点击次数：

论文类型：期刊论文

发表时间：2009-09-01

发表刊物：Sichuan Daxue Xuebao (Gongcheng Kexue Ban)/Journal of Sichuan University (Engineering Science Edition)

收录刊物：Scopus、EI、PKU、ISTIC

卷号：41

期号：SUPPL. 2

页面范围：106-114

ISSN号：10093087

关键字：Australia; Automatic calibration; Automatic optimization; Complex behavior; Hydrographs; Hydrologic models; Low flow; Multi objective; Multi-objective calibration; Multiple objectives; Objective functions; Peak flows; Performance measure; Rainfall-runoff models; Runoff prediction; Shuffled Complex Evolution; Single objective; Ungauged catchments; Xinanjiang model, Catchments; Hydraulic models; Multiobjective optimization; Rain; Runoff, Calibration

摘要：Automatic calibration for hydrologic models with multiple objective capabilities is becoming increasingly popular due to the recognition that a single performance measure is no longer sufficient to characterize the complex behavior of the catchment. This paper presents the multiple objectives calibration of Xinanjiang rainfall-runoff model for 210 relatively unimpaired catchments in the south-east Australia. The calibration includes optimization of multiple objectives that measures different aspects of the hydrograph: 1) peak flow; 2) low flow. An automatic optimization procedure based on the Multi-Objective Shuffled Complex Evolution Metropolis (MOSCEM) algorithm is used for solving the multi-objective calibration problem. The benefits of multi-objective calibration are illustrated through comparison with traditional single objective calibrations. The results show that the single-objective calibration matches one aspect of the hydrograph well but is inadequate in the other aspect; the multi-objective calibration corresponding to a proposed balanced aggregated objective function is seen to provide a proper balance between the two objectives and matches both aspects well. Moreover, multiple objectives optimization provides the smallest relative volume error between the observed and simulated streamflow for calibration, verification and regionalization, which further demonstrates multi-objective calibration outperforming single-objective calibration.