Thirsty Industries: Irrigation

From the window of any aircraft crossing a dry continent, the pattern is the same. Green circles on brown ground, half a mile across, in rows and clusters, hundreds of them between one horizon and the next: in Kansas and Texas, in Saudi Arabia, in the Kalahari, on the plains of Castile. Each one is a field watered by a single machine, a pipe on wheels a few hundred metres long that pivots around a well at its centre and walks slowly round in a day or two, spraying as it goes. The circle is the shape of a well's reach. From the air it is also the shape of the industry that takes more water than every other on this site put together.

About seven tenths of the fresh water that people withdraw from rivers, lakes and the ground goes to irrigating crops. Industry, including every power station, steel mill and chip plant on this site, takes about a fifth; cities and households take about a tenth. This article is about how a field is watered, where the water goes once it is on the field, why most of the world still floods its fields the way Mesopotamia did, why making irrigation more efficient has so often made it use more, and what the water leaves behind.

What a plant does with water

A crop uses water in one way that counts and several that do not. The one that counts is transpiration: the plant draws water from the soil through its roots and lets it out through the pores in its leaves as vapour, and in doing so it pulls the nutrients up with it and keeps itself cool and rigid. A field of maize or wheat in high summer puts five to seven millimetres of water into the air each day, which is fifty to seventy cubic metres a hectare, and over a season it adds up to four to eight hundred millimetres, which is more rain than most of the world's farmland gets in the months the crop is growing. Where the rain falls short, the farmer makes up the difference, and that is irrigation.

The water that does not count, in the plant's accounting, is everything else: what evaporates from the wet soil surface before the roots reach it, what runs off the end of the field, what soaks past the roots into the ground, and what leaks from the canal on the way. Those are the losses that the word efficiency refers to, and the arguments over irrigation are mostly arguments over where that water went.

Flooding the field

Most of the irrigated land on Earth is watered by gravity. Water is let from a canal through a gate onto a field that has been levelled or ridged for it, and it flows down furrows between the rows or across a basin bounded by low banks, and soaks in as it goes. It is the method of the Nile and the Indus and the Yellow River, it needs no pump and no pipe, and it is how the rice of Asia, the cotton of Central Asia and much of the wheat and maize of the world is still grown. Its weakness is that the field drinks unevenly: the top of the furrow is flooded for hours while the bottom waits, the water that reaches the end runs off or sinks, and in a hot dry wind a share of it evaporates from the wet surface before any root can take it. Of the water let onto a flood irrigated field, something between forty and sixty percent is taken up by the crop; the rest goes somewhere else.

Somewhere else is the important phrase. The water that sinks past the roots goes into the ground and, in a river valley, back to the river or the aquifer, where it can be pumped and used again downstream; the water that runs off the end of the field goes into a drain and to the next farm. The losses of a flood irrigated field are a loss to that field and often a gain to the one below it, and in a basin like the Nile's the same water may pass through several fields before it reaches the sea. Only the evaporation, from the wet soil and the open canal, is truly gone.

Sprinklers and drip

The two other methods put the water closer to the plant. A sprinkler, whether a pipe on wheels, a centre pivot or a gun on a reel, throws the water through the air onto the crop, which distributes it evenly across a field of any shape and lets a farmer water on a schedule rather than when the canal runs; it loses to evaporation and wind drift on the way down, and a pivot in a desert at noon can lose a fifth of its water before it lands. A drip line lays the water at the root through a tube with an emitter every few centimetres, often under a plastic mulch, a litre or two an hour per plant, so that the soil surface stays dry and the plant gets almost exactly what it transpires. The crop gets seventy to eighty five percent of the water from a sprinkler, and ninety or more from a drip line, and the drip line can carry the fertiliser too, which is why the orchards and vineyards and vegetable fields of California, Israel and Spain have gone over to it.

The paradox

The obvious conclusion, that switching the world's fields from flood to drip would free up a large share of the world's water, has been drawn by every government that has subsidised the switch, and it has mostly turned out to be wrong, for a reason that took a while to accept. A farmer who installs drip and gets the same crop from half the water does not, as a rule, leave the other half in the river. The farmer plants more land, or a thirstier crop, or a second crop in the dry season, because the water is now cheap enough per tonne to make it worth doing, and the basin's total consumption, the water that goes into the air through plants, goes up rather than down. The flood irrigated field's losses, which had been recharging the aquifer and feeding the downstream farm, stop, and the downstream farm finds its canal lower. Studies of the Ebro in Spain, the Rio Grande, the Murray in Australia and the basins of China have found the same pattern, and the paper that named it in 2018 called it the paradox of irrigation efficiency. Efficiency saves water for the farmer and for the field; it saves water for the river only if someone caps what is taken, and the cap is the hard part.

The circles and the aquifer

Under the Great Plains of the United States, from South Dakota to Texas, lies the High Plains aquifer, water that filled the gravels under the plains over tens of thousands of years, and since the 1950s it has been pumped through the pivots for maize, wheat and cattle feed at several times the rate the rain refills it. The water table has fallen by thirty metres and more in the southern plains, by a metre a year in parts of Kansas and the Texas panhandle, and wells that pumped a thousand gallons a minute in 1970 pump a few hundred, or are dry; the groundwater article on this site has the mechanism. The same is true under the Punjab, under the North China Plain, under Saudi Arabia, which pumped a fossil aquifer to grow wheat in the desert for thirty years and then stopped, and under the Central Valley of California, where the ground itself has sunk by metres as the water was taken from beneath it. About two fifths of the world's irrigation water comes from the ground, and in the dry regions where it matters most, the ground is being emptied.

Salt

The water that the plant transpires leaves the field pure; the salt that was dissolved in it stays in the soil. Every irrigation water carries some, a few hundred milligrams a litre in a good river, more from a well, and a field that is watered for decades without enough drainage to wash the salt below the roots accumulates it until the crop will not grow. It is the oldest failure of irrigation, the reason the fields of Sumer went from wheat to barley to nothing over the centuries, and it is the condition of about a fifth of the world's irrigated land today, from the cotton plains of Uzbekistan that the Aral Sea article describes to the Murray valley and the Indus. The cure is drainage, and more water to flush the salt, which goes down the drain to the river and makes it saltier for the next farm, which is the salt problem of the Colorado at the Mexican border and of the lower Murray. The brine article on this site is about the same salt at the far end of a desalination plant; here it is at the near end, in the ground.

What it teaches

Irrigation is the industry behind nearly every number on this site. The blue water in a tomato, a strawberry, a kilogram of rice or a cotton shirt is a field like these, watered by gravity or sprinkler or drip from a river or a well, and the seven tenths of the world's withdrawals that it takes are the reason a city's water, a factory's water and a power station's water are the smaller arguments. The field is also where the fix is. A cap on the water a basin can take, drainage under the fields that salt is killing, a price on the groundwater that is being mined, and drip lines where they save water rather than move it: those are the four things that would change the number, and the first is the one that nobody has found easy.

Seven tenths of the water, a fifth of the land, two fifths of the food, and a water table falling a metre a year under the circles.