| "For hydrological prediction and estimation in ungauged or data-poor basins, it is required that model parameters can be identified by using basin characteristics rather than calibrations. As the first step of exploring such a physically based identification method for the physically based distributed hydrological model, BTOPMC (Block-wise use of TOPMODEL with the Muskingum-Cunge method), preliminary studies on two aspects were carried out in this paper. First, through floods and annual runoff simulations for the Fuji-kawa and the Nakagawa basin of Japan, the effects of sub-basin scale on runoff simulation were investigated. It was found that smaller average sub-basin size produces higher peak discharge, larger baseflow and larger total runoff for floods, while similar effects on annual runoffs do not show any determined tendency and are not obvious, this is considered as the effect of averaging scale of the topographic index. Furthermore, when average sub-basin area is smaller than about 1/150 of the entire basin, simulation results seem stabilized, indicating a proper sub-basin scale for BTOPMC may exist, at which uncertainty due to subdivision level can be reduced to quite a low level. This subdivision level can be used as a referential value in establishing the relationship between model parameters and basin features. Second, for the four USA basins, comparisons were conducted between the two-year hydrographs obtained separately from the a priori estimation and automatic calibration; trials attempting to relate parameter values to basin features were carried out. The preliminary results of limited study cases suggest that it appears possible to establish the quantitative relations between parameter values and basin features for BTOPMC model." |
