Tamil Nadu is water deficit. A structural deficit and not a seasonal one. The total assessed water resources in the State amount to 1,587 TMC (thousand million cubic feet) while the State government's demand estimate is 1,894 TMC ( Fig.1 ). Demand exceeds supply by 19.3 per cent; this happens when rainfall is "normal".
( Fig.1 )
Consider what gets reported as normal: the aggregate data from across the State that Indian Meteorological Department (IMD) releases. That often does not reflect what went on in the districts. Each district has a different estimate of normal as well, complicating the meaning of a single final figure further. Tamil Nadu's water conservation and reservoir systems are not integrated to the degree that we can take this aggregate number in isolation.
For instance, last season's rainfall for Tamil Nadu was 12 per cent below the estimated normal. The IMD classifies that deviation as normal rainfall. The threshold at which it gets classified as deficit is 20 per cent below estimated normal. Of Tamil Nadu's 34 Meteorological subdivisions, 24 had rainfall below the estimated normal ( Fig.2 ). And of these, 17 had more than 20 per cent deficit. That is, half the State was classified deficit by the IMD measure. Only 4 subdivisions had excess rainfall.
( Fig.2 )
But they seem to have had a disproportionate influence on the aggregate. Tamil Nadu's catchment areas aren't concentrated like it happens in most other geographies. Trichy having a 54-per-cent-deficient rainfall isn't made up by Theni having excess rainfall of 138 per cent. At least not fully and not even if these two districts made up the entire State but retained their relative positions on a map.
Even a marginally deficient rainfall, given the 19.3 per cent of shortage to begin with, is a significant challenge. This challenge is local; making the probability of more than one region being deficient every year very high. In the past five years, the average drop in ground water across Tamil Nadu has been 0.34 metres/year ( Fig.3 ). That is shocking and ranks among one of the most rapid drops in the world.
( Fig.3 )
But what is worse is where these drops have occurred; the average hides the fact some regions/districts have depleted far more than others. The Tamil Nadu PWD maintains observatory wells in 30 districts. 27 of these have shown an increase in depth over a five-year period ( Fig.4 and Fig.5 ). The absolute drops in some cases are 4 meters in five years. What is even more alarming is the rate of relative decline across the State. Even those districts with relatively low depth are showing accelerating trends of depletion.
( Fig.4 )
Districts in the Cauvery Delta region - those that do not come to mind readily as dry regions -- are among those that show the highest relative decline in ground water. Cuddalore, Pudukkottai, Salem, Sivagangai, Thanjavur, Thiruvarur and Thoothukudi have registered drop in depth that exceeds 40 per cent in just five years. In terms of absolute depth, districts in central and western Tamil Nadu dropped their depth well into double digits when measured in metres.
Most of these districts under such stress also happen to have large areas under farmland. It is quite likely that the reason for such an accelerating rate of ground water depletion is agricultural practice. The only piece of data we need to look at to come to that conclusion is the demand estimate of various sectors. Domestic water use, often decried as wasteful, forms three per cent of the total demand. Irrigation takes 93 per cent. Therefore, any attempt to understand Tamil Nadu's water resources should start with agriculture.
( Fig.4 )
Of Tamil Nadu’s total land area, 38.3 per cent is sown. That, to those of us who are not familiar with the data, sounds like a staggeringly high proportion of land under cultivation. With the exception of Chennai, every district has a significant portion of its land sown. For some of the districts, like those in the Cauvery Delta, it’s more than 50 per cent of their total surface area ( Fig.6 ) . As a point of comparison, one of the world’s largest and most productive agricultural States, California, has 27 per cent of its surface area as cropland. And that State is now contemplating reducing that to conserve water.
( Fig.6 )
Tamil Nadu's problem is that in such a water scarce situation, its primary crop is rice -- a hydrophilic plant that requires standing water. Of all land that is sown in the State, 32.3 per cent is sown with rice ( Fig.7 ). Other food and non-food crops do not come close to this dominance. Simple arithmetic tells us 12.16 per cent of the State's total surface area, therefore, is just rice crop. That is over 1.9 million hectares requiring standing water for a significant period of time in hot conditions. With a 19.3 per cent water scarcity to start with under such rapidly depleting ground water resources, is this feasible?
( Fig.7 )
That the rice crop is the reason behind depletion of ground water is a reasonable hypothesis. It is borne out in the distribution of rice crop across districts.
The delta districts have rice in over 60 per cent of their sown land. There are other districts which also reach over 60 per cent. But, as the earlier data revealed, the delta districts also have the maximum total area under cultivation. For example, Kancheepuram and Sivagangai – two districts that aren’t in the delta – have about 20 to 30 per cent of their overall land area sown. And 70 per cent of that is rice. This computes to about 14 to 21 per cent of their total land area coming under rice ( Fig.8 ). Whereas the delta districts of Thanjavur, Thiruvarur and Nagapattinam have 60 to 70 per cent of their total land area sown; and of this, again have 60 to 70 per cent as rice. That is 40 to 50 per cent of their total surface area being occupied by water guzzling crops!
( Fig.8 )
The data above shows regions that grow rice in over 40 per cent of their sown area have either already depleted ground water significantly or are getting there at an accelerated pace. To understand how this happens, it is useful to see what kinds of irrigation methods are used. In Tamil Nadu, surface water irrigation maxed out its capacity in the 1960s. As a result, recent growth in source of irrigation has almost exclusively been open wells, tube wells and bore wells. These kinds of ground water extracting methods now account for 54 per cent of all irrigation.
To understand the spread of such methods further, consider a district-wise mapping of the different kinds of wells that are privately owned ( Fig.9 ). Thanjavur, which has about half of its surface area cropped with rice has the maximum number of bore-wells that suck water out far more effectively. Districts with absolute ground water levels that are not as bad as others seem to have more private wells and bore-wells - which is understandable because they have water at reachable depths and also alarming given it accelerates their depletion further. The regions in western parts of the state that have ground water already at double digit depths (in meters) have far lower levels of rice cultivation.
( Fig.9 )
The government classifies ground water blocks in various categories. Only 145 of the 385 such blocks are classified safe. The others are in various stages of depletion: over-exploited, critical and semi-critical. About 2 per cent of the blocks are already saline.
Tamil Nadu has two immediate questions staring at its face:
a) Can it afford to have 38.3 per cent of its surface area sown?
b) Can it afford to have 32.3 per cent of the sown area be rice?
The answer to both questions seems to be 'no'. The political and policy implications of this are far too complex for anyone to suggest this happen immediately. But a delay in that decision-making only makes the State dry up faster and impact soil fertility negatively.
Desalination, of the type that Israel has mastered, is another option. But productivity of the Tamil farmer does not justify the cost of desalination for irrigation. One hopes the next generation of innovators work to make desalination cheap. And meanwhile, the only reasonable thing is to encourage migration to cities and not decry it. We need less and not more farmland.
(Nilakantan R.S. works as a Data Scientist for a tech startup.)
