摘要
This study examined changes in the division responses caused by imperviousness changes in time. Each response is represented by an instantaneous unit hydrograph with storage parameters. The critical work in this study was to ensure that the calibrated storage coefficients reasonably respond to their corresponding imperviousness. Estimating the spatially different mean rainfall of each division by adopting traditional methods (e.g., the Thiessen polygon method) may be difficult because of the lack of raingauge data in a division. The storages cannot be measured directly and were generally calibrated using an optimization method with parameter bounds. The upper bounds, which imply an ultimate of each corresponding physical property, are particularly difficult to determine directly. This study applied block Kriging and nonlinear programming to estimate spatiotemporal effective rainfall. An approach was proposed to determine suitable upper bounds for each storage parameter in a calibration. An examination of the calibrated storages at three stages revealed that the values resulting from a calibration with suitable bounds are more reliable compared with those based on a calibration without bound limitation. Finally, we concluded that (1) suitable bounds are essential for ensuring that significant calibrated storages can be applied to achieve a favorable regression result; (2) the analytic results indicate that large changes in imperviousness in downstream areas reduced the time to peak by at least 15?%, increasing the peak discharge by nearly 40?%; and (3) hydrograph characteristics of the watershed outlet reduced the time to peak by approximately 10?%, increasing the peak discharge by nearly 15?%.