抄録
Linear or non-linear system equations have been applied for the modeling of hydrological time series. However, the estimation of optimal parameters in the system is one of the most difficult processes in hydrological modeling, because the slight perturbation of the parameters in the system can cause remarkable difference of system solutions. Thereby, in order to assure the reliability of estimated system parameters, it is crucial to investigate the response of the system solutions against perturbation of the system parameters.
In this study, we conducted sensitivity analysis for the perturbation of parameters in non-linear system, using Self-Organizing Map. As the results, the time series generated from semi-optimal parameters were classified depending on their similarities. The parameters which generated approximately actual time series were partially visualized on the 2-Dimentional map.
