A Bayesian total uncertainty analysis framework is presented to assess the model estimates of the effectiveness of watershed management practices in reducing nonpoint source (NPS) pollution. The framework entails a two-stage procedure. First, various sources of modeling uncertainties are characterized during the period before implementing Best Management Practices (BMPs). Second, the effectiveness of BMPs are probabilistically quantified during the post-BMP period. The framework was used to assess the uncertainties in effectiveness of two BMPs in reducing daily total nitrogen (TN) loads in a 54 ha agricultural watershed in North Carolina using the SWAT model. The results indicated that the modeling uncertainties in quantifying the effectiveness of selected BMPs were relatively large. Assessment of measured data uncertainty revealed that higher errors were observed in simulating TN loads during high flow events. The results of this study have important implications for decision-making under uncertainty when models are used for water quality simulation.

提出了贝叶斯总不确定性分析框架，以评估流域管理实践在减少非点源污染方面的有效性的模型估计。该框架包含两个阶段的过程。首先，在实施最佳管理规范（BMP）之前，对建模不确定性的各种来源进行了表征。其次，在BMP后阶段，概率性地量化了BMP的有效性。该框架用于使用SWAT模型评估北卡罗来纳州一个54公顷农业流域中的两个BMP降低每日总氮（TN）负荷有效性的不确定性。结果表明，量化所选BMP有效性的建模不确定性相对较大。对测量数据不确定性的评估表明，在高流量事件期间模拟TN载荷时观察到更高的误差。当模型用于水质模拟时，这项研究的结果对于不确定性下的决策具有重要意义。