抄録
A natural rainfall-runoff process is conceptualized by means of mathematical form reflecting a physical law and hydrologist's perception and it is typically calibrated and verified based on streamflow data, the commonly used catchment response. The streamflow data is obviously required, but is not sufficient to identify a conceptual parameter of a hydrologic model since numerous parameter combinations can often result in quite similar numerical values in terms of objective function and even indistinguishable simulated hydrographs. One of the efficient techniques to resolve this impasse is a combination of the model with streamflow data augmented by other kinds of hydrological information relevant to the prediction. In this study, internal dynamic response of catchment, spatiotemporally separated components of hydrograph, to the various mimic parameter sets is presented. It is concluded that all of simulations based on plausible parameter combinations sampled by using deterministic and stochastic automatic optimization algorithms, are performed equally well in terms of both model performance measure and visual comparison between observed and simulated hydrograph while internal behaviors of catchment show totally different aspects. The spatiotemporal information presented here can be utilized as one of the complementary constraints capable of filtering out non-physical parameter set (s) .
