The Irrigation Myth — Most Of The World Can Grow Food With Rain Alone

The numbers on freshwater use in agriculture are not disputed, though they are rarely presented in terms of their long-run implications. According to the FAO, agriculture accounts for approximately 70 percent of global freshwater withdrawals and over 90 percent of consumptive use — water that is not returned to the source after use but is evaporated or incorporated into crops.

Of the approximately 1.5 billion hectares of cropland globally, about 310 million hectares are equipped for irrigation. That represents about 20 percent of cropland. Yet irrigation accounts for a disproportionate share of total food output in specific commodities: wheat in the Indo-Gangetic Plain, rice in Southeast Asia, corn and soybeans in the U.S. Midwest, cotton across Central Asia. These are the industrial-scale commodity systems whose outputs dominate global trade flows.

The problem is the water source. Irrigation draws from two sources: surface water (rivers, reservoirs, diverted streams) and groundwater (aquifers). Surface water systems are renewable if managed within the constraints of the watershed's natural hydrology. Many are not. The Colorado River in the United States, the Yellow River in China, the Amu Darya in Central Asia — all of which have historically reached the sea — now frequently fail to do so because withdrawals exceed natural flow. The Aral Sea, once the fourth-largest lake in the world, has essentially disappeared due to irrigation diversions from its feeder rivers.

Groundwater is a different category of problem. Most major aquifers are fossil water — accumulated during wetter climatic periods over thousands to millions of years and recharged at rates that range from negligible to a few centimeters per year. Pumping rates in agricultural use typically exceed recharge by factors of ten to hundreds.

The High Plains Aquifer (commonly called the Ogallala) underlies approximately 450,000 square kilometers across eight states. It supports irrigation for about 30 percent of all groundwater-irrigated land in the United States, producing wheat, corn, cotton, and cattle feed. It took roughly 10,000 years to accumulate to its current state. It is being depleted at a rate that will make irrigation economically impractical in the shallowest and most intensively used areas within 25 to 50 years.

Kansas State University researchers published projections showing that if current pumping rates continue, 69 percent of the High Plains Aquifer will be depleted within 50 years. A 2013 study in Nature Climate Change found that the aquifer had already lost an average of 30 percent of its pre-development saturated thickness across the region, with losses exceeding 70 percent in some southern Kansas and Texas counties.

When a well in the High Plains becomes uneconomical to pump — when the water table drops low enough that the cost of electricity to lift the water exceeds the value of the crop — the land reverts to dryland farming or is abandoned. This is already happening. The transition is not gradual in its economic impact: commodity grain farming on the High Plains has largely been profitable because irrigation subsidized both yields and timing flexibility. Without irrigation, yields drop by 40 to 60 percent for corn and sorghum, and drought years become catastrophic rather than merely difficult.

The broader implication is that a significant portion of global grain production — production that sets the price floor for global commodity markets and on which supply chains for meat, processed food, and ethanol depend — is built on a capital stock that is being consumed without replacement.

The Indo-Gangetic Plain is the breadbasket of South Asia, producing the majority of India's and Pakistan's food supply. It is heavily dependent on groundwater irrigation, and it is experiencing aquifer depletion at rates that alarm hydrologists who study it.

NASA's GRACE satellite mission, which measures gravity variations caused by changes in water mass, documented groundwater depletion across the Indo-Gangetic Plain at roughly 18 cubic kilometers per year between 2002 and 2008. A follow-up study extended this analysis and found accelerating depletion in Punjab and Haryana — the states that constitute India's Green Revolution heartland. These states produce approximately 30 percent of India's wheat and rice from a region that holds perhaps 3 to 5 percent of its population.

The Green Revolution itself is partly responsible. The high-yielding varieties that transformed Indian food production in the 1960s and 1970s required irrigation that did not exist from rainfall alone in the Punjab winters. Massive canal and well infrastructure was built. Subsidized electricity made pumping groundwater cheap. Production soared. And for decades the depletion was invisible because the aquifer is deep and the drawdown per year was small enough to be dismissed.

It is no longer dismissible. States are competing over river water rights. Wells in farming villages that previously yielded water at 20 feet now require drilling to 300 feet. Farmers who cannot afford deeper wells are abandoning groundwater irrigation and seeing yields fall. The debt-driven farmer suicide crisis in India is partly a story of groundwater economics — input costs that only made sense when irrigation was cheap and yields were high, now becoming unsupportable as water access deteriorates.

The narrative that rainfed agriculture is necessarily lower-yielding and more precarious than irrigated agriculture is partially true and largely misleading. It is partially true in the specific context of commodity monoculture without appropriate soil and water management — shallow-rooted annual crops planted on exposed soil in semiarid climates will yield erratically in rainfed conditions. It is misleading because the practices that make rainfed agriculture productive are well-documented and widely implementable.

The core practices are:

Deep-rooted cultivars: Traditional and heritage varieties of many staple crops have root systems that extend 1.5 to 3 meters, accessing subsoil moisture unavailable to modern shallow-rooted high-yield varieties selected for performance under irrigation. The International Maize and Wheat Improvement Center (CIMMYT) has breeding programs specifically targeting drought tolerance, with documented results in sub-Saharan Africa.

Soil organic matter: Soil with 3 to 5 percent organic matter retains significantly more water per cubic meter than degraded soil with 0.5 to 1 percent organic matter. Each 1 percent increase in soil organic matter in the top six inches holds an additional 20,000 gallons of water per acre. Regenerative practices — cover cropping, compost, no-till — build organic matter. Conventional tillage destroys it. This differential in water-holding capacity can be the difference between a crop surviving a two-week dry spell and failing.

Earthworks: Contour bunds, swales, terraces, and check dams slow runoff, allowing rainfall to infiltrate rather than shed from the surface. In the semi-arid Sahel, simple earthwork structures called zaï — pits scraped by hand or ox-drawn tool — concentrate rainfall and compost around individual plants, dramatically improving germination and early growth in years with below-average rainfall.

Mulching: A thick layer of mulch on the soil surface reduces evapotranspiration by 30 to 60 percent, maintains soil temperature, and suppresses competition from weeds. This is perhaps the single most cost-effective intervention for improving rainfed productivity in water-stressed regions, and it can be done with crop residues, green chop, or any available organic material.

Crop diversity: Polyculture systems that mix deep-rooted and shallow-rooted species, drought-tolerant and moisture-sensitive varieties, and short-season and long-season crops are inherently more resilient to rainfall variability than monocultures. If one component of the system fails in a dry year, others may succeed. This is not a compromise on productivity — it is a portfolio approach to production risk that has served subsistence farmers throughout history.

The current global investment trajectory in agricultural water management is moving in approximately the wrong direction. Large infrastructure projects — dams, canals, desalination plants — continue to attract development finance because they are legible, controllable, and create business for engineering and construction sectors. Small-scale water harvesting, farmer-led earthworks, and soil organic matter building receive comparatively little institutional support despite demonstrably superior cost-effectiveness in terms of food security per dollar invested.

A 2016 study by IWMI (International Water Management Institute) found that small-scale water harvesting and storage infrastructure — ponds, tanks, check dams, bunds — could cost-effectively provide supplemental irrigation for 200 to 300 million additional hectares of rainfed cropland, dramatically improving yields in marginal rainfall years without depleting groundwater. The total investment required was estimated in the tens of billions of dollars — a fraction of the cost of equivalent large dam infrastructure.

The reorientation toward rainfed systems is not anti-technology. It requires technology: remote sensing to map rainfall patterns and aquifer depletion, precision breeding for drought tolerance, digital tools for farmer-to-farmer knowledge sharing, and materials science for water-efficient earthworks. What it requires abandoning is the assumption that large centralized water infrastructure is the primary solution — because the accounting no longer supports that assumption when the depreciation of the underlying aquifer capital is included in the ledger.

The plan that sustains food production through the next century cannot be built on water sources that will not exist. It must be built on the water cycle that will — and on the soil, seed, and earthwork systems that make that cycle productive.