The uncertainties associated with the simulation models are often ignored in operational hydrology. While many methods are available for evaluation of simulation uncertainty, most of them focus on construction of prediction bands, which alone may not be sufficient to make effective decisions. This is a serious concern in watershed management planning, especially in cases where the models are uncalibrated due to unavailability of observations. This study addressed uncertainty in hydrologic modeling, and its consideration in check dam design decisions. Size of the check dams were determined using a simulation–optimization framework by considering dual objectives of maximizing water availability for agriculture and minimizing the adverse effects on downstream reaches. The optimizer suggested distinct Pareto-optimal-front for different parameter sets of the model (in turn resulting in different simulations), indicating the influence of parametric uncertainty on the design. An analysis of the optimal solutions suggested varying check dam sizes (0.5–1.5 m) for similar objective function value, which plausibly indicate an economic impact. Nonetheless, the effectiveness of the structure (in terms of simulated wet and dry period lengths) did not exhibit significant variability across the designs (average wet period length of > 100 days). The median of the streamflow ensemble provided satisfactory performance (> 100 days wet period length and only 11–25% reduction of flow to downstream) and could be a viable choice for implementation. The results suggest that parametric uncertainty that is propagated to prediction uncertainty significantly influences the final design decisions and calls for careful assessment prior to implementation.

在运营水文学中，与模拟模型相关的不确定性通常被忽略。尽管有许多方法可用于评估仿真不确定性，但大多数方法都着眼于预测带的构建，仅靠预测带可能不足以做出有效的决策。这是流域管理规划中的一个严重问题，尤其是在由于观测不到而无法对模型进行校准的情况下。这项研究解决了水文建模中的不确定性，以及在检查大坝设计决策中的考虑。通过模拟优化框架，通过考虑使农业用水最大化和对下游河段的不利影响最小化的双重目标，确定了检查坝的大小。优化器为模型的不同参数集建议了截然不同的Pareto最优前沿（进而导致了不同的仿真），表明了参数不确定性对设计的影响。对最佳解决方案的分析表明，对于相似的目标函数值，检查坝尺寸（0.5-1.5 m）有所不同，这有可能表明存在经济影响。尽管如此，结构的有效性（根据模拟的湿期和干燥期长度）并未在整个设计中表现出明显的可变性（平均湿期长度> 100天）。流量合计的中值提供了令人满意的性能（> 100天的湿期长度，并且仅减少了11-25％的下游流量），并且可能是实施的可行选择。