue passionately for it, and I am very concerned about India's water issues, which I rate as the single biggest problem facing the country in the medium term. See my previous columns on this topic, Water Wars: Cauvery, 'Chinatown' and 'Cadillac Desert', The River Sutra and The coming water wars.

On the one hand, I fear that there will be disastrous, unimagined end-results to river-linking. On the other hand, the very rationale will go away: there will be no excess water from the Himalayan rivers. The disappearance of the glaciers will mean the Ganga and the Brahmaputra will dry out in the summer, and in any case, the Chinese will divert most of what remains in the Brahmaputra to their thirsty northern areas.

The Sethu Samudram project is another dubious terra-forming project. The negative fallout can be very high; there have been persistent scientific concerns about the flora and fauna (including endangered coral reefs) in the area, the loss of the suspected benign role of the barrier as an absorbent of tsunami surges, and the general patterns of sea-erosion around southern peninsular India. So far as I know, nobody has done detailed modeling in an oceanographic lab.

Let us note that tsunamis are not unknown in this region. There are intriguing submarine structures off Kanyakumari and Mahabalipuram that are almost certainly the remnants of submerged cities. A future tsunami surge barreling through and accelerated by the funnel of a defunct Rama's Bridge may wreak tremendous damage on the Kerala [Images] coast, which, ironically, rose up from the sea as the result of an underwater tectonic movement in pre-historic times.

For the Sethu Samudram, the sums simply do not add up. The alleged benefits are illusory: That a few ships can avoid going around Sri Lanka en route from the west coast of India to the east coast. But coastal shipping in India is insignificant, and the major shipping lanes in the Indian Ocean skirt the east coast: they go from the Persian Gulf down the Arabian Sea and clear across the Ocean to the Straits of Malacca. Not even 1 per cent of this traffic will brave a newly excavated channel in untested waters which would, incidentally, add to the length of their journey; besides, most of the traffic consists of tankers (VLCCs) and other large ships that need unobstructed deep water anyway.

The Sethu Samudram project is a prestige issue for so-called 'Dravidian' politicians. It is a matter of hubris waiting for nemesis.

Just as the Communists did with the Aral Sea and the Aswan Dam and the Three Gorges, the 'Dravidians' of Tamil Nadu will also act in haste and repent at leisure. Alas, the rest of peninsular India will have to pay for their folly.

Comments welcome at my blog at http://rajeev2007.wordpress.com