Research report: A genetic algorithm-based framework for scenario-driven optimization of cropping patterns using water footprint economic value

A new study from Alborz Province, Iran, has made that hidden economics explicit by building an optimization framework that scores every possible allocation of wheat and barley acreage according to how much water each crop consumes per dollar of net profit. The work, published in the journal Environmental Challenges, combines the water footprint accounting method developed by Arjen Hoekstra and colleagues with a genetic algorithm that searches for the cropping pattern with the lowest economic water footprint across six counties.

The researchers, led by Mehdi Saravani and colleagues, defined the economic water footprint as the volume of water required to generate one unit of net economic benefit, expressed in cubic meters per dollar. This is a subtle but important reframing of conventional water footprint analysis. A traditional water footprint measures water consumed per kilogram of crop, which says nothing about whether that water produced a worthwhile economic return. Two crops can drink identical amounts of water yet deliver wildly different profits. By dividing the physical water footprint by the net economic benefit per kilogram, the new indicator rewards crops that convert scarce water into money efficiently, and it penalizes those that do not.

The study area covers roughly 5,833 square kilometers around 35.48 degrees north and 50.58 degrees east, with an average annual precipitation of only about 250 millimeters. Agriculture there has historically depended on the Karaj River, but increasing competition for water, recurrent droughts, declining rainfall, unauthorized groundwater abstraction, and climate change have squeezed supply. According to Ministry of Agriculture Jihad statistics, the long-term average cultivated area of irrigated crops in the province is about 6,405 hectares, of which wheat and barley together occupy roughly 1,655 hectares.

To quantify water consumption, the team used the CROPWAT 8.0 framework from the Food and Agriculture Organization, feeding it twenty-year average monthly climate data from synoptic stations assigned to each county. Reference evapotranspiration was computed with the FAO Penman-Monteith equation, and crop evapotranspiration followed from crop coefficients drawn from Iran’s national water document. Effective precipitation, the fraction of rainfall actually available to crops, was separated from total precipitation so that the green water footprint, supplied by rain, could be distinguished from the blue water footprint, supplied by irrigation. The resulting physical water footprints ranged from about 0.57 to 1.07 cubic meters per kilogram for wheat and from about 0.58 to 0.83 cubic meters per kilogram for barley, with the lowest values in Karaj and Nazarabad, where shorter growing periods reduce evapotranspiration.

Under current conditions, the economic water footprint of the two crops combined varied sharply by county, from 1.11 cubic meters per dollar in Karaj to 2.14 cubic meters per dollar in Taleghan. The authors estimate that wheat and barley production alone consumes about 35 percent of the province’s renewable water resources, which are drawn from Ministry of Energy statistics and total between roughly 10.9 million cubic meters in Fardis and 45.6 million cubic meters in Karaj per county per year. That heavy dependence on a constrained resource is precisely what motivated the optimization exercise.

The optimization model treats the cultivated area of each crop in each county as a decision variable, twelve variables in all, and minimizes the aggregate ratio of total green and blue water consumption to total net economic benefit. Three families of constraints can be switched on or off: a cap on cultivated area, a limit on available blue water, and a minimum production requirement for each crop. Four scenarios probe different combinations. Scenario 1 frees land allocation while keeping water and production constraints. Scenario 2 removes the water constraint but keeps land and production limits. Scenario 3 keeps land and water limits but drops the production floor entirely. Scenario 4 allows expansion to the maximum available cultivated area while retaining water and production constraints.

The genetic algorithm, implemented in MATLAB R2022a, produced strikingly different answers depending on which constraints were active. In Scenario 1, wheat cultivation fell to zero in five of the six counties, with only Nazarabad retaining 7,056 hectares of wheat because of the production requirement, while barley expanded to 2,971 hectares in Karaj and smaller allocations elsewhere. Scenario 2 shifted wheat production almost entirely to Nazarabad, which the authors attribute to its more favorable climate, lower evapotranspiration, and higher yields, while barley expanded into Savojbolagh with 4,284 hectares. Scenario 3, with no production floor, produced a near-complete shift to barley in every county, eliminating wheat altogether even in Nazarabad. Scenario 4, the most resource-feasible configuration, retained wheat in Savojbolagh and Nazarabad while expanding barley elsewhere.

One of the most counterintuitive findings deserves careful reading: in every scenario, the county-level economic water footprint values were higher than under the baseline condition, rising as high as 3.17 cubic meters per dollar in Taleghan under Scenario 3. Because lower values indicate better economic water efficiency, the authors are explicit that these increases should not be read as improvements. The scenario allocations represent model-derived resource-allocation experiments, not verified gains in water productivity. Scenario 3 in particular is presented as a theoretical benchmark showing what the optimizer would do if food-production requirements were ignored, and the authors warn that its barley-dominated corner solution should not be interpreted as an agronomically optimal or implementable cropping pattern, nor as evidence of a new water-food trade-off mechanism.

The study is candid about its limitations. The model covers only two crops, excludes crop rotation, market dynamics, price variability, farmer preferences, production risk, and policy constraints, and relies on deterministic long-term average inputs rather than interannual variability. The 2012 to 2022 agricultural baseline paired with twenty-year climate averages introduces a potential temporal mismatch. More unusually for a computational study, the original MATLAB source code and random-seed configuration were not retained, so the population size, generation count, crossover and mutation rates, and stopping criterion of the genetic algorithm could not be retrospectively verified, and repeated independent runs could not be reproduced. The authors also note that the problem is small enough that a deterministic linear-fractional programming approach, transformable via the Charnes-Cooper method, could in principle solve it exactly, and they frame the genetic algorithm as a flexible, extensible choice rather than a computational necessity. The reported solutions are therefore feasible scenario outcomes rather than independently verified global optima.

Despite those caveats, the framework offers a template for water-stressed agricultural regions well beyond Alborz. Its transferable element is the analytical structure linking crop water requirements, effective precipitation, water footprint estimation, yield, net economic benefit, and resource-constrained land allocation; the numerical parameters must be recalibrated locally. The central lesson for planners is the trade-off the scenarios expose: maximizing the economic efficiency of water use pushes the system toward crop specialization, while maintaining minimum production of strategically important crops constrains that optimum in exchange for production stability. The authors position the results as decision-support benchmarks rather than prescriptions, and they recommend future work incorporating formal sensitivity analysis, climate projections, multi-objective optimization, and validation against observed cropping patterns before any real-world reallocation of land and water is attempted.

Subject of Research: Scenario-based optimization of wheat and barley cropping patterns in Alborz Province, Iran, using the economic water footprint and a genetic algorithm

Article Title: A genetic algorithm-based framework for scenario-driven optimization of cropping patterns using water footprint economic value

Article References: Saravani, M., Najafyan, D., Pourgholam-Amiji, M., & Hajirad, I. (2026). A genetic algorithm-based framework for scenario-driven optimization of cropping patterns using water footprint economic value. Environmental Challenges, 25, Article 101706. https://doi.org/10.1016/j.envc.2026.101706

Image Credits: AI Generated

DOI: 10.1016/j.envc.2026.101706

Keywords: water footprint, economic water footprint, genetic algorithm, cropping pattern optimization, Alborz Province, Iran, wheat, barley, blue water, green water, irrigation, water scarcity