Development of reservoir operation functions in SWAT+ for national environmental assessments

Authors: WU, JINGWEN - Texas Agrilife Research; YEN, HAW - Texas Agrilife Research; Arnold, Jeffrey; YANG, ETHAN - Lehigh University; CAI, XIMING - University Of Illinois; White, Michael; SANTHI, CHINNASAMY - Independent Ecological Researcher; MIAO, CHIYUAN - Beijing Normal University; SRINIVASAN, RAGHAVAN - Texas Agrilife Research

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/7/2020
Publication Date: 1/9/2020
Citation: Wu, J., Yen, H., Arnold, J.G., Yang, E., Cai, X., White, M.J., Santhi, C., Miao, C., Srinivasan, R. 2020. Development of reservoir operation functions in SWAT+ for national environmental assessments. Journal of Hydrology. 583:124556. https://doi.org/10.1016/j.jhydrol.2020.124556.
DOI: https://doi.org/10.1016/j.jhydrol.2020.124556

Interpretive Summary: To accurately simulate stream flow and quality across the U.S., an accurate reservoir release model is required. In this study, a new reservoir release model was developed and tested using daily reservoir volumes and releases from 123 reservoirs in different ecoregions of the conterminous U.S. The reservoir model parameters were categorized based on: 1) reservoir capacity, 2) reservoir release rates, and 3) climate conditions, to allow categorization and parameterization of all reservoirs in a national model. Results showed the simulated daily reservoir outflows were accurately simulated when compared to measured outflows as evidenced by commonly used statistical measures. The resulting model will provide more realistic results when used for national conservation and environmental assessments.

Technical Abstract: The Soil and Water Assessment Tool (SWAT) model is one of the most extensively used watershed-scale models across the world. In recent years, in order to enhance the potential of SWAT code maintenance and future development, a completely reconstructed version of SWAT, dubbed SWAT+, has been just released. In addition to the new model structure in SWAT+, reservoir operation functions have been added to improve the performance of model simulations. Before conducting watershed-scale model simulations, reservoir functions of SWAT+ were further coupled with the automatic calibration tool, IPEAT, to better consider the interactions among reservoir parameters and the corresponding hydrologic processes. In this study, a total of 123 reservoirs were simulated at daily scale across the contiguous United States (CONUS). A series of reference values of reservoir parameters (total 15) were categorized based on different classifications: (i) reservoir capacity; (ii) reservoir release; (iii) climate conditions; and, (iv) reservoir storage and release. Overall results showed that the performance of reservoir simulations at daily scales is satisfactory by commonly adopted statistical measures. In different scenarios, SfNFD (seasonal flood control drawdown days during non-flood season) (from 5 to 126 days), EfFLD (exclusive flood control drawdown days) (from 0.1 to 2.1 days) and EfFLC (exclusive flood control constant flow) (from 4.8 to 6.5 ha-m) are sensitive parameters and the converged parameter ranges were close to each other but not for other parameters. The recommended parameter ranges may be embedded by varying sources of uncertainty. However, the proposed calibration guidelines provide fast and accessible reference for future SWAT+ studies especially in subjects related to reservoir operations.