Particle Swarm Optimization and Shuffle Complex Evolution for Calibrating Xinanjiang Model Parameters

Abstract

Xinanjiang model, a conceptual hydrological model, is well known and widely used in China since 1970s. Xinanjiang model consists of large number of parameters that cannot be directly obtained from measurable quantities of catchment characteristics, but only through model calibration. Parameter optimization is a significant but time-consuming process that is inherent in conceptual hydrological models representing rainfall–runoff processes. This study presents newly developed Particle Swarm Optimization (PSO) and compared with famous Shuffle Complex Evolution (SCE) to auto-calibrate Xinanjiang model parameters. The selected study area is Bedup Basin, located at Samarahan Division, Sarawak, Malaysia. Input data used for model calibration are daily rainfall data Year 2001, and validated with data Year 1990, 1992, 2000, 2002 and 2003. Simulation results are measured with Coefficient of Correlation (R) and Nash-Sutcliffe coefficient (E²). Results show that the performance of PSO is comparable with the famous SCE algorithm. For model calibration, the best R andE² obtained are 0.775 and 0.715 respectively, compared to R=0.664 and E²=0.677 for SCE. For model validation, the average R=0.859 and average E²=0.892 are obtained for PSO, compared to average R=0.572 and average E² =0.631 obtained for SCE.