测量数据的不确定性对城市雨洪模型的影响

作者：C.B.S. Dotto , M. Kleidorfer, A. Deletic, W. Rauch, D.T. McCarthy

期刊：《水文学杂志》，2014年1月，508卷：28–42

关键字：输入和校准数据；城市排水；模型测量误差；敏感性分析；贝叶斯理论；参数概率分布

摘要：完善城市排水模型在实践中的应用关键在于，评价由于不同来源的误差导致模型的不确定性。本文探讨了输入和校准数据的误差对参数敏感性的影响，以及预测通过城市雨洪模型传递这些误差的不确定性。该城市雨洪模型是降雨径流模型和积聚/冲洗水质模型相互耦合作用的。本文建立误差模型来干扰测量数据的输入和校准，从而反映出这些类型数据集中共同的系统和随机不确定性。并且利用贝叶斯方法来分析模型的敏感性和不确定性分析。结果表明，测量数据的随机误差对模型的性能和灵敏度的影响较小。一般情况下，数据的输入和校准产生的系统误差影响参数分布，例如，改变参数分布的形状和峰值的位置。在大多数存在系统误差的情况下，尤其是通过最优假设代表输入和校准数据的不确定性的情况下，测量数据中误差完全有参数来补偿。但是，在利用极端的最坏的情况表示输入和校准数据的系统不确定性，参数是无法补偿的。因此，在这些少数最坏的情景下，模型的性能会大大的减弱。

Impacts of measured data uncertainty on urban stormwater models

Authors: C.B.S. Dotto , M. Kleidorfer, A. Deletic, W. Rauch, D.T. McCarthy

Journal: Journal of Hydrology，Volume 508, 16 January 2014, Pages 28–42

Keywords: Input and calibration data; Urban drainage; Modelling measurement errors; Sensitivity analysis; Bayesian inference; Parameter probability distributions

Abstract: Assessing uncertainties in models due to different sources of errors is crucial for advancing urban drainage modelling practice. This paper explores the impact of input and calibration data errors on the parameter sensitivity and predictive uncertainty by propagating these errors through an urban stormwater model (rainfall runoff model KAREN coupled with a build-up/wash-off water quality model). Error models were developed to disturb the measured input and calibration data to reflect common systematic and random uncertainties found in these types of datasets. A Bayesian approach was used for model sensitivity and uncertainty analysis. It was found that random errors in measured data had minor impact on the model performance and sensitivity. In general, systematic errors in input and calibration data impacted the parameter distributions (e.g. changed their shapes and location of peaks). In most of the systematic error scenarios (especially those where uncertainty in input and calibration data was represented using ‘best-case’ assumptions), the errors in measured data were fully compensated by the parameters. Parameters were unable to compensate in some of the scenarios where the systematic uncertainty in the input and calibration data were represented using extreme worst-case scenarios. As such, in these few worst case scenarios, the model’s performance was reduced considerably.

文献来源：http://www.sciencedirect.com/science/article/pii/S0022169413007440

翻译：王丽婷   校核：刘淼