Hoseingholi et al. (2026) Optimal Hedging Rules Determination for Dam Reservoir Operation Under Climate Change

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-02-01
• Authors: Pegah Hoseingholi, Ramtin Moeini, Mehry Akbary
• DOI: 10.1007/s11269-025-04479-x

Research Groups

• Department of Civil Engineering and Transportation, University of Isfahan, Isfahan, Iran
• Kharazmi University, Tehran, Iran

Short Summary

This study proposes a novel approach combining hydrological modeling, optimization, and climate projections to determine optimal discrete and continuous hedging rules for the Marun Dam reservoir in Iran under climate change. It finds that the discrete hedging rule significantly outperforms standard operation policy and continuous hedging rules in maintaining reservoir storage and reducing severe water shortages, offering a more robust strategy for future water resource management.

Objective

• To determine optimal discrete and continuous hedging rules for the Marun Dam reservoir operation in Iran, considering future climate change conditions and drought scenarios, by integrating hydrological modeling, optimization approaches, and climate projections.

Study Configuration

• Spatial Scale: Marun Dam basin, located in Khuzestan and Kohgiluyeh, Boyar Ahmad provinces, Iran (between 49°35′ and 51°09′ E longitude, and 30°28′ and 31°21′ N latitude). The dam is at 50°21′54″ E longitude and 30°42′36″ N latitude. The dam has a total capacity of 1274.17 x 10^6 m^3 and supplies water for irrigating approximately 5.5 x 10^8 m^2 of agricultural land, provides flood control, and produces 1.368 x 10^15 J of energy annually.
• Temporal Scale:
  • Historical Observation: 1961–2005 (for climate model calibration), 2001–2021 (for current HR evaluation).
  • Future Projection: 2020–2050 (for climate change impact and future HR determination), specifically periods 2021–2030, 2031–2040, and 2041–2050.
  • Drought Indices: Calculated over multiple time scales (1, 3, 6, 9, 12, and 24 months).

Methodology and Data

• Models used:
  • Climate Model: CanESM2 (Canadian Earth System Model version 2) under RCP 8.5 scenario.
  • Downscaling Model: SDSM (Statistical Downscaling Model) software.
  • Inflow Prediction Models: Genetic Programming (GP), Artificial Neural Network (ANN).
  • Optimization Algorithm: Genetic Algorithm (GA) for determining hedging rule coefficients and storage volumes.
  • Reservoir Operation Policies: Standard Operation Policy (SOP), Discrete Hedging Rule (HR), Continuous Hedging Rule (HR).
  • Drought Indices: Standard Precipitation Index (SPI), Standardized Runoff Index (SRI).
  • Spatial Interpolation: Inverse Distance Weighting (IDW) method.
• Data sources:
  • Observed precipitation, maximum, and minimum temperature data from four synoptic stations (Borujen, Shahreza, Lordegan, Yasouj) in the reservoir's upstream.
  • Historical inflow and release values for Marun Dam.
  • NCEP (National Centers for Environmental Prediction) data (large-scale low-resolution).
  • CMIP5 (Coupled Model Intercomparison Project Phase 5) outputs.

Main Results

• Improved Reservoir Storage: Discrete HR resulted in significantly higher average reservoir storage (546.6 x 10^6 m^3) compared to continuous HR (469.12 x 10^6 m^3) and SOP (285.48 x 10^6 m^3) during the historical operation period (2001-2021).
• Reduced Water Shortages: Discrete HR reduced the maximum monthly shortage volume from 424 x 10^6 m^3 (SOP) to 338.1 x 10^6 m^3, outperforming continuous HR (433.66 x 10^6 m^3).
• End-of-Period Storage: Discrete HR achieved the highest reservoir storage at the end of the operation period (655 x 10^6 m^3), compared to continuous HR (469.12 x 10^6 m^3) and SOP (285.48 x 10^6 m^3).
• Performance Trade-offs: While discrete HR excelled in storage preservation and shortage reduction, its reliability (time) and sustainability index were slightly lower than SOP for historical conditions (e.g., Rel_t of 0.78 for discrete HR vs 0.82 for SOP; SI of 0.64 for discrete HR vs 0.68 for SOP).
• Climate Change Projections: Simulations under the RCP 8.5 scenario indicate a gradual decline in precipitation and an increasing temperature trend for the Marun Dam basin until 2050. This suggests an increased probability, frequency, and persistence of drought events in the future, with short-term droughts becoming more frequent but shorter in duration, and long-term droughts becoming less frequent but longer.
• Inflow Prediction Accuracy: The Genetic Programming (GP) model demonstrated slightly better performance (RMSE = 1.09, R^2 = 0.98) in predicting reservoir inflow compared to the Artificial Neural Network (ANN) model (RMSE = 1.12, R^2 = 0.97).

Contributions

• Proposes a novel integrated approach combining hydrological modeling, climate projections, and optimization (GA) to determine both discrete and continuous hedging rules for dam reservoir operation under climate change.
• Demonstrates the superior performance of discrete hedging rules over continuous hedging rules and standard operation policy in maintaining reservoir storage and mitigating severe water shortages, particularly under drought conditions.
• Integrates advanced data-driven models (GP, ANN) for accurate reservoir inflow prediction based on downscaled climate model outputs.
• Provides valuable insights for water resources planning and policy-making in drought-prone regions facing climate change impacts, specifically for the Marun Dam in Iran.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sector.

Citation

@article{Hoseingholi2026Optimal,
  author = {Hoseingholi, Pegah and Moeini, Ramtin and Akbary, Mehry},
  title = {Optimal Hedging Rules Determination for Dam Reservoir Operation Under Climate Change},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-025-04479-x},
  url = {https://doi.org/10.1007/s11269-025-04479-x}
}