WHAT comes to mind immediately when we talk of traditional water harvesting systems are the tanks and eris that punctuate south India. But these systems are not restricted to the south alone: local variations are to be found, both functioning and dry, all across the country.

Funnily enough, estimates of the number of tanks in the country are at best whimsically sketchy, at worst wide of the mark. The Minor Irrigation Census that was conducted about 6 or 7 years ago tentatively pegs the number of tanks at something like a quarter of a million.

We have some (inaccurate) estimates at the time of Independence. The storage capacity of these tanks countrywide was estimated then at 15 billion cubic metres (cum), when the total storage capacity of large reservoirs was estimated at about 18 billion cum. Thus, a large part of the water storage was then in the form of traditional water harvesting structures -- ubiquitous tanks, most of them still functional legacies of the past, with their basic design tried and tested, evolved and adapted, down the centuries. The same technology, ancient as it is, is still in use today.

At the time of Independence, tanks irrigated a sizeable area of about 22 million-odd ha. But there has been no wide-ranging or comprehensive statistical enumeration exercise; in fact, no one knows what the current total tank storage capacity is.

On the other hand, the total capacity of large storages (medium and mega projects) shot up from 18 billion cum to 146 billion cum in 1970 and to an estimated 230-240 billion cum thereafter.

The ongoing national water crisis gave birth to 2 mutually exclusive groups of experts: those who believe that without large storages no appreciable quantity of water can be impounded; and those, especially from ecologically sensitive groups, who insist that small storage tanks make far more sense: their contention is that small tanks are cheaper, more amenable to local control and cause far less ecological damage.

In Tamil Nadu and Andhra Pradesh, among other states, we find that tank storages are shallow -- on an average, about 1.5 metres deep. Another feature is the ratio of the water spread -- the area submerged by the water storage -- to the area that can be served by that water is close to 1:1 in some cases; in others, it varies from 0.7-0.8:1 to as much as 1.1-1.2:1.

The tank advantage
All this is relevant to arguments about the benign aspects of the economical, ecological and technical advantages of tanks over other kinds of storages. There are, in fact, 2 types of tanks: the first is what is called the "isolated, purely rainfed tank" (the term speaks for itself). The sole source of supply is the water run-off from the catchment that feeds the tank.

Second, it is typical in south India for a stream with a tank at its head to run for up to 40 km. The water that is caught in the catchment is obviously a mere fraction of what is caught in the entire stream. The surplus flows down into the stream; plus, the rainfall below the first tank and the catchment of the next tank will feed the second tank and so on. This is a chain that promotes a kind of unidirectional interdependence. Water supply to the downstream tanks depends very much on what happens to the "upstream rainfall", how it is impounded and how it is utilised.

System tanks are not limited by the rainfall in the immediate catchment: they are linked to another source -- either a larger reservoir or a diversion across a river. In this case, both the quantum and the assurance of rainfall is substantially delinked -- increased, in fact -- from what rain the immediate catchment happens to grab.

Of the nearly 40,000 tanks in Tamil Nadu, 20,000 serve less than 40.5 ha each. About 5,000 serve larger areas. About 4,000 of them are really system tanks.

There are also zameen tanks, which cause a lot of trouble because they are tied to a particular form of land tenure and are privately owned by zamindars. All the other types of tanks in Tamil Nadu and in much of the ex-Madras Presidency are Ryotwari tanks, which are owned by the community of irrigators. A 40-ha tank can help only a fraction of the total cultivated area of the village. But the entire picture is not as dismal as the statement suggests: there are, of course, villages that are blessed with more than one small tank. Often enough, there are 3 or 4 tanks impartially irrigating as many as 15 villages in Tamil Nadu. These tanks are relatively big.

However, the norm is the single village tank. Because of the fact that it has to be shared, there is usually a formal, stiffish bonhomie within the community that is served by the tank. The consequence of this sharing of goodies has, among other things, institutionalised social ties into hardline convention.

The management of these tanks is embedded deep in the social structure. The tanks are, technically, privately owned, but the street-smart insistence of the villagers has ensured decentralisation -- if not autonomy -- of decision-making about the traditionally inflexible rules of upkeep and repair.

The tanks vary widely in terms of the frequency and number of fillings. For instance, if a tank with a 30,000 cum capacity is topped only once a year, common sense says that you will get 30,000 cum of water per annum. At first glance, this is an impressive figure, till you come across tanks that providence fills thrice a year. (There are system tanks that get filled 10 times a year.)

The ex-Madras Presidency classified tanks from class I to class VI, according to the duration and assurance of supply. Perennial tanks were in the top-level category, and the seasonal, very unreliable tanks -- the local tanks and the rainfed tanks -- were in the lowest. There are, of course, complex equations involved in the estimation of irrigation water available. In a normal year, if you multiply the storage capacity by the number of fillings and then divide it by the ayacut (command area), you get the irrigation water available, which is about 1.5 metres per ha.

To come to the institutions: in Tamil Nadu, particularly, temples have historically played a very important role in investing their surplus in tanks. We have located tanks in North Arcot, which are called mamulnamas. These were enumerated sometime in 1800. The texts essentially put down in vernacular Tamil a summary of what the traditional practices were. Who was entitled to what water, which sluice would be opened, who would be the functionaries, how would one ration water in case of a shortage? -- every niggling detail was explained.

The zamindar owned the tank and it was in his interest to see that the tank was maintained well, purely because his ryot, his tenant, would be able to cultivate better and could then be squeezed for a higher rent. The British did away with all this: after they took over, there was a decimation of the oligarchs.

What the English did not anticipate was that the disappearance of the zamindars would lead to a disintegration of the services they commanded and ensured. In the Ryotwari areas, the mirasi system ensured that land control was concentrated in a few families. There was a traditional system of sharing the total produce of the land of the mirasi among various groups in the community, including labourers, temple workers, fisherfolk, carpenters and others.

The other feature of this social organisation were the engineers. The controllers decided to make the Harijans the bureaucrats of the water systems. They were the ones who were supposed to open and close the sluice, make announcements for the maintenance works and enforce water rationing (using the rotation system or whatever else they had recourse to). The system was clearly not equitable. But it was workable because it was embedded in a social and political structure that had very distinct inegalitarian characteristics.

New order
But things have changed considerably. As of today, control over land is no longer concentrated as it was in the past. It has passed into the hands of the middle and lower castes. There is also a disjunction between the land control structure and the political structure of the village. A major chink in the social armour is that the local power system is no longer invincible or supreme: intervention comes from government functionaries, local political bigwigs and so on. Often enough, the local MLA awards the contracts for the maintenance of the tanks, effectively taking the wherewithal out of the hands of the people.

There are two ways in which the people have coped with this intrusion and erosion of traditional authority. Where the systems are still in use, we have documented cases of self-regulated labour contribution for cleaning up the inlets and the distribution channels.

But where there are wells, the situation changes drastically in 2 ways: first, the people who have the wells usually take the lead in enforcing the collective regulations. But once they can appropriate the water they need (by pumping), their interest in getting involved in the communal maintenance and the enforcement of regulations flags. It weakens even more when the social imperatives of caste -- the composition, the numbers involved, the kind of conflicts -- barricade the path of "social service".

The second situation arises where the wells do their job so well that tank water becomes insignificant. I know of one case where the tank channels have been closed and the tank is purely a percolation device.

I"d have thought that in the face of this "obsolescing" of a traditional device, the entire edifice would have collapsed. It has not. In fact, what came up was a formal organisational structure which was enforced by some government order. That organisational structure has gone kaput, no elections have been held for 20 or 30 years -- but the tank itself is working.

Disturbing neglect
The neglect is, nonetheless, disturbing. When the dams are breached, for instance, no one bothers about repairs, which effectively means that the tank becomes unusable. Compounding this is the fact that as urban settlements expand, people fill up the tank beds and erect houses on the landfill. In the process, drainage channels are wiped out.

Unfortunately, there is still no counter to this kind of encroachment. Even the statistics vary: the Minor Irrigation Census gives one number, the Public Works Department another. The only way to get around this lethargy is to pick a district and count the number of tanks it has, and how many have gone out of use and why. I would even like to use satellite imagery. Ideally, the local people ought to be involved in such an inventory, largely because it will have educative value: they would know, at first hand, exactly where they stand.

There can be no argument that discharges have to be reduced. I, for one, do not see why denudation should reduce discharge, even if the catchment is fragmented and interrupted. Denudation will certainly increase the flood flow, but the moot question is: where do you keep the surplus? What denudation does is increase the intensity of the discharge and shorten its duration -- too much released in too short a period.

Also, inlet channels are noticeably narrowing. (In most places, inlet channels have been encroached upon and are choked with weeds.) The problem, however, is eminently susceptible to exaggeration: people do come forward to do their bit; and let us not forget that the government does carry out some repair work in bits and pieces. Much of this was done in the late 19th and early 20th centuries, but with a rider attached: the colonial administration said that it would pass irrigation investment only if there were assured financial returns.

Confusion compounded
Compounding this confusion is the fact that, in the Minor Irrigation Programme, tank construction falls under 5 different programmes, none of which accepts responsibility for headcounts. In fact, while the Planning Commission claims that many tanks have been constructed and that their potential has increased, the land use statistics of the net area irrigated by the tanks is evidently declining. So who"s telling the truth?

What they have done in Tamil Nadu in the name of tank modernisation reads like a bureaucratic horror story. If you look at the old survey maps, the field channels are articulated to each survey number. They were articulated to the plots which are to be numbered.

The Tamil Nadu government is trying to scuttle this established technique and is trying to introduce a water-use cooperative. There was no consultation with the people about problems and constraints. Yet the answers lay with the community ensconced right where the action was. The government ended up with monsters that cost upward of Rs 1 lakh per ha. And it is but predictable that 6 months hence, the farmers will make mincemeat of the concrete-lined channels.

There are reasons enough to push for NGOs that have a distinct local identity. This can be replicated and doesn"t require a high level of expertise. The questions that hit you in the face are: What is the present state of the tanks? What are the possibilities of integrated watershed management of the tank command system in order to clear up the inlet channels, improve the cover in the upstream catchment to augment the effective utilisation of local rainfall for impounding in the tanks?

This ought not be limited to just afforestation programmes and building new percolation points but upgrading and maximally utilising existing facilities. Further, the only thing necessary is a single, integrated land, water and watershed-centred programme in any given local area. This would go a long way in cutting down overheads, reducing fragmentation and waste and avoiding duplication.

Another aspect of this total programme is finding out how many of the existing rainfed tanks can be converted to system tanks by linking them to reservoir systems or river diversion systems.

The first thing that the government has to realise is that tanks are not land-savers. Quite the contrary, in fact: if you have shallow storages, large areas will be submerged. Then, it ought to be made clear that functional tanks need a particular kind of topography and rainfall pattern. It is not accidental that many tanks are concentrated in what I call the "2-peak rainfall regimes": they tap a bit of the rainfall that comes along with the first monsoon -- the southwest monsoon -- and use it for nursery raising. The second crop is fed by the main rains.

And this is one of the primary reasons why large storages are necessary: in our kind of rainfall and climatic regime, where groundwater comes at a premium, there are no alternatives to the big dams.

And quite a bit of the groundwater recharge depends upon surplus water. One reason why the Central Groundwater Board keeps "upping" the recharge statistics without quite spelling out the methodology is that they do not separate straight recharge and the induced recharge from canals. It is clear enough that the notion that tanks can be a major substitute for large storages is completely misplaced.

What would be interesting is the integration of small storages with large storages, the kind of thing that they have been talking of in the Narmada dam. Although the tank storage capacity in Saurashtra is a small fraction of the area"s total storage, if the number of fillings can be increased, the effective quantum of water they can handle will shoot up.

In any case, local management is the key. Wherever you find system tanks, you will also find very strong local institutions to manage canal water. Tanks allow the people flexibility enough to use the water when they please.

Where to start
But again, while this is all very well, where do improvement and modernisation begin? We can start with the village community, who are familiar with the changes that have taken place down the decades. We don"t need an army of engineers. The concepts of water balance, the relationship between water use and productivity, the scheduling of water, moisture trends, among others, are established and already in place.

In Bhutan, for instance, this tradition is long, firmly entrenched and virtually impregnable: the involvement of their local communities is amazing. These people are not engineers, they are pure and simple farmers. This is the kind of decentralisation of authority and activity that can haul us out of the morass we are in. There is no way an outside engineer can discover the reasons for inadequacy, distribution problems and so on through the vehicle of a survey.

These matters apart, what is imperative is "user education": all we need is 5 years of hyperactive networking and the dissemination of ideas through NGOs. The communities will have to learn, perhaps evolve through their own internal logic, different ways of managing the water. In the pragmatic nomenclature of this day and age, it"s all a question of producing higher returns per unit of water.

A Vaidyanathan, former member of the planning Commission, is professor Emeritus at the Madras Institute of Development Studies.