Sreeraj et al. (2025) Data-driven parameterization of SWAT+ reservoir module without access to operation rules

Identification

• Journal: Environmental Modelling & Software
• Year: 2025
• Date: 2025-12-25
• Authors: S. Sreeraj, P. Athira, Kristin Peters, Jens Kiesel
• DOI: 10.1016/j.envsoft.2025.106852

Research Groups

• Department of Civil Engineering, Indian Institute of Technology Palakkad, Palakkad, Kerala, India
• Environmental Science and Sustainable Engineering Centre, Indian Institute of Technology Palakkad, Palakkad, Kerala, India
• Institute of Natural Resource Conservation, Department of Hydrology and Water Resources Management, Christian Albrechts University of Kiel, Kiel, Germany
• Stone Environmental, Montpelier, VT, USA

Short Summary

This study developed a data-driven approach using simulated annealing to parameterize the SWAT+ reservoir module without requiring operational rules. The method significantly improved the accuracy of modeled reservoir outflow, storage, and evaporation, enabling reliable hydrologic simulations in data-scarce environments.

Objective

• To develop and apply a data-driven approach using simulated annealing for parameterizing the reservoir module in the SWAT+ hydrological model when information on reservoir operation rules is unavailable.

Study Configuration

• Spatial Scale: Cedar Creek Reservoir catchment, United States.
• Temporal Scale: Daily time steps for observed release flow, precipitation, and temperature data.

Methodology and Data

• Models used: SWAT+ (Soil and Water Assessment Tool Plus) hydrological model; Custom R scripts employing simulated annealing Optimization Module.
• Data sources:
  • Daily precipitation and temperature (PRISM)
  • Land Use/Land Cover (LULC) for 2014 (CropScape)
  • Digital Elevation Model (DEM) (Shuttle Radar Topography Mission - SRTM)
  • Soil data (Soil Survey Geographic Dataset - SSURGO)
  • Streamflow at gauges 08062895 and 08062800 (USGS waterdata)
  • Hydrologic inflow to Cedar Creek reservoir (Tarrant Regional Water District, Texas)
  • Evaporation loss from Cedar Creek reservoir (Observed data)
  • Observed release flow time series (Used for optimization)

Main Results

• The data-driven approach substantially improved modeled reservoir outflow by 163.7 % (KGE = 0.72), storage by 588.2 % (KGE = 0.83), and evaporation by 85.3 % (KGE = 0.63) compared to default parameterizations.
• The method successfully enabled accurate SWAT+ simulations even without access to explicit reservoir operation rules.

Contributions

• Presents a novel data-driven methodology for parameterizing the SWAT+ reservoir module using simulated annealing, specifically addressing the common challenge of unavailable reservoir operation rules.
• Demonstrates significant improvements in the accuracy of simulated reservoir outflow, storage, and evaporation using only observed release flow time series for optimization.
• Provides a practical solution for hydrologic engineering and water resources planning in data-scarce regions by enabling reliable reservoir simulations.

Funding

• Not specified in the provided text.

Citation

@article{Sreeraj2025Datadriven,
  author = {Sreeraj, S. and Athira, P. and Peters, Kristin and Kiesel, Jens},
  title = {Data-driven parameterization of SWAT+ reservoir module without access to operation rules},
  journal = {Environmental Modelling & Software},
  year = {2025},
  doi = {10.1016/j.envsoft.2025.106852},
  url = {https://doi.org/10.1016/j.envsoft.2025.106852}
}