Study on parameter uncertainty of hydrological model based on factorial analysis and variance decomposition method

Hydrological model is an important tool for simulating hydrological process and understanding hydrological law. It has important practical significance for runoff-yield calculation and hydrological forecasting. But the uncertainty of model parameters reduces the adaptability of simulation results. Therefore, this paper uses factorial analysis variance decomposition method to quantitatively explore the influence of uncertainty and transfer of SWAT model parameters in low dimensional space. The results show that the parameters uncertainty is larger in the lower flow than in the higher flow; Precipitation and underlying surface affect the hydrological response and thus the most sensitive parameters are TRANSRCH, CH_K2 and SOL_K in turn; the interaction effect between TRANSRCH and SOL_K, CH_K2 and SOL_K also has significant on runoff simulation. The results can provide a theoretical guidance for the study on dynamic parameters of model, in the hope of reduce the parameter uncertainty of hydrological model.