The geographical distribution of the irrigation appears definitely unequal. The availability of surface water is, in a first approximation, the main factor explaining this geography : first of all, irrigation is less important on the dry inland Deccan or in the moutainous areas of the Ghats - naturally very well watered- that in the plains of Tamil Nadu and Andhra Pradesh. Second, high percentage of irrigated lands is located preferentially along big rivers and in their deltas (that of Cauvery and Godavari take shape clearly). Then runoff water and rivers appear crucial for irrigation, as they feed the canals, supply the tanks [cf. definition in the legend of the Karnataka map] and fill the large dams..

On the whole, it is remarkable to note that most of the densely irrigated zones result from major engineering works such as digging of hundreds kilometers long canals, improvement of water diversion structure (anicut) and building of large dams. The Nagarjunasagar project on the Krishna river in Andhra Pradesh, started in 1956 and planned to provide water to 830 000 hectares, the Pochampad project on the Godavari river in North of Andhra Pradesh and the Tungabahdra project in Karnataka that provides water to 350 000 hectares are among the biggest irrigation projects. As a result of this development, canal irrigation has now become the first source of irrigation in South India (39% of irrigated lands) although it was introduced on a large scale during the last 150 years only. It is followed by the use of the subsurface water (30%) which increased a lot since the years 1960, because of the political efforts of the Green Revolution to develop it and to electrify the countryside.

The percentage of irrigated lands does not systematically predict cultures grown, because of the diversity of types of irrigation. However in the zones with great density of irrigation, rice or sugarcane are preferentially grown, while in the drier zones dominate less consuming water plants, such as sorghum, eleusine (a kind of millet), etc.

The data of the 1991 census do not inform us either about the intensity of the irrigation, i.e. the number of irrigated cultures per year, knowing for example that an irrigation by tubewell makes it more possible to get rid of the climatic constraints than an irrigation by tank.

O.A.