|YEN, HAW - Texas Agrilife Research|
|JEONG, J - Texas Agrilife Research|
|ARABI, M - Colorado State University|
Submitted to: International Journal of Agricultural and Biological Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/1/2013
Publication Date: 5/25/2015
Citation: Yen, H., White, M.J., Jeong, J., Arabi, M., Arnold, J.G. 2015. Evaluation of alternative surface runoff accounting procedures using the SWAT model. International Journal of Agricultural and Biological Engineering. 8(3):54-68.
Interpretive Summary: The Soil and Water Assessment Tool (SWAT) model uses the curve number (CN) procedure to predict surface runoff. The CN calculation method can be based on either soil moisture or plant evapotranspiration (ET). The primary objective of this study is to evaluate uncertainty associated with these differing methods on SWAT predictions. The methods were evaluated both individually and combined against a test watershed dataset with measured discharge and nitrate loss. The plant ET method and the combined approach were found superior to the traditional soil moisture method.
Technical Abstract: For surface runoff estimation in the Soil and Water Assessment Tool (SWAT) model, the curve number (CN) procedure is commonly adopted to calculate surface runoff by utilizing antecedent soil moisture condition (SCSI) in field. In the recent version of SWAT (SWAT2005), an alternative approach is available to apply CN method by implementing information from plant evapotranspiration (SCSII). Improved surface runoff predictions using SCSII has been shown in previous studies. However, few quantitative comparison of model performance has been made between the two CN approaches alone or in combination. In addition, the effect of SCSII on water quality responses (e.g. total nitrate, pesticide) in SWAT has not been evaluated. The primary objective of this study is to evaluate the role of structural uncertainty associated with these differing methods on hydrologic and water quality predictions. The analysis hinges on the evaluation of: (i) To characterize improvements in hydrologic and water quality predictions may be made by utilizing different surface runoff estimation techniques alone and in combination; and (ii) To investigate how model uncertainty may be affected by combining such techniques. Two approaches are combined by the Bayesian model averaging (BMA) method in multi-site, multiple-response case studies at the Eagle Creek watershed, Indiana. Results show that SCSII and BMA associated approaches exhibit outstanding performance in both discharge and total NO3 predictions compare to SCSI. In addition, applications of BMA have positive effect on the increase of inclusion rate but the predictive uncertainty is not evidently reduced/enhanced. Therefore, we recommend additional future SWAT calibration/evaluation research with an emphasis on the impact of SCSII on the prediction of other pollutants.