A probabilistic appraisal of rainfall-runoff modeling approaches within SWAT in mixed land use watersheds

A probabilistic approach is presented to assess the performance validity of the empirical Curve Number (CN) and physically-based Green and Ampt (G&A) rainfall-runoff methods in the SWAT model. Specifically, the effects of modeling uncertainties on characterization of the hydrologic budgets and streamflow regimes at various spatial scales and upstream land use conditions are investigated. A Bayesian total uncertainty assessment framework, which explicitly accounts for uncertainties from model parameters, inputs, structure, and measurement data, was employed to explore uncertainties in streamflow simulation using SWAT with different rainfall-runoff methods in a mixed-land use watershed. While the models were trained for streamflow estimation only at the watershed outlet, the performances of the models were compared at different stream locations within the watershed. At the watershed outlet, the CN method had a slightly better, but not significant, performance in terms of streamflow error statistics. Similar results were obtained for the predominantly forested and agricultural tributaries. However, in tributaries with higher percentage of developed land, G&A outperformed the CN method in simulating streamflow based on various performance metrics. In general, the 95% prediction intervals from the models with G&A method covered a higher percentage of observed streamflow especially during the high flow events. However, they were approximately 20-45% wider than the corresponding 95% prediction intervals from the CN methods. Using 95% prediction interval for estimated flow duration curves, results indicated that the models with CN methods underestimated high flow events especially in tributaries with highly developed land use. However, the CN methods generated higher water yields to streams than the G&A method. The results of this study have important implications for the selection and application of appropriate rainfall-runoff methods within complex distributed hydrologic models particularly when simulating hydrologic responses in mixed-land use watersheds. In the present study, while CN and G&A methods in the SWAT model performed similarly at the outlet of a mixed-land use watershed, G&A captured the internal processes more realistically. The subsequent effects on the representation of internal hydrological processes and budgets are discussed.