Rice is one of the thirstiest crops on the planet, consuming roughly 800 to 1,200 millimeters of water over a single growing season just through evaporation from the soil and transpiration through the plant’s leaves. But that figure only accounts for the water the crop itself uses. Factor in the water that seeps through the soil, evaporates from standing paddies, and gets lost in conveyance, and a flooded rice field can demand anywhere from 1,200 to well over 2,000 millimeters of water from planting to harvest. The range is enormous because it depends on soil type, climate, how the field is managed, and whether the rice is grown under standing water at all.
Where All That Water Actually Goes
To understand why rice needs so much water, it helps to break the total into its components. The biggest chunk is evapotranspiration, the combined water lost from the soil surface and through the rice plant’s leaves. In a study of irrigated lowland rice under semiarid conditions, seasonal evapotranspiration came out to about 840 to 855 millimeters across two growing seasons, with daily rates swinging between roughly 4.5 and 10.5 millimeters depending on the growth stage.1Italian Journal of Agronomy. Actual evapotranspiration and crop coefficients of irrigated lowland rice (Oryza sativa L.) under semiarid climate The peak water demand happens during the mid-season stage, when the canopy is fully developed and temperatures are high.
The second major drain is percolation and seepage, water that moves downward through the soil or sideways out of the field. This is where the numbers get wildly variable. On heavy clay soils with a compacted plow layer, percolation losses can be near zero. On sandy soils with a permeable subsoil, losses can reach one to five centimeters per day or more.2Agricultural Water Management. Water use efficiency of flooded rice fields II. Percolation and seepage losses Over a growing season lasting 100 to 120 days, even a modest percolation rate of one centimeter per day adds up to another 1,000 millimeters on top of what the crop itself transpires. That is why rice grown in sandy lowlands requires dramatically more irrigation than rice grown in clay-bottomed paddies, even though the plants themselves use roughly the same amount of water.
The practical takeaway is that the rice plant’s biological water need is not as extreme as the total field water use suggests. A huge share of the water applied to a rice paddy never passes through the plant. It seeps away, evaporates off the standing water surface, or runs off. This distinction matters because it means many of the water savings possible in rice come from managing those non-crop losses rather than starving the plant.
Why Rice Is Grown in Standing Water
Rice is semi-aquatic, but it does not strictly need to stand in water to survive. Upland rice varieties have been grown in rain-fed, non-flooded conditions for centuries, and some wild rice populations carry stacked alleles for robust root systems adapted to aerobic soils.3PLoS Genetics. Loci and natural alleles underlying robust roots and adaptive domestication of upland ecotype rice in aerobic conditions So why do most rice farmers flood their fields?
The biggest practical reason is weed control. Standing water suppresses the germination and growth of many weed species that would otherwise compete fiercely with the rice crop. Early flooding reduces the costs of hand weeding and cuts dependence on herbicides, which is especially important in smallholder farming systems across Asia and Africa.4PubMed Central. Adaptation to flooding during emergence and seedling growth in rice and weeds, and implications for crop establishment Flooding also stabilizes the soil chemistry in ways that benefit nutrient availability, particularly for phosphorus and iron. In other words, the traditional paddy system is less about meeting the rice plant’s water needs and more about creating growing conditions that suppress competitors and enhance fertility. That understanding opens the door to strategies that use far less water without necessarily hurting yields.
Alternate Wetting and Drying
The most widely studied water-saving approach for lowland rice is alternate wetting and drying, commonly called AWD. Instead of keeping the field flooded for the entire season, the farmer allows the water to recede and the soil to dry partially before re-flooding. The result is a cycle of wet and dry spells rather than continuous inundation.
A review of AWD adoption across multiple countries found that it has reduced irrigation water input by up to 38 percent with no yield reductions when implemented correctly.5Field Crops Research. Review Adoption and economics of alternate wetting and drying water management for irrigated lowland rice The catch is in the phrase “when implemented correctly.” The threshold matters a lot. A meta-analysis pooling data from many trials found no meaningful yield penalty when the water table was allowed to drop only to about 15 centimeters below the soil surface. But when fields were dried more aggressively, with the soil drying well beyond that threshold, yields dropped by roughly 23 percent.6Field Crops Research. Rice yields and water use under alternate wetting and drying irrigation: A meta-analysis The line between “safe” AWD and “too dry” AWD is not large, which means farmers need a practical way to monitor how far the water table has fallen. In practice, many programs use a simple perforated tube pushed into the paddy soil so the farmer can see the water level at a glance.
Some field evaluations have reported even more dramatic savings. One study found AWD reduced irrigation water by 40 to 45 percent while improving grain yield by 28 to 40 percent over conventional flooding, and a related system known as the System of Rice Intensification (SRI) saved up to about half the water with 30 to 38 percent higher yields.7Physics and Chemistry of the Earth. Evaluation of water use efficiency and rice yield under different water saving irrigation practices Results like these tend to come from well-managed experimental stations rather than average farm conditions, so the real-world savings are often more modest. But the direction is clear: rice does not need to stand in water all season long, and cutting back on flooding can actually improve yields in some settings by encouraging stronger root growth and better soil aeration.
Drip Irrigation and Aerobic Rice
A more radical departure from the paddy system is growing rice under aerobic conditions with drip irrigation, essentially treating rice more like a row crop. This eliminates standing water entirely and delivers water directly to the root zone.
Research on drip-irrigated aerobic rice has shown water savings of around 50 percent compared to traditional flooding, with a roughly 29 percent improvement in grain yield and a twofold increase in water productivity, the amount of grain produced per unit of water used.8Agronomy Journal. Evaluation of Drip Irrigation System for Water Productivity and Yield of Rice The challenge is cost. Drip systems require tubing, emitters, pumps, and more management skill than flood irrigation. But studies have found that with optimized lateral spacing, drip-irrigated aerobic rice can achieve about 42 percent irrigation water savings relative to surface flooding while remaining economically viable.9Field Crops Research. Optimizing the lateral dripline spacing of drip-irrigated aerobic rice to increase water productivity and profitability under the water-limited condition
For now, drip-irrigated rice remains a niche practice, mostly tested in research stations and in areas where water scarcity makes conventional paddy farming untenable. It is not likely to replace flooded paddies across the major rice-growing regions of South and Southeast Asia anytime soon, but it represents a proof of concept that rice can be productive on dramatically less water if the economics justify it.
The Methane Problem
Flooded rice paddies are one of the largest human-caused sources of methane, a greenhouse gas far more potent than carbon dioxide over a 20-year horizon. Waterlogged soil creates the oxygen-free conditions that methane-producing microbes thrive in. Anything that disrupts those conditions tends to reduce methane emissions, and that is one of the biggest environmental arguments for AWD beyond water savings alone.
Periodic drying interrupts the anaerobic environment and can cut methane emissions by up to about 48 percent compared to continuous flooding.10PubMed Central. Advances in mitigating methane emissions from rice cultivation: past, present, and future strategies Field experiments have confirmed this range: one study found a 31 percent reduction in methane under AWD,11Environmental Progress & Sustainable Energy. Assessing the efficacy of water management and wheat straw addition in mitigating methane emissions from rice paddy fields while another testing different drying intervals found reductions of 24 to 39 percent depending on how long the soil was allowed to dry between irrigations.12Agriculture, Ecosystems & Environment. Methane and nitrous oxide emissions in rice fields influenced with duration of cultivars and irrigation regimes
There is a tradeoff, though. When the soil dries, a different set of microbes produces nitrous oxide, another potent greenhouse gas. The same study that found a 39 percent methane reduction under a three-day drying interval also reported a 29 percent increase in nitrous oxide emissions. On balance, the net greenhouse gas impact was still substantially lower with AWD, but the nitrous oxide offset means that the climate benefit is never quite as large as the methane reduction alone would suggest. And for deep-water rice or varieties that depend on continuous flooding for optimal performance, intermittent drainage may not be a workable option at all.13Rice Science. Methane Emission from Rice Fields: Necessity for Molecular Approach for Mitigation
Arsenic and How Water Management Affects What Ends Up in the Grain
Water management in rice is not just about quantity and greenhouse gases. It also directly affects the concentration of arsenic in the grain you eat. Arsenic occurs naturally in many soils, and when those soils are flooded, the chemistry changes in ways that release arsenic from soil particles into the water, where rice roots readily absorb it. Under flooded conditions, arsenic concentrations in the soil solution can be roughly 7 to 16 times higher than when the same soil is kept aerobic.14Environmental Science & Technology. Growing Rice Aerobically Markedly Decreases Arsenic Accumulation That translates into grain arsenic levels 10 to 15 times higher in flooded rice versus rice grown under aerobic conditions.
This finding has significant implications for food safety, particularly in regions where soils are naturally high in arsenic or where historical contamination from mining or pesticides has raised levels. Even with aerobic water management, some arsenic still enters the grain, and the proportion of inorganic arsenic (the more toxic form) can increase under aerobic conditions, though the total concentration remains lower than in flooded rice.15PubMed. Mitigation of arsenic accumulation in rice with water management and silicon fertilization
The flip side is that drier conditions can increase the availability of cadmium, another toxic metal. Field experiments with dry direct seeding, where rice is planted into non-flooded soil, found that the drier conditions reduced iron plaque formation on roots and elevated cadmium availability in the soil.16PubMed. The impact on Cd bioavailability and accumulation in rice (Oryza sativa L.) induced by dry direct-seeding cultivation method in field-scale experiments So the relationship between water regime and grain safety is not straightforward: flooding raises arsenic, drying raises cadmium. In contaminated soils, the optimal water management strategy for food safety depends on which contaminant poses the greater risk.
Breeding Rice That Needs Less Water
While irrigation engineers work on delivering less water to the field, plant breeders are working on rice that can get by with less. One promising avenue involves root architecture. A gene called DRO1 (Deeper Rooting 1) controls the angle at which rice roots grow downward. Higher expression of DRO1 drives roots into a steeper, more downward-pointing growth pattern, helping the plant access deeper soil moisture during dry spells.17PubMed. Control of root system architecture by DEEPER ROOTING 1 increases rice yield under drought conditions Researchers have identified upregulation of DRO1 under drought stress across multiple rice varieties, and mining for favorable alleles of this gene among drought-tolerant donors is an active area of work.18PubMed Central. Allele mining for a drought responsive gene DRO1 determining root growth angle in donors of drought tolerance in rice (Oryza sativa L.)
A longer-term and more ambitious effort aims to install elements of the C4 photosynthetic pathway into rice, which naturally uses the less efficient C3 pathway. C4 plants, including maize and sugarcane, concentrate carbon dioxide more effectively inside their leaf cells, which reduces water lost through their pores during photosynthesis. If rice could be engineered to use a similar mechanism, it could potentially produce the same or greater yields while losing less water through its leaves, particularly in hot, dry environments.19PubMed Central. Improvement of photosynthesis in rice (Oryza sativa L.) by inserting the C4 pathway This project has been underway for years and remains in the research phase, but it illustrates the scale of what breeders are attempting.
Groundwater Depletion in Major Rice Regions
All these water-saving strategies exist against a backdrop of accelerating pressure on freshwater supplies. In northwestern India, the shift to high-yielding, water-intensive crop varieties during and after the Green Revolution drove a massive expansion in groundwater pumping. In states like Punjab, Rajasthan, and Haryana, groundwater extraction now far exceeds natural recharge rates.20Environment, Development and Sustainability. Combatting groundwater depletion in North-Western India: Obstacles, consequences, and water-saving agronomic approaches Rice is a major driver of that imbalance, though not the only one.
The problem is not confined to South Asia. In northeast China, which has become a key rice production region, modeling under climate change scenarios projects a 29 to 52 percent increase in irrigation water demand and a 24 to 53 percent increase in the groundwater depletion ratio by mid-century, depending on the emissions pathway.21Agricultural and Forest Meteorology. Rice water requirement exacerbates groundwater depletion in Northeast China under a changing climate The Sanjiang Plain, which has the largest concentration of rice paddies in the region, is especially vulnerable. Even AWD irrigation, while helpful, does not fully offset the projected increases under more extreme warming scenarios.
Virtual Water and the Global Rice Trade
When a country imports rice, it is also importing the water that was used to grow that rice. Economists call this “virtual water,” and the numbers for rice are staggering. Between 2001 and 2022, the global rice trade moved an estimated 1,476 billion cubic meters of virtual water between countries. About 63 percent of that was rainwater captured by the soil, about 29 percent was irrigation water drawn from rivers or aquifers, and the remainder was the water that would be needed to dilute the pollution generated during production.22Sustainable Water Resources Management. Modifying the virtual water trade of rice for global sustainability of freshwater resources
The largest virtual water exporters in rice are Thailand, India, Pakistan, Vietnam, and the United States. The largest importers include China, Iran, Saudi Arabia, and the Philippines. Roughly 59 percent of this virtual water trade was classified as unsustainable, meaning the exporting regions were drawing water faster than it could be replenished. India alone has been exporting enormous volumes of embedded water through its rice trade, with one estimate putting the virtual water flow at over 24,000 million cubic meters per year, much of it from groundwater percolation.23Journal of Applied and Natural Science. India rice export and virtual water trade In effect, countries importing cheap rice are offshoring their water footprint onto regions that can least afford it.
This dynamic adds an uncomfortable layer to the question of how much water rice needs. The biological answer depends on climate and soil. The real-world answer depends on trade policy, groundwater regulation, farmer incentives, and whether the true cost of water is reflected in the price of a bag of rice. In most of the world’s major rice-exporting regions, it is not.