Received: 14 Jul 2020 – Discussion started: 17 Aug 2020
Abstract. A new Bayesian framework for automatic calibration of SWMM, which simultaneously considers both parameter uncertainty and input uncertainty, was developed. The framework coupled an optimization algorithm DREAM and the Storm Water Management Model (SWMM) by newly developed API functions used to obtain and adjust parameter values of the model. Besides, a rainfall error model was integrated into the framework to consider systematic rainfall errors. A case study in Guangzhou, China was conducted to demonstrate the use of the framework. The calibration capability of the framework was tested and the impacts of rainfall uncertainty on model parameter estimations and simulated runoff boundaries were identified in the study area. the results show that calibration considering both parameter uncertainty and rainfall uncertainty captures peak flow much better and is more robust in terms of the Nash Sutcliffe index than that only considering parameter uncertainty.
How to cite. Gao, X., Yang, Z., Han, D., Huang, G., and Zhu, Q.: A Framework for Automatic Calibration of SWMM Considering Input Uncertainty, Hydrol. Earth Syst. Sci. Discuss. [preprint], https://doi.org/10.5194/hess-2020-367, 2020.
Input errors and parameter errors are two main sources of uncertainties in hydrological model calibration. We developed a new Bayesian framework for automatic calibration of the Storm Water Management Model (SWMM), simultaneously considering parameter and input uncertainties and verified the framework with a case study. The results shows that calibration considering both parameter and input uncertainties captures peak flow much better that only considering parameter uncertainty.
Input errors and parameter errors are two main sources of uncertainties in hydrological model...