Dogouri et al. (2025) Development of a stochastic multi-objective optimization model for managing the water, food, and energy nexus in agriculture

Identification

• Journal: Scientific Reports
• Year: 2025
• Date: 2025-12-07
• Authors: Mahshid Ahmadipour Dogouri, Somaye Janatrostami, Afshin Ashrafzadeh, Nader Pirmoradian
• DOI: 10.1038/s41598-025-31197-6

Research Groups

• Department of Water Engineering, Faculty of Agricultural Sciences, University of Guilan, Rasht, Iran
• Department of Irrigation and Reclamation Engineering, Faculty of Agriculture and Natural Resources, University of Tehran, Karaj, Iran

Short Summary

This study develops a novel stochastic multi-objective optimization model, integrating Chance-Constrained Programming (CCP) with the NSGA-II algorithm, to manage the water-food-energy nexus in the Sefidroud River basin, aiming to minimize agricultural water shortages and maximize hydropower production under uncertainty. The model provides a quantifiable framework for decision-making by analyzing trade-offs between these competing objectives across various confidence levels.

Objective

• To develop a novel multi-objective optimization model based on the water–food–energy nexus, incorporating stochastic uncertainty through chance-constrained programming, to simultaneously minimize agricultural water shortages and maximize hydropower production in the Sefidroud irrigation network.

Study Configuration

• Spatial Scale: Sefidroud River basin, specifically the Sefidroud irrigation and drainage network in Guilan Province, northern Iran, encompassing three irrigation areas: Fomanat (F), Central (G), and East Guilan (D).
• Temporal Scale: Agricultural season (first six months of the year, from mid-April to mid-September) for optimization. Historical data period of 25 years (1992-2017) for meteorological and hydrological data, and 33 years of monthly flow data for hydrometric stations. Sefidroud Dam reservoir records from 2000 to 2017.

Methodology and Data

• Models used:
  • Stochastic Multi-objective Optimization Model
  • Chance-Constrained Programming (CCP)
  • Non-dominated Sorting Genetic Algorithm-II (NSGA-II)
  • SIMDualKc model (for estimating crop evapotranspiration)
  • Penman-Monteith method by ASCE-EWRI (2005) (for reference evapotranspiration)
  • USDA-NRCS curve number method (for surface runoff)
  • Tennant method (for environmental water requirement)
  • Kolmogorov-Smirnov test (for selecting probability distributions)
  • Monte Carlo simulation (for model validation)
• Data sources:
  • Meteorological data (temperature, sunny hours, humidity, wind speed, dry temperature, daily cumulative precipitation) from national meteorological organization (six synoptic and five rainfall stations).
  • Water resource data (Sefidroud irrigation and drainage network, Sefidroud Dam inflow/outflow, energy production levels, local river flows, allowable groundwater extraction, rainwater harvesting wetland capacities) from the regional water organization in Guilan province.
  • Cultivated area data for tea and rice.
  • Soil texture map of the region.
  • Historical flow data (33 years) from hydrometric stations.
  • Sefidroud Dam reservoir records (2000-2017).
  • 30-year records of allowable groundwater extraction.
  • Reports from the Guilan Regional Water Authority.

Main Results

• The model successfully generated Pareto-optimal solutions, illustrating trade-offs between minimizing agricultural water shortages (F1) and maximizing hydropower production (F2) at confidence levels of 80%, 90%, 95%, and 99%.
• At a 90% confidence level, a 14.8% reduction in water scarcity requires a 3.15% decline in energy production. Conversely, a 3.25% increase in energy production leads to a 17.4% increase in water shortage.
• Peak hydropower generation (60,000 MWh) was observed in May. However, June experienced the highest agricultural water shortage (127.4 million m³) and the lowest hydropower output (13,127 MWh), particularly in irrigation area D.
• Water deficit generally decreased with higher confidence levels, except for region D in June, where the deficit at 90% confidence level (62.51 million m³) exceeded that at 80% confidence level (56.94 million m³), indicating increased uncertainty.
• The Sefidroud Dam's downstream channels are the primary water source, contributing 58.02% to 83.55% of agricultural water supply. Increasing confidence levels led to reduced reliance on primary sources and increased contributions from secondary sources (rivers, groundwater, wetlands, drainage).
• Groundwater is the sole source for tea irrigation, with optimal allocations of 87.15 million m³ for region D and 3.9 million m³ for region F.
• For rice, at a 90% confidence level, the largest water allocation (6.546 million m³) supported region G (meeting 93% of demand), followed by F (1.388 million m³, 88% met), and D (400 million m³, 80% met).
• Monte Carlo validation confirmed that higher confidence levels lead to greater system stability and lower sensitivity to environmental fluctuations, with water-limited regions (D) showing stronger responses than stable ones (F).

Contributions

• Development of a novel stochastic multi-objective optimization model for the Water-Food-Energy (WFE) nexus by integrating Chance-Constrained Programming (CCP) with the NSGA-II algorithm.
• Explicitly accounts for uncertainties in critical parameters (water availability, agricultural demands, energy consumption) in WFE planning, addressing a gap in previous studies on the Sefidroud network.
• Provides a quantifiable framework for decision-making under uncertainty, allowing policymakers to set flexible risk thresholds and analyze trade-offs between competing objectives (water scarcity vs. hydropower).
• Offers a transferable framework for sustainable development and resource management in other water-scarce regions facing similar WFE challenges.
• Advances nexus-based optimization by integrating stochastic methods and multi-objective trade-offs.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Citation

@article{Dogouri2025Development,
  author = {Dogouri, Mahshid Ahmadipour and Janatrostami, Somaye and Ashrafzadeh, Afshin and Pirmoradian, Nader},
  title = {Development of a stochastic multi-objective optimization model for managing the water, food, and energy nexus in agriculture},
  journal = {Scientific Reports},
  year = {2025},
  doi = {10.1038/s41598-025-31197-6},
  url = {https://doi.org/10.1038/s41598-025-31197-6}
}