Algorithm Finds the Sweet Spot for Water and Nitrogen in Winter Wheat

Winter wheat feeds hundreds of millions of people, and nowhere is that more true than in the Yellow River irrigation district of northern China, where farmers coax grain out of semi-arid soil with heavy doses of irrigation water and nitrogen fertilizer. A new three-year field study published in Plant and Soil suggests that the region’s wheat growers have been flying blind on one of agriculture’s most consequential questions: how much water and nitrogen is actually enough? Using an evolutionary algorithm borrowed from engineering, researchers led by Tengfei Liu of the North China University of Water Resources and Electric Power have pinpointed a precise combination — irrigating when soil moisture falls to 72 percent of field capacity and applying 250.19 kilograms of nitrogen per hectare — that maximizes yield while keeping nitrogen waste and residual soil pollution in check.

The stakes could hardly be higher. The Yellow River irrigation district is one of China’s largest and oldest irrigation systems, and like much of the North China Plain it faces a dual crisis: groundwater tables that have been dropping for decades, and nitrate pollution that seeps into aquifers when fields receive more nitrogen than crops can absorb. Excessive fertilizer inputs do not simply vanish; they accumulate in the soil as residual nitrate, where they can leach into groundwater or escape into the atmosphere as nitrous oxide, a greenhouse gas nearly 300 times more potent than carbon dioxide. Meanwhile, every unnecessary irrigation event draws down water reserves that future growing seasons will need.

Wide-ridge furrow irrigation, the technique at the heart of the study, is a localized surface irrigation method that has been widely adopted for winter wheat in the region because it saves water and boosts yields compared with conventional flooding. Instead of inundating an entire field, water flows down furrows between wide raised ridges, wetting the root zone in a controlled pattern. But the technique’s water-saving potential depends entirely on how it is managed — when irrigation is triggered, and how much nitrogen rides along with the water. Until now, no study had systematically quantified how these two inputs interact under wide-ridge furrow irrigation to shape wheat productivity, nitrogen uptake, and resource-use efficiency.

To answer that question, the research team ran a split-plot field experiment from 2020 to 2023, testing three irrigation lower limits — triggering irrigation when soil moisture dropped to 60, 70, or 80 percent of field capacity — crossed with three nitrogen application rates of 120, 220, and 320 kilograms per hectare. That nine-treatment matrix allowed them to isolate the effects of water, nitrogen, and their interaction on yield, yield components, water use efficiency, agronomic nitrogen efficiency, nitrogen accumulation, aboveground biomass, and residual soil nitrogen. The results were unambiguous: water and nitrogen interacted significantly, meaning the optimal amount of one depended on the amount of the other.

The interaction revealed a clear pattern. At any given nitrogen rate, the middle irrigation level — 70 percent of field capacity — consistently produced the highest yield, water use efficiency, agronomic nitrogen efficiency, and nitrogen accumulation. Irrigating too conservatively at 60 percent stressed the crop, while irrigating too liberally at 80 percent wasted water without improving performance. Similarly, at any given irrigation level, the middle nitrogen rate of 220 kilograms per hectare generally delivered the best overall outcome. The extremes told a cautionary tale: excessive water and nitrogen inputs did increase nitrogen accumulation in the crop, but they simultaneously reduced nitrogen-use efficiency and promoted the buildup of residual nitrogen in the soil — a lose-lose scenario for both farmers and the environment.

But the researchers went beyond identifying which of the nine treatments performed best. They wanted to find the true optimum, not just the best of a limited set. To do that, they built response surface models — statistical functions that describe how each output variable responds continuously to changes in water and nitrogen inputs — and then fed those models into NSGA-II, the non-dominated sorting genetic algorithm II, a multi-objective optimization tool first developed by Deb and colleagues in 2002. NSGA-II works by evolving a population of candidate solutions through selection, crossover, and mutation, sorting them into Pareto fronts that represent trade-offs between competing objectives. In this case, the algorithm had to balance six objectives simultaneously: maximizing yield, water use efficiency, agronomic nitrogen efficiency, nitrogen accumulation, and aboveground biomass, while minimizing residual soil nitrogen.

The algorithm converged on a solution that no single field treatment had achieved: irrigating at 72 percent of field capacity with 250.19 kilograms of nitrogen per hectare. Under this optimized regime, the predicted yield reached 10,012 kilograms per hectare, with water use efficiency of 1.84 kilograms per cubic meter, agronomic nitrogen efficiency of 21.45 kilograms per kilogram, nitrogen accumulation of 172.52 kilograms per hectare, aboveground biomass of 27,318.46 kilograms per hectare, and residual soil nitrogen of just 48.23 kilograms per hectare. The nitrogen rate sits between the 220 and 320 kilogram treatments tested in the field, suggesting that the true optimum lies in a narrow window that conventional trial-and-error approaches would struggle to identify.

What makes this study notable is not just the numbers but the framework. Multi-objective optimization has been applied to water and nutrient management before, but rarely with this level of integration — combining response surface models with an evolutionary algorithm to balance productivity, uptake, efficiency, and environmental performance in a single framework. The approach acknowledges that there is no single best answer in agricultural management; every decision involves trade-offs. A farmer who maximizes yield alone will waste nitrogen. A farmer who minimizes residual nitrogen alone will sacrifice productivity. The optimized solution represents the Pareto-optimal balance point where no objective can be improved without degrading another.

The findings arrive at a moment when China is grappling with the legacy of decades of intensive fertilizer use. National nitrogen-use efficiency in Chinese crop production remains low by global standards, and the environmental costs — groundwater nitrate contamination, greenhouse gas emissions, and soil acidification — are increasingly well documented. The study’s authors argue that their optimized water-nitrogen regime provides a scientific basis for sustainable winter wheat production in Yellow River irrigation districts, where the tension between food security and environmental protection is at its sharpest. Whether the framework can be scaled to other crops, other irrigation systems, and other regions remains an open question, but the principle — that evolutionary algorithms can find management sweet spots invisible to conventional field trials — is likely to resonate far beyond one wheat-growing district.

Subject of Research: Optimization of water and nitrogen management for winter wheat in wide-ridge furrow irrigation systems in the yellow river irrigation district

Article Title: Optimization of water and nitrogen management for winter wheat in wide-ridge furrow irrigation systems in the yellow river irrigation district

Article References: Liu, T., Wang, S., Wang, A., & Zhao, B. (2026). Optimization of water and nitrogen management for winter wheat in wide-ridge furrow irrigation systems in the yellow river irrigation district. Plant and Soil. https://doi.org/10.1007/s11104-026-08870-4

Image Credits: AI Generated

DOI: 10.1007/s11104-026-08870-4

Keywords: Optimization, water, nitrogen, management, winter, wheat, wide-ridge, furrow, irrigation, systems, yellow, river