Hao et al. (2025) A Flexible Python Module for Reservoir Simulations with Seasonally Varying and Constant Flood Storage Capacity

Identification

• Journal: Water
• Year: 2025
• Date: 2025-12-25
• Authors: Xiaodong Hao, Yali Hao, Xiaohui Sun, Li Tang
• DOI: 10.3390/w18010068

Research Groups

• Shanxi Institute of Geological Survey Co., Ltd., Taiyuan, China
• College of Geological and Surveying Engineering, Taiyuan University of Technology, Taiyuan, China
• School of Management, Wuhan Textile University, Wuhan, China
• Shanxi Center of Technology Innovation for Mining Groundwater Pollution Prevention and Remediation in Karst Area, Taiyuan, China

Short Summary

This study presents an open-source Python module integrating three storage-oriented reservoir schemes to compare constant versus seasonally varying flood storage capacity (FSC) strategies and assess operational zone parameter uncertainty. It reveals that constant FSC significantly outperforms seasonal variation for outflow simulation and reduces storage errors across 289 global reservoirs, recommending constant FSC with H22 or S25 as the default for large-scale modeling.

Objective

• To assess the comparative performance between constant flood storage capacity (FSC) and seasonally varying FSC strategies in storage-oriented reservoir simulation schemes.
• To systematically investigate the impact of empirically determined parameters used to define the four operational zones on reservoir simulation accuracy.

Study Configuration

• Spatial Scale: Global, evaluated using daily observations from 289 globally distributed reservoirs.
• Temporal Scale: Daily records, with at least 10 years of continuous daily data prior to 2020 for each reservoir.

Methodology and Data

• Models used: S25, Z17, H22 (three leading storage-oriented reservoir operation schemes). A new open-source Python module was developed to integrate these models with both constant and seasonally varying FSC options.
• Data sources:
  • GRanD dataset (Global Reservoir and Dam database) for reservoir locations and total storage capacities.
  • Daily in situ observations of inflow, outflow, and storage for 289 reservoirs (compiled from previous studies).
  • Continuous monthly satellite area time series.
  • Continuous naturalized streamflow time series data at dam sites.
  • Flood Storage Capacity data (publicly available on Zenodo).

Main Results

• The constant Flood Storage Capacity (FSC) strategy consistently and substantially outperforms seasonally varying approaches for outflow simulations. Median outflow Nash-Sutcliffe Efficiency (NSE) increased by 0.18 to 0.47, and constant FSC was selected as the optimal strategy for 83.7% of reservoirs when outflow accuracy was prioritized.
• Constant FSC significantly mitigated storage simulation errors, reducing the magnitude of storage errors by 38% to 61%.
• Outflow simulation performance is only moderately sensitive to operational zone parameterization under constant FSC, with median NSE values ranging from 0.188 to 0.360.
• Storage simulation is far more sensitive to operational zone parameterization, with median NSE spanning from -13.9 to -2.26.
• Among the integrated models, H22 and S25 emerged as the strongest performers under the recommended constant-FSC framework, with H22 showing particular strength in storage simulation due to its satellite-constrained active-storage estimates.
• The original published default operational zone schemes of H22 (Zoning IV) and S25 (Zoning VIII) perform near the global performance frontier under constant FSC.

Contributions

• Provides the first global comparison of constant versus seasonally varying flood storage capacity strategies in storage-oriented reservoir schemes.
• Conducts the first systematic assessment of the impact of empirically determined operational zone parameters on reservoir simulation accuracy.
• Introduces and releases an open-source Python module that integrates three state-of-the-art storage-oriented reservoir operation schemes, offering a flexible and reproducible platform for the hydrological modeling community.
• Offers clear practical guidance for large-scale hydrological and land-surface modeling, strongly recommending constant FSC with H22 or S25 as the default for uncalibrated global applications.

Funding

• Key Research and Development Project of Shanxi Province, grant number: 202202020101007.
• Shanxi Province Water Conservancy Science and Technology Research and Promotion Project, grant number 2025GM18.

Citation

@article{Hao2025Flexible,
  author = {Hao, Xiaodong and Hao, Yali and Sun, Xiaohui and Tang, Li},
  title = {A Flexible Python Module for Reservoir Simulations with Seasonally Varying and Constant Flood Storage Capacity},
  journal = {Water},
  year = {2025},
  doi = {10.3390/w18010068},
  url = {https://doi.org/10.3390/w18010068}
}