Neri et al. (2026) On the Choice of Optimal Reservoir Operating Rules in a Changing Climate for the Sustainable Management of Drinking Water Sources

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-03-01
• Authors: Mattia Neri, Elena Toth
• DOI: 10.1007/s11269-025-04473-3

Research Groups

• Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, Bologna, Italy

Short Summary

This study develops a multi-objective optimization framework to define optimal reservoir withdrawal rules for a multi-basin drinking water supply system in Northern Italy, assessing their adaptation to climate change under historical and future meteorological forcings to maximize production and minimize deficits. The research quantifies future expected relative changes in optimal operating rules and outlines corresponding patterns of withdrawal volumes and potential water system failures throughout the century.

Objective

• To optimize multi-objective reservoir withdrawal rules for a drinking water supply system in Northern Italy under historical and future climate conditions, aiming to maximize water supply from the reservoir and minimize water deficits, thereby informing long-term resource management and reducing reliance on less sustainable sources.

Study Configuration

• Spatial Scale: Romagna region, Northern Italy, focusing on the Ridracoli multi-basin reservoir system. The reservoir is fed by a direct catchment of approximately 37 square kilometers and four indirect catchments totaling around 50 square kilometers.
• Temporal Scale:
  • Historical reference period: 1982–2010 (with 1981 as warm-up).
  • Future period for optimization: 2071–2100.
  • Simulations for withdrawal and deficit patterns: 1982–2100.

Methodology and Data

• Models used:
  • Hydrological model: TUW model (a lumped version of the HBV model) for rainfall-runoff simulation.
  • Evaporation model: Modified Penman formulation (Linacre 1977).
  • Optimization algorithm: NSGA-II (Non-dominated Sorting Genetic Algorithm-II) for multi-objective optimization.
  • Overall approach: Parameterization-Simulation-Optimization (PSO).
  • Reservoir system model: Hourly water balance equation incorporating inflows, evaporation, spillway outflow, bottom outlet releases, and withdrawal outflow.
• Data sources:
  • Historical meteorological data: E-OBS dataset (gauge-based, gridded, 0.1° spatial resolution, daily minimum/maximum temperatures, daily precipitation depths from 1950).
  • Future climatic forcings: Ensemble of ten GCM-RCM modelling chains from EURO-CORDEX, forced by the RCP 8.5 emission scenario, bias-corrected by Dosio (2016) using E-OBS as reference.
  • Reservoir management history, technical constraints, historical water demand, and information on local water sources provided by RomagnaAcque – Società delle Fonti.

Main Results

• Future Climate Projections (RCP 8.5): Monthly mean areal precipitation is projected to decrease by up to 30% in summer months (March-October) by the end of the century, with a slight increase (5–10%) in winter. Average daily minimum temperature is projected to increase by more than 4 °C in summer for the long-term future.
• Optimal Operating Rules: Future optimal withdrawal rules are generally more conservative than historical rules, prescribing less water withdrawal due to expected reduced hydrological availability. Minimum Deficit Solutions (MDS) are more conservative than Trade-Off Solutions (TOS). MDS rules show a stronger shift towards conservative operation in future scenarios, particularly from autumn to spring, while TOS rules exhibit smaller changes limited to spring and summer.
• Withdrawal and Deficit Volumes: The magnitude of deficit volumes increases throughout the century, especially for less conservative (TOS) rules designed under historical conditions. When adapting management rules to future conditions, the relative decrease in yearly production by the end of the century is approximately 12%, irrespective of the adopted management policy (TOS or MDS). Adopting TOS rules leads to a slightly more limited decrease in potable volumes but an expected increase of 0.4 million cubic meters per year in deficit volumes compared to MDS. Highly risk-averse policies aiming for zero-deficit conditions imply a substantial reduction in total supplied volumes from the reservoir, increasing reliance on less sustainable alternative sources. Conversely, policies allowing a slightly greater risk of deficit can partially counterbalance the expected future decrease in water availability, mitigating stress on alternative sources.

Contributions

• First-ever optimization of the drinking and multi-source water system for the Romagna region, addressing the challenge of highly seasonal water demand.
• Implementation of a Parameterization-Simulation-Optimization (PSO) approach with a comprehensive model for multi-basin reservoir operations, accounting for complex real-world constraints.
• Utilizes a full ensemble of ten EURO-CORDEX climate projections (GCM-RCM chains) for robust assessment of adapted operating rules, overcoming limitations of previous studies that used smaller ensembles.
• Quantifies the expected relative changes in optimal reservoir operating rules under different management policies (Trade-Off and Minimum Deficit Solutions) due to climate change.
• Provides insights into the expected patterns of withdrawal volumes and potential water system failures throughout the century, aiding decision-makers in long-term water resource management planning and reducing reliance on less sustainable water sources.

Funding

• European Union Next-GenerationEU (National Recovery and Resilience Plan – NRRP, Mission 4, Component 2, Investment 1.3 – D.D. 1243 2/8/2022, PE0000005) within the RETURN Extended Partnership.

Citation

@article{Neri2026Choice,
  author = {Neri, Mattia and Toth, Elena},
  title = {On the Choice of Optimal Reservoir Operating Rules in a Changing Climate for the Sustainable Management of Drinking Water Sources},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-025-04473-3},
  url = {https://doi.org/10.1007/s11269-025-04473-3}
}